By Nic Lewis
Introduction
Many people, myself included, thought that in the many regions where COVID-19 infections were consistently reducing during the summer, indicating that the applicable herd immunity threshold had apparently been crossed, it was unlikely that a major second wave would occur. This thinking has been proved wrong. In this article I give an explanation of why I think major second waves have happened.
Key points
- The herd immunity threshold (HIT) depends positively on the basic reproduction number R0 and negatively on heterogeneity in susceptibility.
- Since neither of the factors on which the HIT depends are fixed, the HIT is not fixed either.
- R0 depends on biological, environmental and sociological factors; colder weather and the evolution of more transmissible strains likely both increase R0; more (less) cautious behaviour and social distancing / restrictions on mixing reduce (increase) R0.
- Second waves were due primarily to changes in these factors increasing R0 and thus the HIT from below to above the existing level of population immunity.
- Heterogeneity in susceptibility is partly biological, but social connectivity differences are key.
- The effect of heterogeneity in susceptibility on the HIT can be represented by a single parameter λ.
- λ will always exceed 1 (its level in a homogeneous population); pre-epidemic λ may be ~4. The higher λ is, the lower the HIT for any given R0.
- The natural infection HIT is hence bound to be below the level of {1 – 1/R0} quoted by ‘experts’.
- Government restrictions reduce λ as well as R0, so the HIT falls less than it would if λ were fixed.
- The final size of an uncontrolled epidemic will substantially exceed the HIT, due to overshoot, so high reported seroprevalence levels can be consistent with a much lower HIT.
The herd immunity threshold (HIT) for a disease epidemic is the proportion of the population needing to have been infected, and thereby no longer susceptible to infection, before the rate of new infections starts to decline. The HIT depends both on the basic reproduction number for infections (R0) – the number of other people that at the start of an epidemic an infected person will on average infect – and the degree of heterogeneity in individuals’ likelihood of being infected (their susceptibility). That likelihood in turn depends on both their social connectivity and biological susceptibility to infection. Neither R0 nor the degree of heterogeneity in susceptibility is fixed in value, so the HIT is not fixed either.
Changes in population behaviour – whether arising from government interventions or in response to increasing disease incidence – affect both R0 and heterogeneity in susceptibility. In addition, R0 (which is proportional to how readily infection is on average transmitted between individuals) may vary seasonally, and change as the virus or other infectious organism mutates.
The resurgence of COVID-19 infections in a second wave after the summer ended is almost certainly due to some combination of the foregoing sociological and biological factors. It has been claimed that the influence of weather on its transmission is relatively minor,[1] and it has so far proved difficult to detect seasonality for COVID-19.[2] However, common colds caused by other coronaviruses are highly seasonal and I now think that it is reasonable to work on the basis that COVID-19 shares that behaviour.
I focus in this article on the mathematical dependence of the HIT to R0 and heterogeneity in susceptibility, and on the factors influencing those controlling variables. I also touch on difference between the HIT and the final size of an uncontrolled epidemic. I discuss in an appendix how, in my view, changes in the factors influencing R0 and heterogeneity in susceptibility likely shaped the evolution of the epidemic in western Europe
How the HIT varies with R0 and population heterogeneity
Table 1 illustrates how the herd immunity threshold varies with R0 and population heterogeneity in susceptibility to infection. The effect of such heterogeneity on transmission of infection and on the HIT can be represented by a single parameter λ, the heterogeneity factor (Tkachenko et al. 2020)[3], which is a function of population variability in both social connectivity and in biological susceptibility.[4] The reproduction number at any time, Rt, and the HIT are related as follows to R0 and λ:
Rt = R0 × Sλ
HIT = 1 – (1/R0)1/λ
where S is the proportion of the population that remains susceptible to infection. For a homogeneous population, these formulae reduce to the classical results Rt = R0 × S and HIT = 1 – 1/R0. With heterogeneity in susceptibility to infection, Rt falls more than pro rata to the susceptible proportion S decreases. Initially, Rt falls λ times as fast with S as in the homogeneous case.
Note that an epidemic takes some time to die out after the HIT is reached, since at that point many people will be infected and will go on to infect others, albeit at a declining rate. Therefore, the final size of the epidemic (FSE) – the attack rate (the ultimate proportion of the population that has been infected) – will exceed the HIT. The columns to the right of each HIT column show (in italics) the FSE if social and biological factors remain unchanged throughout the epidemic.[5] As shown in a previous article,[6] well timed short term restrictions to reduce transmission as the HIT is approached can prevent the FSE from significantly overshooting the HIT.
Table 1. Relationship of each of the herd immunity threshold (HIT) and the final size of the epidemic (FSE) with the basic reproduction number R0, at varying levels of heterogeneity factor λ that arises from heterogeneity in susceptibility (assumed gamma-distributed) across the population, from none (λ = 1) to an estimated normal level (λ = 4). The FSE values assume that the same R0 and λ value applied throughout the epidemic.
Since a person’s social connectivity, which reflects their average rate of contacts with others, equally affects their infectivity, variability in it has a more powerful effect than variability in biological susceptibility.[7] Note that heterogeneity in infectivity that is uncorrelated with susceptibility does not affect the overall progression of an established, large epidemic, although it may affect smaller scale features such as clustering of cases.
For a population that is homogeneous in both biological and social components of susceptibility, λ = 1 (pink columns). In that case, the ‘classical’ formula HIT = 1 – 1/ R0 is valid. This formula also applies to immunity gained through vaccination at random, since such vaccination – unlike natural disease progression – does not preferentially confer immunity on individuals who are more susceptible to infection (and also more likely to infect others).
Analyses of contact networks indicate that, in normal circumstances, the coefficient of variation (standard deviation / mean) for social connectivity in a population is about 1, while biological susceptibility is likely to have a coefficient of variation of about 1/3 or more (Tkachenko et al). Use of those figures implies that λ = 4 (green, rightmost columns).
The effect of government social distancing measures on R0 and the heterogeneity factor
It has been estimated that, prior to significant social distancing taking place, 80% to 90% of all transmission of infection is caused by circa 10% of infected individuals, often at superspreading events where a large number of people are present. When restrictions on gatherings, bars and other venues are introduced, non-household social mixing generally is reduced and superspreading opportunities fall even further, while household mixing will be little affected. The result will be a reduction in R0, but also reduced heterogeneity in social connectivity and hence λ. A further reduction in both these factors can be expected to occur when a lockdown (stay-at-home order) is introduced.
The effects of such government measures, for a range of resulting R0 values, are illustrated by the two middle sets of columns. These both assume the same 1/3 coefficient of variation for biological susceptibility, but a reduction in the coefficient of variation for social connectivity to 0.625, resulting in λ = 3 (yellow columns) or to 0.25, resulting in λ = 2 (salmon columns).
Even in the absence of legal restrictions being imposed, people can be expected to significantly change their behaviour when an epidemic involving severe disease takes hold. The resulting reduction in λ, for any given resulting reduction in R0, might however be less than under an enforced reduction in mixing, since more gregarious people may be less cautious and reduce their high social mixing proportionately less than more cautious, less gregarious people do – the opposite relationship to that arising from restrictions on gatherings, bars and other venues.
How a high seroprevalence level can arise even in the presence of substantial heterogeneity
It might be thought that a high attack rate is incompatible with significant population heterogeneity in susceptibility and hence a moderate HIT. An attack rate of 76% has been claimed for the city of Manaus.[8] However, the weighted measured seroprevalence on which that estimate was based was not from a random sample nor representative of the population,[9] and never exceeded 44%[10]. A random population survey found seroprevalence in Manaus to be only between one-quarter one-third the level claimed in the foregoing study, casting severe doubt on its claim.[11]
The first mentioned study also estimated that in or just after mid-March, near the start of the epidemic in Manaus, Rt – which at that point would not have been far short of R0 – was approximately 2.5, suggesting R0 was in the 2.6 to 2.8 range. The extent of physical distancing that they estimated applied then was moderate, similar to that near the end of the main epidemic. In a relatively poor city like Manaus with household and transport crowding it seems quite likely that in normal circumstances there is lower population heterogeneity in social connectivity than in a high income city, indicating an heterogeneity factor λ perhaps more like 3 than 4 (yellow not green columns). And under moderate social distancing the heterogeneity factor λ might be closer to 2 than 3. For an R0 of 2.6, λ = 2 implies an HIT of 38% but a final epidemic size (FSE) of 64%[12]. Even at λ = 3, the FSE would be 49% (with an HIT of 27%).[13]
To summarize, it seems doubtful that the attack rate in Manaus in fact exceeded 50% – it may have been no more than 20-25% – and an attack rate of 50% is fully compatible with the HIT being below 30%.
Appendix – Changes in R0 and population heterogeneity during the epidemic
The following discussion, which represents my semi-quantitative broad brush analysis of what has occurred, relates primarily to the progress of the epidemic in western Europe. However, it may also be somewhat applicable to the north east United States, where the epidemic took off only slightly later than in western Europe and where the seasonal variation in climate is also large.
In the initial stages of the first wave, which generally started in major cities, in early spring 2020, infections appear to have been doubling every three days or so prior to governments imposing restrictions or people becoming significantly more cautious. Depending on the assumed distribution of the generation interval (from one infection to those it directly leads to), that implies an R0 value of between 2 and 4.[14] I will assume a middle of the range R0 value of 3 for illustrative purposes. That would imply a HIT of 67% for a homogeneous population, reducing to 24% for a population with the highest degree of heterogeneity illustrated in Table 1, which might be expected to apply before people started behaving more cautiously and mixing less.
When people started mixing less, voluntarily or by government fiat, R0 would have reduced, but as discussed above λ will also have fallen. The combined effect of these changes can be visualised as moving diagonally upwards and leftwards in Table 1, from the green columns to the yellow columns and then to the salmon columns. The resulting reduction in the HIT would therefore be somewhat smaller than that implied by the reduction in R0 alone.
By late spring or early summer the first wave had largely faded, and it generally continued to decline after restrictions on mixing were at least partially relaxed. As summer progressed, people’s behaviour unsurprisingly returned closer to pre-epidemic norms. I will assume for illustrative purposes that the yellow columns (λ = 3) were representative of that period. Since by midsummer the epidemic appears to have been declining even where only a minor first wave had occurred, it seems that R0 must generally have declined to 1 or below, so that population immunity levels would everywhere have exceeded the HIT (which is only positive if R0 > 1).
As autumn arrived, infections and then serious illness started to rise again, although where testing was increasing the rise may have been exaggerated. It follows that R0 must have risen again, resulting in the HIT increasing to above the level of population immunity. An obvious explanation for the rise in R0 is seasonally reduced sun and cooler weather, with more contact occurring indoors, where almost all COVID-19 transmission appears to take place. A major increase in mixing among young people as school and, particularly, university terms started likely also boosted R0 and the level of infections in the autumn; young adults have generally had the highest incidence rates during the second wave.[15] In some places the rise in infections appears to have occurred slightly earlier, perhaps as a result of holidaymakers returning infected from areas where COVID-19 was more prevalent.[16]
Initially it seemed that some large cities where a significant proportion of the population had been infected in the first wave might be spared, but in most cases the increase in R0 evidently became sufficiently large to raise the HIT to above the level of population immunity. As a result of increasing infections, government-imposed restrictions were generally increased, which as well as reducing R0 will also have reduced the heterogeneity factor λ. This can be visualised as a move diagonally upwards from the yellow columns to the salmon columns. Those actions appear typically to have pushed Rt down to about 1, or slightly lower, which in the presence of a reasonable degree of existing population immunity implies an R0 level significantly above 1. With reduced heterogeneity, the existing level of population immunity causes a lesser reduction in Rt, relative to R0, but Rt will still be a smaller fraction of R0 than the proportion of the population that remains susceptible to infection.
In the UK, and possibly various other countries, a new lineage (B.1.1.7) of the SARS-CoV-2 virus has now emerged[17] and grown faster than existing ones, as discussed in a previous article[18]. Since writing that article, some further data has provided less indirect evidence that B.1.1.7 is 25–50% more infectious than pre-existing variants.[19] On the other hand, recent data from the regions where B.1.1.7 has become dominant suggests that it may now be growing no faster than other variants.[20] It has been suggested that the fast growth in the regions where B.1.1.7 now dominates may have been at least partly due to it spreading there in schools.[21] However, making for illustrative purposes the assumption that B.1.1.7 is actually 25–50% more infectious, R0 will have been increasing, perhaps typically reaching somewhere in the range1.5 to 2.0 once B.1.1.7 becomes the dominant variant, if R0 was previously in the 1.2 to 1.4 range.
Tougher restrictions that have been introduced in a number of countries in response to infection rates increasing, whether due to the spread of the B.1.1.7 lineage, to cold winter weather or to greater mixing, will have reduced population heterogeneity in social connectivity further. In these circumstances, is unclear whether existing levels of population immunity will suffice to prevent further growth of the B.1.1.7 lineage, or the rather similar one that has emerged in South African, even with severe restrictions being introduced. However, increased population immunity resulting from some combination of further spread of infections and vaccination programmes, the combination varying from one country and region to another, should bring COVID-19 epidemics under control within the next few months.
Nicholas Lewis 10 January 2021
[1] “All pharmaceutical and non-pharmaceutical interventions are currently believed to have a stronger impact on transmission over space and time than any environmental driver.” Carlson CJ, Gomez AC, Bansal S, Ryan SJ. Misconceptions about weather and seasonality must not misguide COVID-19 response. Nature Communications. 2020 Aug 27;11(1):1-4. https://doi.org/10.1038/s41467-020-18150-z
[2] Engelbrecht FA, Scholes RJ. Test for Covid-19 seasonality and the risk of second waves. One Health. 2020 Nov 29:100202. https://doi.org/10.1016/j.onehlt.2020.100202
[3] Tkachenko, A.V. et al.: Persistent heterogeneity not short-term overdispersion determines herd immunity to COVID-19. medRxiv 29 July 2020 https://doi.org/10.1101/2020.07.26.20162420 They use the term ‘immunity factor’ for λ. Equations [11)], [12] and [13] and intervening paragraph. I adopt their assumption that there is negligible correlation across the population between biological susceptibility to infection and either social connectivity or biological infectivity.
[4] I make from here on the common assumption that a gamma distribution can well represent variation within the population in both social connectivity and biological susceptibility, on which basis λ = (1 + 2 × CVs2) × (1 + CVb2) where CVs and CVb are respectively the social and biological coefficients of variation (standard deviation / mean) for the population.
[5] The FSE (1 – S∞) depends on the sum of the squared coefficients of variation η = CVs2 + CVb2 as well as on λ. It is given by the solution to the equation S∞ = (1 + R0 η [1–S∞λ–η]/[ λ–η])–1/η. See Tkachenko et al. equation [17].
[6] https://www.nicholaslewis.org/when-does-government-intervention-make-sense-for-covid-19/
[7] Variability in infectivity that is uncorrelated with susceptibility in the population has no overall effect in a sizeable epidemic.
[8] Buss, Lewis F., et al. “Three-quarters attack rate of SARS-CoV-2 in the Brazilian Amazon during a largely unmitigated epidemic.” Science (2020).
[9] It was a convenience sample, comprised entirely of blood donors.
[10] That maximum seroprevalence estimate was adjusted upwards to 52% to account for test for sensitivity and specificity. The attack rate estimate further assumed that antibodies would no longer be detectable in a proportion of previously infected individuals.
[11] Hallal, P.C. et al:SARS-CoV-2 antibody prevalence in Brazil: results from two successive nationwide serological household surveys. Lancet, 8(11), e1390-e1398,, September 2020 https://doi.org/10.1016/S2214-109X(20)30387-9
[12] Actually slightly lower, as the stricter social distancing measures in the middle part of the epidemic would have reduced the excess of the FSE over the HIT.
[13] If R0 = 2.0, which is possible if a shorter estimate of the generation interval is used, the corresponding FSE sizes would be 52% or 38%, with the HIT being respectively 29% or 21%.
[14] Assuming a gamma distributed generation interval with a mean in the range 4 to 6.5 and a coefficient of variation between 0.37 and 0.74.
[15] Aleta A, Moreno Y. Age differential analysis of COVID-19 second wave in Europe reveals highest incidence among young adults. medRxiv. 13 November 2020. https://doi.org/10.1101/2020.11.11.20230177
[16] It is also possible that, notwithstanding a published finding to the contrary, the A20.EU1 variant that was brought back from Spain by people infected on holiday there may have been somewhat more infectious than existing variants.
[17] Other evidence that has now become available suggests that a similar variant arose in Italy prior to B.1.1.7 being detected in the UK.
[18] https://www.nicholaslewis.org/the-relative-infectivity-of-the-new-uk-variant-of-sars-cov-2/
[19] The observed 50–70% increase in weekly growth rate corresponds to roughly a 25–50% increase in infectivity (and hence in R0), assuming a generation interval with a 4–6 day mean and a reasonable CV, if R0 was previously not substantially above 1.
[20] https://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/adhocs/12722estimatesofcovid19casesto02januaryforenglandregionsofenglandandbycasescompatiblewiththenewvariant
[21] Loftus (2021, Jan. 1). Neurath’s Speedboat: Did the new variant of COVID spread through schools? Retrieved from http://joshualoftus.com/posts/2021-01-01-did-the-new-variant-of-covid-spread-through-schools/
Originally posted here, where a pdf copy is also available
But did not happen where they use Ivermectin.
FYI –
The Association of American Physicians and Surgeons posted on their website a paper/brochure advising treatments at home for Covid-19 patients (a step-by-step doctors’ plan).
They recommend two treatments; one features Hydroxychloroquine and the other features Ivermectin.
https://aapsonline.org/CovidPatientTreatmentGuide.pdf
do they give the statistical evidence that it “works”.?
“Do they give statistical evidence that it works”?
Would they?
If “they” can do all sorts of “studies” to prove that HCQ is dangerous and doesn’t work, and if they can publish a paper in the Lancet on HCQ (proving how terrible it is) that was so bad that it had to be retracted, then why would we believe that anyone would care to do honest studies on either HCQ or Ivermectin, both of which are relatively cheap and have been used safely for years? https://ahrp.org/how-a-false-hydroxychloroquine-narrative-was-created/
The point seems to be to discredit any alternative therapies that would take glory, and profits, away from vaccines.
Ivermectin – one might hope it works better on viruses than on demodex. I used that for six months intermittently, to no effect. Not bloody likely.
Oxford University has just started a series of trials on this as yet unproven drug
https://lockdownsceptics.org/2021/01/24/latest-news-264/#ivermectin-miracle-cure-or-snake-oil
tonyb
Wish someone could translate to common understandable words.
> This thinking has been proved wrong.
I [Nic Lewis] was wrong.
Can’t seem to get this to post in the right place.
ATTP’s reference shows that herd immunity is indeed properly used as I have used it. VTG’s insistence I’m wrong is not rational and contradicted by all the evidence.
“The term “herd immunity” is widely used but carries a variety of meanings [1–7]. Some authors use it to describe the proportion immune among individuals in a population. Others use it with reference to a particular threshold proportion of immune individuals that should lead to a decline in incidence of infection. Still others use it to refer to a pattern of immunity that should protect a population from invasion of a new infection. A common implication of the term is that the risk of infection among susceptible individuals in a population is reduced by the presence and proximity of immune individuals (this is sometimes referred to as “indirect protection” or a “herd effect”). We provide brief historical, epidemiologic, theoretical, and pragmatic public health perspectives on this concept.”
VTG has nothing on IFR either other than word salads unsupported by any references or any science. 0.03% is easy to explain in a place like Africa where the average age is 20 or so.
ATTP’s reference shows that herd immunity is indeed properly used as I have used it. VTG’s insistence I’m wrong is not rational and contradicted by all the evidence.
“The term “herd immunity” is widely used but carries a variety of meanings [1–7]. Some authors use it to describe the proportion immune among individuals in a population. Others use it with reference to a particular threshold proportion of immune individuals that should lead to a decline in incidence of infection. Still others use it to refer to a pattern of immunity that should protect a population from invasion of a new infection. A common implication of the term is that the risk of infection among susceptible individuals in a population is reduced by the presence and proximity of immune individuals (this is sometimes referred to as “indirect protection” or a “herd effect”). We provide brief historical, epidemiologic, theoretical, and pragmatic public health perspectives on this concept.”
VTG has nothing on IFR either other than word salads unsupported by any references or any science. 0.03% is easy to explain in a place like Africa where the average age is 20 or so.
VTG is part of the despicable public attack on Ioannidis. These attacks are personal, nasty, and political. Partisan hacks should be ashamed of themselves for attacking one of the lions of medical science.
Layman’s summary: Epidemeology is very uncertain and the problem of predicting the progress of any epidemic is ill-posed and highly sensitive to parameters and assumptions. Nic did a good job of showing this in this post.
Nic Lewis chose a set of model parameters which could fit then current data and claim herd immunity based on a particular locality.
Other localities and indeed choices of parameters could not support this claim, and as expected by almost the entire epidemiological community it has indeed proved false.
In the meantime, right wing media outlets used this and other similar analyses to push a false and damaging claim that the pandemic could be ended through infection aquired herd immunity.
How many lives have been lost as a result is a matter for speculation.
Parallels with the climate change public debate are notable.
What a silly nonscientific comment. There has been no proof that herd immunity was not achieved in many places this summer and then R increased because of winter and a more contageous strain and HIT went up too. HIT can vary a lot as the current post shows.
It’s the simplest explanation of the common pattern of most European countries and many US states of an initial wave which then abated in the summer (in many places going down to negligible deaths) and then a strong second wave. For example California and New York have had increasingly severe lockdowns and show this patter with California in a very strong growth phase right now.
In any case, there are by now many many papers showing that strong mitigation measures don’t make a significant difference anyway. I cited a new one below. One could just as easily argue that strong mitigation has caused more excess deaths than it has prevented.
dpy,
Thank you for your technical comment. The mathematic treatment is compelling.
In a more serious note, your unscientific attempt to redefine herd immunity is noted.
The Ionnadis paper you cited on suppression is as bad an attempt to critique virus suppression as his repeated failed attempts to deny reality on IFR. That’s a high bar! I note even you seem to have given up defending his IFR position. There’s a lesson for you there if you choose to find it.
Another vacuous response. Herd immunity is reached when R is less than 1 and infections trend toward zero. It varies a lot depending on many things. There are large numbers of papers digging into the details. Why do you bother with such blatantly political attacks? If you think you are convincing anyone you are wrong.
You have no credibility in judging any science statistical result. Nic Lewis is a top flight professional stastician,
Nic Lewis was wrong.
Herd immunity was not reached in Sweden.
Nobody else defines herd immunity as reducing cases whilst restrictions are still in place. It’s a ludicrous attempt at justifying a failed hypothesis. Cases rose throughout Europe during summer once restrictions were lifted.
Ionnadis was wrong about IFR and is now putting forward a risible proposition that South Korea and Sweden had equivalent responses(!)
There’s nothing political here except those promoting these failed ideas. Like the promoters of Great Barrington Declaration, a document sponsored by a libertarian think tank, lest we forget.
Learn your lessons: cherry picking science based on your politics leads to failure. Go apply it to climate change.
A quick google search shows how completely wrong and simple minded about science you are. I found one that is for laymen from Nature:
“Reaching herd immunity depends in part on what’s happening in the population. Calculations of the threshold are very sensitive to the values of R, Kwok says. In June, he and his colleagues published a letter to the editor in the Journal of Infection that demonstrates this4. Kwok and his team estimated the Rt in more than 30 countries, using data on the daily number of new COVID-19 cases from March. They then used these values to calculate a threshold for herd immunity in each country’s population. The numbers ranged from as high as 85% in Bahrain, with its then-Rt of 6.64, to as low as 5.66% in Kuwait, where the Rt was 1.06. Kuwait’s low numbers reflected the fact that it was putting in place lots of measures to control the virus, such as establishing local curfews and banning commercial flights from many countries. If the country stopped those measures, Kwok says, the herd-immunity threshold would go up.”
Ioannidis’ peer reviewed paper on this subject gives a wide range of IFR’s based on seroprevalence studies from 0.03% to 0.5%. It is simple minded and unscientific to try to find a single number. For a disease like this that is strongly age related, the profile of the population will give a wide range of numbers. There are by now a multitude of papers and studies that show that in healthy populations under 70 (for example blood donors) the IFR is in the 0.07% range. For the elderly and ill, its much higher. But they have a very high mortality rate in any case.
You are the politically motivated activist here. The scientists who you slander are vastly more honest than you. And they actually know statistics and science.
I personally don’t even know why you persist in this concensus enforcement activity as you will convince no one with your totally unsupported wrong assertions.
dpy,
drop the victimhood? Pointing out you’re wrong isn’t enforcing anything. Relying, as you consistently do, on fringe science which aligns with your politics is a recipe for being consistently wrong.
Ionaddis IFR numbers are, and have been throughout, an outlier. 0.03%!
The interpretation of herd immunity you copy clearly is not compatible with no second wave hypothesis, so if used needs to be clearly in that context.
FWIW, the herd immunity threshold is typically defined in terms of R0. If we implement interventions that change our behaviour and, consequently, change the effective R number and limit the spread of the infection, the herd immunity threshold is still defined by R0, not by the effective R. The latter would only be the case if the interventions were essentially permanent and, hence, had changed R0. For example, this paper.
https://academic.oup.com/cid/article/52/7/911/299077
David –
> “Reaching herd immunity depends in part on what’s happening in the population.
Imagine a scenario where a boat with 1000 people aboard get stranded on a desert isle.
1 has COVID at the time of the shipwreck and then infects three others and then each of those 3 others infect two others after that. So a total of 10 have been infected.
So then everyone isolates and 2 of those infected died and the rest recovered and developed antibodies and immunity and there’s no more virus on the island.
So, according to your thinking, 0.1% of the population on the island has been infected and the island population has reached herd immunity.
Basically, you’re equating “herd immunity” with infection level. You could call any prevalence of infection “herd immunity” depending on the conditions.
Nic stated, that Stockholm had (likely, but with high confidence) reached “herd immunity,” specifically because he had identified the people in Stockholm as engaging in near normal-level behaviors. He has re-stated such in this thread. If you have reached “herd immunity” in near-normal level behaviors, changes in behavior (presumably because of season) would not result in the subsequent level of growth in infection rate that they’ve seen in Sweden since the date when Nic said they’d reached “herd immunity.” Well, unless those changes in behaviors included something like going around and deliberately infecting people
For the term “herd immunity” to have some meaning it needs to mean more than “any level of prevalence of infection I want it to mean no matter the context.” To be meaningful, it needs to have some implication w/r/t the level of confidence people can feel that they won’t likely get infected under normal (or near-normal) conditions (with the understanding of course, that with overshoot there’s still some level of potential to get infected).
Referring to an out of context “definition” is a weak rhetorical device.
ATTP’s reference shows that herd immunity is indeed properly used as I have used it. VTG’s insistence I’m wrong is not rational and contradicted by all the evidence.
“The term “herd immunity” is widely used but carries a variety of meanings [1–7]. Some authors use it to describe the proportion immune among individuals in a population. Others use it with reference to a particular threshold proportion of immune individuals that should lead to a decline in incidence of infection. Still others use it to refer to a pattern of immunity that should protect a population from invasion of a new infection. A common implication of the term is that the risk of infection among susceptible individuals in a population is reduced by the presence and proximity of immune individuals (this is sometimes referred to as “indirect protection” or a “herd effect”). We provide brief historical, epidemiologic, theoretical, and pragmatic public health perspectives on this concept.”
VTG has nothing on IFR either other than word salads unsupported by any references or any science. 0.03% is easy to explain in a place like Africa where the average age is 20 or so.
VTG is part of the despicable public attack on Ioannidis. These attacks are personal, nasty, and political. Partisan hacks should be ashamed of themselves for attacking one of the lions of medical science.
Posting in the correct place.
ATTP’s reference shows that herd immunity is indeed properly used as I have used it. VTG’s insistence I’m wrong is not rational and contradicted by all the evidence.
“The term “herd immunity” is widely used but carries a variety of meanings [1–7]. Some authors use it to describe the proportion immune among individuals in a population. Others use it with reference to a particular threshold proportion of immune individuals that should lead to a decline in incidence of infection. Still others use it to refer to a pattern of immunity that should protect a population from invasion of a new infection. A common implication of the term is that the risk of infection among susceptible individuals in a population is reduced by the presence and proximity of immune individuals (this is sometimes referred to as “indirect protection” or a “herd effect”). We provide brief historical, epidemiologic, theoretical, and pragmatic public health perspectives on this concept.”
VTG has nothing on IFR either other than word salads unsupported by any references or any science. 0.03% is easy to explain in a place like Africa where the average age is 20 or so.
VTG is part of the despicable public attack on Ioannidis. These attacks are personal, nasty, and political. Partisan hacks should be ashamed of themselves for attacking one of the lions of medical science.
dpy,
you’re hilarious, honestly.
Where to start.
IFR mainstream estimates:
we estimate the overall IFR in a typical low-income country, with a population structure skewed towards younger individuals, to be 0.23% (0.14-0.42 95% prediction interval range). In contrast, in a typical high income country, with a greater concentration of elderly individuals, we estimate the overall IFR to be 1.15% (0.78-1.79 95% prediction interval range).
https://www.imperial.ac.uk/mrc-global-infectious-disease-analysis/covid-19/report-34-ifr/
Herd immunity: even when presented with a reference you claim it says the opposite of what it actually does.
“Despicable” attacks on Ionaddis is laughable. The guy has published endless dubious studies throughout the pandemic and has been proved wrong again, and again and again. He’s zero credibility left in the field but is the darling of the right wing media. That you find pointing this out “despicable” is a measure of how indefensible your position is.
Some academic critique here. It’s not comfortable reading. For instance on that infamous Santa Clara study:
The sampling method, the statistical tests, the sensitivity/specificity of the tests, are all reported to be seriously flawed
Ouch.
https://arxiv.org/pdf/2012.12400.pdf
Joshua,
indeed, dpy’s definiton of herd immunity means that anywhere in the world where cases are dropping has reached herd immunity.
Summarizing: As proven by the citations I was right about the term herd immunity. VTG was wrong. And it is such an obvious error on his part that a minute of research would show it is an error. That indicates something rather dark.
There is by now a vast literature on IFR for covid-19. Values are all over the place and that’s the expected result for this disease. Insisting on a single number is literal minded and wrong too.
The controversy you cite is over the Santa Clara paper. By now, its largely irrelevant to the science as there are a huge number of seroprevalence studies, many of them in the same ball park as the Santa Clara paper. Los Angeles and Miami Dade to name two. Ioannidis has at least one paper doing a meta-analysis of 50 of these studies.
Your vague pseudoscience attack on Ioannidis’ work on IFR is also wrong especially as you obviously haven’t read any of the papers. Again, it is pretty dark stuff.
VTG wrong on all counts. That happens when the scientifically uninformed activist tries to comment on science.
For the medically illiterate who might be frightened by media headlines:
Covid-19 is not very dangerous for healthy individuals under the age of 70. It is dangerous for those over 75 or those who are seriously ill. And yes for the uninformed, metabolic syndrome is a serious illness. It dramatically increases the risk of a heart attack for example. I did some calculations using the MESA study risk calculator. For an 80 year old man, entering typical values for metabolic syndrome resulted in a 3 fold increase in heart attack risk over entering normal values. If you smoke in addition, your chances of a heart attack are very high. Any infection could easily push you over the edge.
After along research effort into the medical literature, I believe that the most important public health priority is to get people to exercise, lose weight, and stop smoking.
VTG, I quoted the entire abstract of ATTP’s reference in a comment above. It says my definition is widely used. My reference shows an instance of scientists using it the way I did.
That you can’t admit you are wrong on such an obvious point shows something dark.
Let’s all remember what Ioannidis says about denialists and science:
“The anti-vaccine movement and climate emergency deniers are already drawing ammunition from the reversals of opinion and policy during the pandemic.”
https://www.bmj.com/content/371/bmj.m4048.full
“Many fields lack the high reproducibility standards that are already used in fields such as air pollution and climate change.
[…]
It is a scandal that the response of governments to climate change and pollution has not been more decisive.”
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5933781/
Ioannidis saying that anthropogenic climate change on same level of certainty as smoking killing people:
17:17 to 18:22 of:
http://rationallyspeakingpodcast.org/show/rs-174-john-ioannidis-on-what-happened-to-evidence-based-med.html
Yet right-wing contrarians act like Ioannidis is on their side. Priceless.
Dpy,
The abstract doesn’t support you, and neither does the paper.
The darkness is all yours.
Quoting the abstract:
“Others use it with reference to a particular threshold proportion of immune individuals that should lead to a decline in incidence of infection.”
That’s exactly how I used the term. You really are entering dishonest denial territory here.
For those whose reading comprehension is very low, I quote again from the abstract.
“Others use it with reference to a particular threshold proportion of immune individuals that should lead to a decline in incidence of infection.”
That’s exactly how I used the term. You are obviously wrong VTG.
David –
It’s certainly possible that anyone could use the term “herd immunity” to mean “the degree of prevalence of infection in a given community at a given time.”
In the current context, that seems to me like a rather meaningless definition. And defining some arbitrary “herd immunity threshold” as some point at which the rate of infections drops, irrespective of context such as interventions in place, or whether it might rise again dramatically without significant changes in the mitigating circumstances as happened in Sweden, is similarly meaningless.
You might as well just say “the level of the population that is infected.”
But it’s clear what Nic meant when he said that Srockholm had hit a “herd immunity threshold”
8 months ago – as we have seen in how Rand Paul referenced Nic’s conclusions about his toy model.
And Nic was wrong. That doesn’t diminish his skills as a statistician. But it also doesn’t diminish the errors he made in applying his skills to a particular context, errors that may not been because of technical flaws, but flaws in his scientific process of application.
The same would Ioannidis’ errors, such as when he went on a national publicity campaign and wrongly suggested we’d top out at perhaps 40,000 deaths in the US, or confidently said that COVID is essentially like the seasonal flu, or when in the Santa Clara study they employed statistical methodology that has been severely criticized by expert statisticians, or when they extrapolated from unrepresentative concenience sampling, or when they promised potential study participants “immunity passports” if they were tested for antibodies when there was a good chance that they’d return a false positive, etc.
And neither does your appeal to Nic’s and Ioannidis’s authority take away your laughable over-evaluation of your own competence for assessing the pandemic, such as that led you to argue that it was unrealistic to think that vaccines would meaningfully mitigate the reach of COVID, or that led you to declare that the pandemic was over in the US in the summer, or to declare that the summer surge wouldn’t result in a spike in deaths, only a spike in infections among young people or a merely spike in positive test results that is a function of a spoke in tests.
But I do appreciate your sense of irony, expressed by your posting here at Climate Etc, with your constant appeals to authority, and your constant assertions that asking for people to allow greater respect for uncertainties is tantamount to vocusous personal attacks that have no merit.
So thanks. Each and every time you post a comment you’re helping to strengthen arguments, and I wouldn’t want you to think I don’t appreciate it.
Josh, I don’t know why you post such long comments that are largely meaningless.
Using a commonly used (by scientists) definition of herd immunity, there is no evidence Sweden did not reach herd immunity for summer time. You provided no evidence of any kind and just asserted something. Coming from a person who has demonstrated a deep ignorance of science and mathematics, I doubt anyone will bother to read what you just wasted a significant amount of life energy typing in. Nic is vastly more credible.
Your criticism of Ioannidis is likewise simply a recitation of what others said and your own completely unsupported assertions. You have no way to verify who is right. Ioannidis has published hundreds of excellent contributions and has the most cited medical paper of the last 20 years. He’s a lion of the field. You are an anonymous nonscientist internet persona who has absolutely no qualifications or expertise to say anything about science.
David –
> Using a commonly used (by scientists) definition of herd immunity,
People use different definitions if terms and words in different contexts.
Yes, “herd immunity” could theoretically mean effectively “the % of people in a given community who are immune.”
But that meaning is useless in this context.
And more to the point it is clearly not what Nic meant when he was talking about “herd immunity” in Sweden. Also, by “herd immunity threshold” he didn’t just mean the point at which, in some generic sense, there was a drop in the infection rate.
That is made clear by his comments after he said they’d reached a “herd immunity threshold” and then he looked at a flat rate of infections for period and said it confirmed his earlier asssertion of having crossed that point.
IOW, it’s clear his asu
Repeating a falsehood doesn’t make it true. You still have presented nothing supporting your nonsense assertions about Nic being wrong about herd immunity in Sweden. There was evidence they reached herd immunity, I.e., R<1.
OK, dpy, so now you’re clear.
You’re claiming that anywhere R<1, he'd immunity has been reached.
So, in the UK, we reached herd immunity in May.
By August, we'd lost herd immunity.
In November, we gained it again.
December, lost it again.
January, we've got herd immunity again.
Glad that's clear.
It’s certainly the definition used by Kwok in the Nature piece I quoted from.
“Though an important paper by Fox et al in 1971 [1] argued that emphasis on simple thresholds was not appropriate for public health, because of the importance of population heterogeneity, assumptions of homogeneous mixing and simple thresholds have persisted. ”
Like many very simple models of complex systems, epidemiological models assuming a homogeneous dynamic are not very useful in explaining an ill-posed system like the course of any epidemic. Insistence on simple minded verbal formulations is not helpful for understanding and is not science.
So yes, epidemiology is a field fraught with huge uncertainties. Sniping at respected scientists by anonymous nonscientists who don’t even know the basics is not helpful to anyone. You know who you are.
Insisting on using R0 to define HIT seems overly simplistic and not very useful. I don’t even think R0 is well defined or useful for modeling an epidemic. R obviously changes over time. It’s obviously seasonal. Viruses typically mutate and new strains will be more or less contagious. I’ve seen some evidence that in Britain there is a new strain that is much more contagious than the older strains. It seems reasonable to say that a new strain would change R0. In fact I’m not sure that R0 is really a very meaningful concept. Is it the R at the start of the epidemic or at some vaguely defined point in the epidemic. If its R at the beginning, its not a useful concept because then the HIT would depend strongly on what season the epidemic started.
Its not an insult to truthfully point out that someone has no qualifications to be taken seriously on scientific questions. It’s childish to comment here without even doing some research on the topic at hand.
David –
Here is what Nic wrote:
> I also projected, based on their declining trend, that total COVID-19 deaths would likely only be about 6,400. Subsequent developments support those conclusions. Swedish COVID-19 deaths have continued to decline, notwithstanding a return to more travel and less social distancing, and are now down to 10 to 15 a day.
Based on HIS DEFINITION of “herd immunity,” Nic protected at total of 6,400 death. Today their reporting surpassed 10,000 dead and it will clearly go higher. I’ll guess that his projections might be off by as much as 200% by the time they get to herd immunity though used if vaccines (that you said wouldn’t matter) by the time all is said and done.
Once again, his projections were based in HIS DEFINITION of “herd immunity.” this there is no question that his modeling if “herd immunity” was wrong as was his assertion that Sweden had crossed a “herd immunity threshold” as HE CONCEIVED THE ISSUES…
He also said it was likely that NO COUNTRY would surpass a population fatality rate of 0.085%. There are some 30 countries that have passed that mark – some by more than double. Obviously, his erroneous thinking in that regard was to some large degree based on his misunderstanding of the mechanisms of “herd immunity.”
There no crime in his having been wrong, but your insistence that he wasn’t wrong and that I’m incapable of pointing out his error because I’m not a scientist is just silly.
You should stop with that.
Josh, I’m simply not going to read your comments as they are very long, totally redundant and completely free of any evidence or documentation. We’ve been over your distortions before many times. That you need to gain attention by such personal attacks is a sad commentary. Don’t you have a day job or relatives and friends to interact with?
No one is right 100% of the time. That’s especially true in your case. Why you can’t just agree to disagree is very childish
Josh you are such an ankle biter. Nic’s projection of deaths was wrong. Winter came, R went up and there were new strains. His statement about herd immunity was in my view correct at the time it was made. Epidemiology is a field with vast uncertainties so most projections will be wrong.
Do you have anything interesting to say that might help people learn? So far, your ratio of words to actually relevant science is almost infinite. Instead you spend time obsessing over minor points in an otherwise excellent body of work by some very competent people.
If this is the only way you can get people to pay attention to you, you are pathetic. You must spend virtually all your time on the internet picking at people vastly more credible than you are. On your tombstone will be written: “Josh: who burned his chest when he fell asleep with his laptop open and obfuscated everything he ever participated in”
Yes Josh, I’m glad to see you admit you are childish.
What I have said here is true about herd immunity. I provided supporting evidence. What you have said is not helpful to anyone trying to understand anything. It’s simply you trying to tear down and call into question the credibility of people who know vastly more than you do.
If you have any friends ask one of them to read some of your comments. You will be amazed that most will agree with my impression.
DPY,
No, it doesn’t really. It goes on to show that the herd immunity threshold is typically defined in terms of the basic reproduction number (R0) not the effective reproduction number.
So then ATTP you disagree with the abstract of your reference and Kwok as quoted in the Nature piece.
“Others use it with reference to a particular threshold proportion of immune individuals that should lead to a decline in incidence of infection” is not ambiguous.
I posted below some reasons why R0 is pretty much useless. Just as an example because of seasonality R0 will be a strong function of the season the outbreak starts. R0 is not like the universal gas constant.
…or some of those states that had a “very high number of cases” actually _didn’t_.
A grossly high false positive rate would seem to show a high level of infections in the early stages (without actual cases), but still allow for a high positive rate in later waves – because the population didn’t actually reach the herd immunity threshold.
Likewise, if a population had a very high “retest” rate, with a significant number of people being counted as positive for each test (once a day in some places, according to some reports), you’d be well below the HIT even if you had a giant pile of true positive tests.
“The FSE values assume that the same R0 and λ value applied throughout the epidemic.”
Which is a extremely supremely BAD assumption! /:
No doubt in practice, but this is for the case of an uncontrolled epidemic.
It wouldn’t make much difference if R0 and λ were different in the relatively early stages of the epidemic from thereafter, and for an uncontrolled epidemic it isn’t obvious that it is a bad assumption after that point.
“uncontrolled epidemic”
Unfortunately, and in practice, that would never happen (in a modern society with near instant global communications and existing codependent technologies (our health care systems and R&D thereof)). And in all countries concerned, that did not happen.
I’ve used the Groundhog Day analogy three times now. Every day of the pandemic all previous days are forgotten wrt the current day, Noone notice that a pandemic has been going on but do notice missing people and they conclude that those missing people are somewhere else on vacation.
Anyone noticing a pandemic would employ some methods of protection in an attempt to minimize (or reduce) further spread, infections and deaths. I am pretty sure that is what has happened to date.
niclewis: It wouldn’t make much difference if R0 and λ were different in the relatively early stages of the epidemic from thereafter, and for an uncontrolled epidemic it isn’t obvious that it is a bad assumption after that point.
With all due respect to your efforts in these essays, those are unsupportable statements. A model and its assumptions have to be shown to be accurate before any consequences are believed or depended on. Given that models are simplifications, the user of the model has the burden of proof that the simplifications do not cause important errors. These points have been made about GCMs and other models.
My simple theory is that, since there were reports of antibodies being shorter-lived as well as mutation, makes it much like seasonal Influenza. I think herd immunity is achieved each season, whether by immunization or affliction.
Antibodies have been shown to last 9+ months in some cases, and memory cells provide resistance even when they fade. Reinfection thus far remains quite rare and serious re-infection even less so[1][2].
Immune resistance to SARS-CoV1 was shown to last up to a decade.
[1] https://www.nejm.org/doi/full/10.1056/NEJMoa2034545
[2] https://bnonews.com/index.php/2020/08/covid-19-reinfection-tracker/
Nic –
Over a period of months, you continually downplayed accumulating and clear evidence that you were wrong in your confident predictions about populations crossing a HIT some 8 months ago.
You tried to say that rising infections were a “blip.” You tried to shift to an excuse that deaths weren’t rising in parallel with infections (by wrongly not allowing sufficient time for an established pattern of lags to equalize).
You basically dismissed the possibility that the drop in infection rates in Sweden were due to seasonality when you argued that infections had dropped there contemporaneous with effectively normal behaviors. If a sufficient % of the population has been infected and recovered or died under normal behaviors, then their “herd immunity” will not surge again in the magnitude we have seen (unless the virus mutates or immunity from infection wanes). On that basis you strongly argued that other countries should follow Sweden’s COVID policies.
You even went on to basically make up a theory as to why populations had reached a “herd immunity threshold” despite relatively low %’s of seropositivity from testing (because of T-cell reactivity that you inferred caused “T-cell immunity”)
I’ll remind you, once again, that along with Willis you looked at some lines on a computer screen and tried to reverse engineer about the obviously complex realities in how disease spreads in real context. Because of your confidence and lack of respect for uncertainties in that regard, you agreed with Willis that COVID fatality would likely top off at 0.085% of the population. There are near 30 countries that have passed that population fatality percentage and some countries are close to double that figure.
It seems rather clear that all along the problem was a lack of sufficient allowance on your part in your calculations for the effects of uncertainties.
Now you want to focus on seasonality. If guess because you think that will give you some way of having been less wrong. But I would suggest that you’re making the same mistake again. You aren’t accounting for enough of the uncertainties related to seasonality, either.
Perhaps having some faith in expert epidemiologists who many decades of experience isn’t always misplayed?
> However, increased population immunity resulting from some combination of further spread of infections and vaccination programmes, the combination varying from one country and region to another, should bring COVID-19 epidemics under control within the next few months.
And here you ignore another uncertainty: The possibility that with a greater number of infections the greater the possibility of mutations that are vaccine resistant.
With NPIs, we increase the chance of lowering rate of infections so as to allow for enough time to reach herd immunity through vaccinations as opposed to infections (or to least shift the % of total immunity that would be attributable to vaccinations as opposed to infections).
Perhaps a problem here is that it seems each time you underestimate the uncertainties related to herd immunity, it’s in one direction only. That could just be coincidence, but at some point it might be worthwhile for you to ask yourself the question as to whether your political biases might be influencing your analytical processes.
I don’t see that Nic has argued against NPI’s? Though it’s becoming better understood that more vs. less restrict NPI’s make little to no difference in spread[1].
As a passive observer, your case is lessened when you make ad hominem attacks. I’d suggest eliminating that approach from your strategy if you’d like to more effectively persuade.
[1] https://onlinelibrary.wiley.com/doi/10.1111/eci.13484
Variant –
> As a passive observer, your case is lessened when you make ad hominem attacks
Fair enough. Point taken.
If you look back over Nic’s series of comments you will see a clear advocacy against NPIs as a way to reduce fatality from COVID.
I’ll look at your link, thanks. But I do doubt that one link will prove conclusive as you suggest. From what I’ve seen there’s a lot of uncertainty regarding the efficacy (or lack thereof) of NPIs that can’t be resolved until we have much better data.
Given the uncertainties, before immunity though vaccination that was at least understandable even it proved to be wrong.
Oops. Last paragraph should be…
Given the uncertainties, before immunity though vaccination was a realistic possibility the fast herd immunity as opposed to NPIs was at least an understandable option even if it proved to be wrong in the end. Perhaps it was true that in the end, the same number would get infected either way despite a slowing effect from NPIs. Pending your link proving conclusive, I’d say that it’s just common sense that NPIs will lower the rate of infections. If you want to argue that they differentially cause more deaths from knock-on effects, I’d love to read it. I’ve yet to se anyone actually take that on as opposed to just conflating correlation with causation and making unsupported counterfactual assumptions about what would be if things were different.
I’ll just note here that Josh’s long and boring litany is not supported by any citations or evidence. It’s just his impressions which in the past have proven to be often wrong.
Nic’s posts are generally informative and interesting.
I see no other viable explanation for the summer near extinction of the epidemic in many areas except herd immunity. Many epidemiologists were fearful that a second winter wave would happen and it did. Variant’s paper is interesting. My comparison of US states also suggests that very strong restrictions have little additional benefit over voluntary actions. The very strong growth in the last week in California is not explainable otherwise.
Josh, You are just so wrong and biased. The data showed that in fact it was under control this summer in many if not most places.
You have a pattern of obsession with scientists and trying to tell them how to think and how they are biased. That’s called mind reading and its unethical for any professional. Why would anyone with scientific credentials even take the time to read your truly vacuous statements?
“And here you ignore another uncertainty: The possibility that with a greater number of infections the greater the possibility of mutations that are vaccine resistant.”
The mutations that are worrisome appear to have arisen not from an increased number of infections, but rather from well meaning (but it seems highly dangerous) health system actions that inadvertently acted rather like gain of function-like experiments carried out in non-biosafe hospitals, in the UK, Italy and quite likely South Africa.
Thank you for your work, Nick. It’s always a pleasure to read it.
If possible, could you please post a link to an article that deal with:
“The mutations that are worrisome appear to have arisen not from an increased number of infections, but rather from well meaning (but it seems highly dangerous) health system actions that inadvertently acted rather like gain of function-like experiments carried out in non-biosafe hospitals, in the UK, Italy and quite likely South Africa.”
> This variant is a wake up call that we should try to really decrease transmission of Sars-Cov-2,” he said. “It is clear that if you leave it circulating, the virus has the ability to outsmart us and become better at transmission and evasion of the antibody response.”
https://www.telegraph.co.uk/global-health/science-and-disease/third-concerning-coronavirus-variant-should-wake-call-world/
Pingback: COVID-19: why did a second wave occur even in regions hit hard by the first wave? – Climate- Science.press
In terms of the second wave in some areas, I wonder what role the additional six or eight months of aging and deteriorating health has on the susceptibility of already vulnerable people. It is obviously not simple to evaluate given the plethora of restrictions and relaxations, seasonality in one form or another, etc. In Saskatchewan, the death rate was linear but very low for months, but after rising in November is now linear again, but steeper. Care homes and northern communities seem to be bearing the brunt (although the toll is not high by any means). This trend is not all that different from Sweden over the past couple of months it seems. We just never had a first wave.
The most successful predictions regarding this disease have been those that assumed it would behave like every other seasonal respiratory infection.
Except ALL other respiratory infections have dropped to near zero. Year over year comparisons will tell us a lot about public health.
You two amuse me. I was quite confident that there would be a resurgence in fall and winter. No one could predict with any confidence the size of that resurgence. There is just as much value in understanding what can’t be predicted with any accuracy as what can be.
And a reduction of other seasonal infections is normal during a pandemic.
There’s plenty of evidence that shows the picture is complicated:
First identified in Wuhan, China, in December 2019, a novel coronavirus (SARS-CoV-2) has affected over 16,800,000 people worldwide as of July 29, 2020 and was declared a pandemic by the World Health Organization on March 11, 2020. Influenza studies have shown that influenza viruses survive longer on surfaces or in droplets in cold and dry air, thus increasing the likelihood of subsequent transmission. A similar hypothesis has been postulated for the transmission of COVID-19, the disease caused by SARS-CoV-2. It is important to propose methodologies to understand the effects of environmental factors on this ongoing outbreak to support decision-making pertaining to disease control. Here, we examine the spatial variability of the basic reproductive numbers of COVID-19 across provinces and cities in China and show that environmental variables alone cannot explain this variability. Our findings suggest that changes in weather (i.e., increase of temperature and humidity as spring and summer months arrive in the Northern Hemisphere) will not necessarily lead to declines in case counts without the implementation of drastic public health interventions.
https://www.nature.com/articles/s41598-020-74089-7
–snip–
However, in comparison to other respiratory viruses such as the flu, COVID-19 has a higher rate of transmission (R0), at least partly due to circulation in a largely immunologically naïve population. This means that unlike the flu and other respiratory viruses, the factors governing seasonality of viruses cannot yet halt the spread of COVID-19 in the summer months. But, once herd immunity is attained through natural infections and vaccinations, the R0 should drop substantially, making the virus more susceptible to seasonal factors.
https://healthcare-in-europe.com/en/news/covid-19-will-probably-become-seasonal-but.html
The infection will wane during the coming summer just as it did last summer. That is my prediction.
As of December 12th in Japan:
-snip-
During the first ten weeks of flu season this autumn, Japan saw just 148 cases of common influenza, or less than 1% of the five-year average for the same period (17,000)
-snip-
https://amp.economist.com/asia/2020/12/12/the-japanese-authorities-understood-covid-19-better-than-most?__twitter_impression=true
doubgadgero comment – “The most successful predictions regarding this disease have been those that assumed it would behave like every other seasonal respiratory infection.”
To a large extent, covid has behaved similarly to prior seasonal respiratory infections. The timing of the surges have followed somewhat closely with the Hopes Simpson analysis (from 1990’s) with the dividing line with covid being approximately the 33 latitude instead of 30 latitude. The other notable exceptions being Canada, Norway and finland, all of which have much harder lockdowns and substantially reduced cross border travel.
https://www.bmj.com/content/370/bmj.m3563
I don’t know Nic if you have seen this. It seems to indicate that there is cross over T cell reactivity fo covid19 in many places. So far, its not really quantified well.
Thanks. Yes, I saw that article at the time. I discussed it in an earlier article of mine: https://judithcurry.com/2020/09/22/herd-immunity-to-covid-19-and-pre-existing-immune-responses/
Did Sweden, which apparently did not use draconian government controls, also suffer a severe 2nd wave?
Am not a statistical or medical expert, but the general thought: better to take the damage early or drag it out? Do you ultimately end up at the similar medical and economic endpoints? These are more pragmatic queries…. and yes I am an engineer.
I have not formed a firm conclusion, one way or the other. My hunch is that the law-of-unintended consequences rears it’s ugly head when governments think they can control outcomes that involve a vast armadas of variables and unknowns. Aim for the middle and hope for the best might be a better approach.
Swedish data takes up to a month to be finalized. But yes, it appears that they are having a second wave almost as strong as the first.
“almost as strong”
Swedish deaths are above 1st wave and still rising.
when you look at morality that s not clear..
https://www.euromomo.eu/graphs-and-maps/
The plain fact is that Sweden, which didn’t have the huge infringements on rights that most of the world had, fared just about as well as the rest of the western world. They did this without the huge harms to the social fabric and businesses that other western countries inflicted on their populations.
All in all, was it really worth it? NY State, land of draconian measures, has suffered about the worst in the US; Florida, without such measures and with many more older people, is doing much better.
If you compare Sweden to the countries which are the most similar, and which have structural elements in terms of health outcomes predictors that would suggest similar results, it has done very poorly.
It’s a complicated process. Isolating the predictive variables is complicated. But generally if you’re trying to assess the efficacy of interventions you’d want to compare situations where you’ve controlled for a any confounds as possible. Therefore, comparing to the other Nordic countries, rather than “the rest of the western world, ” is probably more instructive. Just as you wouldn’t compare to South Korea or Taiwan to assess outcomes in Sweden.
Why do you think Sweden has fared so poorly compared to its Nordic neighbors?
And btwz do you think that citizens of its Nordic neighbirs generally feel that they’ve had their rights trampled on relative to Sweden?
What evidence would you use to assess
disruptions to social fabric and businesses in Sweden relative to their Nordic neighbors?
Florida is not doing “much better” than NY in terms of cases per capita. In fact, it’s doing worse. Considering the improvements in our understandjmg of how to treat people ill with covid and a number of other factors, and the timing of the surges in NY as opposed to Florida, comparing outcomes in those two states gets complicated. Consider that Florida, if you look at population density combined with population distribution (Iow, compare pop density in NY and Miami Dade), as well as the impact of seasonality, has important balancing factors that in contrast to median age, cthat work to the advantage of Florida with respect to expected outcomes.
Related to herd immunity, IOW rate of infections (as opposed to deaths), their second wave is considerably larger than their first. Although ascertainment may also be higher. And their testing rate may be higher. But their fatality rate may well be lower (maybe they’ve gotten at least somewhat better at keeping COVID out of senior congregate living and aren’t just leaving infected older people to die as much as they did during the first wave).
IOW, When someone who has no idea what he’s talking about says “it appears that they are having a second wave almost as strong as the first.” you should prolly assume he doesn’t know what he’s talking about.
Nic – take a look at our cases here in the low deserts of Arizona. Most of our population is at low altitudes and fairly low latitude. This means that right now is an optimal time to be outside rather than inside. The summer is the time we stay indoors.
And yet, we had three peaks.
The first was in the spring, was met with relatively strong mitigations, and was not very large.
The second was in mid-summer (early July), when our high temperatures were way over 100F and our lows over 85F. And yet, it was a much higher peak, to the extent the it made world headlines.
It went away – perhaps due to stronger mitigations put in place at state and also local levels than existed during its rise.
Now we are in our winter peak, and we are at world headline levels again. Our 7 day average is 140 cases per day per 100,000. We still have the same mitigations that went into place just before the summer peak. A few of our schools are open.
A couple of factors might affect this, but it’s hard to explain our current very high level (and current Rt of 1.5): our winter visitors arrived in the autumn and are still with us; and, schools have been open, more or less.
But, the winter visitors tend to be elderly, and likely to take more precautions than younger people, due to their higher risks. And, the schools have been closing back down.
Our hospitals are stressed, and if that 1.5 persists, will be grossly overloaded within two weeks.
How do you explain the shape of this?
http://91-divoc.com/pages/covid-visualization/?chart=states-normalized&highlight=Arizona&show=highlight-only&y=both&scale=linear&data=cases-daily-7&data-source=jhu&xaxis=right&extra=&extraData=deaths-daily-7&extraDataScale=separately#states-normalized
Meso,
Arizona has seasonal influenza outbreaks just like other states have them. Whatever factors apply to seasonal outbreaks apply to Arizona also.
I’m trying to explain the unusual seasonal influences here. Why should we have two large peaks at opposite peaks of our seasons. In other words, one when the temperatures are too hot to do much outdoors, and the humidities are very low, and the other when the temperature are very nice outdoors, and the humidities are low.
I don’t see how the weather/climate is the big factor.
Meso – the two peaks in AZ (not three) are following the Hopes Simpson curve with the exception that the divide appears to be the 33rd/34th parallel instead of the 30th parallel. The same is true for rest of the southern states. California being an exception. Granted it is not a complete explanation, but it does provide some insight. The exceptions to the hope simpson curve are those regions/countries that had an extremely hard lockdowns.
“Meso – the two peaks in AZ (not three) are following the Hopes Simpson curve with the exception that the divide appears to be the 33rd/34th parallel instead of the 30th parallel. The same is true for rest of the southern states. California being an exception. Granted it is not a complete explanation, but it does provide some insight. The exceptions to the hope simpson curve are those regions/countries that had an extremely hard lockdowns.”
I apologize, but I have no idea what explanation you are offering. Could you explicate it a bit?
Thanks Joe. Interesting:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2134066/
“Meso, Arizona has seasonal influenza outbreaks just like other states have them. Whatever factors apply to seasonal outbreaks apply to Arizona also.”
Yes, it does. But we don’t get influenza in the summer, just the winter. But COVID19’s main spikes here were in early July and now (early January). It is possible that our winter influenza peak is imported from other states, especially with our snowbird visitors (part year residents who come from cold climates).
So, something is affecting COVID19 differently, which is one reason I am puzzled.
dougbadgero – figure 2 on page 39 of your link
Figure 2 matches very closely with the with what is happening in the US and Europe. The top graph in Fig 2 is matches the northern states, while the second graph matches the southern states. The only large difference is the dividing line is 30 parallel in Hopes simpson while covid dividing line seems to be approximately the 33rd-34 parallel. ( I cant explain the difference in the 3degrees of latitude)
I cant explain the reason, nor will I attempt any explanation other than nature having a bigger hand in the curve than man’s ability to control it.
Numerous individuals have been critical of Nic for being “wrong” about herd immunity. Possibly, Sweden appeared to have reached herd immunity based on the reduction in the IFR in the late spring and during the summer. Sweden’s curve likewise matches very closely with the Hope Simpson curve for the northern latitudes along with the rest of europe,with the exception of norway and finland. Note that I am neither praising or condemning Nic. (he is vastly more knowledgable than I am with both stats and epidemiology, so it is certainly not my place to condemn his work)
Meso – The hope-simpson curve shows for regions south of the 30th parallel a summer peak approximately 1/2 of the winter peak, while regions north of the 30th parallel to have a single peak in the winter months with no summer surge.
All the southern states such as AZ, TX, GA, FL, AL, MS, CA. The difference in the southern states surge vs the historical trends detailed by Hope-simpson is the dividing line for the summer surge is the 30th parallel vs the dividing line to be approx the 32/33/34th parallel with covid. I cant explain the reason, only pointing out the similarities.
https://youtu.be/FfaU-9tpQlM
I’m thinking… it’s the result of the onset of winter and a mobile population…
What people are still flying here, although flying is down quite a quite a bit, are from all over the world…
For example, I talked to a couple who called off a family gathering prior to Thanksgiving because a couple of members had come down with covid-19 and they blamed that on a local restaurant in their area but as it turned out the wife of the couple also worked for a company that had been visited by persons from Turkey with whom they spent the day and night with and that country was on a government warning list and they had probably flown to the US through Paris or Frankfurt and no telling how many days layover they might have taken to enjoy the trip so… If you’re older and no longer working you probably are safer as far as being insulated from potential carriers unless you’re in a rest home in which case you probably have no control over people coming in and bending over you and coughing in your face…
In 2020, we had 1001 cases on Vancouver Island (Population 870,000) and 12 deaths. We actually beat back introduced viruse here 3 or 4 times during the summer, and we had one streak of 100 days with no infections. That was pretty good! Vancouver (population 3 million’ish) didn’t do so good even though we are under a little less restrictions than they are. Lots of people retire here so we have been masking up a lot, so the old people don’t bite the dust. Also, its a bit harder to travel here (ferries or planes) and people here tend to key (or generally vandalize) the cars of people who pop over for a visit. BUT now the UK variant is already here so our winning streak is over. The R value will obviously increase now the more transmissible version is here. So it spreads quicker AND the herd immunity value goes up from maybe 70% to say 90%. BUT when the South African, Nigerian, and Brazilian variants mix, then “survival of the fittest” means an even more transmissible, or longer lasting or more gradual (or more rapid) Super infection will develop from the combination. Maybe a variant will develop that leaves minimal immune response. Viruses don’t have sex per say, but they do the recombination thing if someone gets infected by 2 strains at once, so a new strain or combination of strains may win the race. This is inevitable. I don’t get why we western humans make it so easy for the different strains to mix. We couldn’t have done it better if we tried. Planes come in to Victoria (where I live) every week with Covid. Container Trucks come up from the usa every day too. I have no idea why they don’t change “tractors” at the border. That would really reduce mixing. Its Crazy. Anyway, the “math” Nick is doing is a curiosity. It means nothing, mainly because western countries don’t have common sense leadership who think to reduce “mixing” of the viral races by banning international travel, sensible truck rules, etc etc. It really isn’t that hard. A lockdown needs to be real, not half hearted. We never had a real one. It may be too late now. .
https://twitter.com/DrTomFrieden/status/1347711472282046473?s=20
https://twitter.com/youyanggu/status/1347266544946929665?s=19
https://twitter.com/youyanggu/status/1347266560021340160?s=19
https://twitter.com/youyanggu/status/1347266561082421248?s=19
A lot more discussion there – including in the follow-on comments.
https://youtu.be/yMBNCpAW7cw
Wow. In just the first 2:30, “same IFR as the seasonal flu,” and “There’s already herd immunity.”
Remarkable.
Have you seen people dropping dead like flies on the streets in NSW?
I guess not. => Herd immunity!
It is time for you to read AP biology to understand how immune system works.
Here, a regular commuting day for folks who are not paranoid. Zero covid deaths in more than a month.
https://ibb.co/crwh1CN
SMH
https://twitter.com/NSWHealth/status/1348421479931682816?ref_src=twsrc%5Egoogle%7Ctwcamp%5Eserp%7Ctwgr%5Etweet
Thanks for posting. I can’t help but think it’s too late for anyone to listen to the Big Pharma issue anymore, governments have taken the path of least resistance.
I cant help but think, maybe erroneously, that 3000IU of vitamin D daily, whilst not stopping Covid-19 would serious reduce its adverse effect
Really can’t hurt. Being at risk of vitamin D deficiency I take 1000 IU of vitamin D along with a gram of calcium a day. I will increase that. Thanks.
https://watertechbyrie.files.wordpress.com/2021/01/nihms436898f1.jpg
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3559187/
China deletes Wuhan lab research data in desperate cover up of source of covid-19:
https://youtu.be/M2YBUFpnnbs
Pingback: COVID-19: why did a second wave occur even in regions hit hard by the first wave? – Climate- Science.press
Many? Almost nobody thought this in the scientific community.
Many people in the right wing libertarian political community and their well financed media friends *wanted* to believe it, sure.
Untrue. If you bothered to look at the references that I cited, you would see that they are both by mainstream academic authors. One of them is in Nature Communications, which to my knowledge is not somewhere the ‘right wing libertarian political community’ (whoever that may be) are normally able to publish.
Untrue.
You don’t specify which references you rely on, but the nature Comms article absolutely does not support your “unlikely that a major second wave would occur” assertion in any way whatsoever.
Essentially what you’re claiming here is that many in the scientific community believed that herd immunity was what drove the reduction in infection seen in many regions.
That’s just not true at all.
According to the govt dashboard and research by a third party, the levels of infection are thought to be far higher than originally thought in many parts of the country
https://www.dailymail.co.uk/news/article-9127059/One-FIVE-people-coronavirus-modelling-suggests.html
I don’t know either way, so just putting it out there as maybe relevant
tonyb
Tony,
No idea what relevance that has to previous views on herd immunity and 2nd waves. Those figures aren’t far off what was previously thought btw, I haven’t looked into what’s behind them, and I doubt the Mail will help that…
@verytallguy
Re: “Essentially what you’re claiming here is that many in the scientific community believed that herd immunity was what drove the reduction in infection seen in many regions.
That’s just not true at all.”
Exactly. It was mostly an idea peddled by right-wing politically-motivated non-experts, and people who’s ideas were so obviously wrong that they couldn’t get through peer review or any review by informed experts. A second wave was shocking to those ideologues + non-experts, but not to those of us who knew better than to think Sweden, Stockholm, New York City, Geneva, etc., reached the herd immunity threshold.
Instead, informed people were pointing out that the herd immunity threshold was not very low, one should not allow rely on natural infection (as opposed to vaccination) as means of achieving some fictionally low herd immunity threshold, and that one should pursue the sort of mitigation policies Lewis opposed (instead of allowing people to get infected to achieve herd immunity).
It is simply absurd for Lewis to act as if much of the scientific community was behind him on his obvious errors:
“The arrival of a second wave and the realisation of the challenges ahead has led to renewed interest in a so-called herd immunity approach, which suggests allowing a large uncontrolled outbreak in the low-risk population while protecting the vulnerable. Proponents suggest this would lead to the development of infection-acquired population immunity in the low-risk population, which will eventually protect the vulnerable.
This is a dangerous fallacy unsupported by scientific evidence.”
https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)32153-X/fulltext
“In summary, there are large differences in patterns of per-capita deaths in different countries that are difficult to reconcile with herd immunity arguments but are easily explained by the timing and stringency of interventions. Seroprevalence studies also provide an independent source of information that is highly consistent with mortality data. The herd immunity argument is therefore at odds with both mortality and seroprevalence data, whereas the intervention argument provides a parsimonious explanation for both.”
https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)31357-X/fulltext
“Countries should not rely on herd immunity by natural infection to suppress the epidemic. The disease and death that would accompany natural infection rates to reach herd immunity, typically estimated as 40–60% of the population infected, would be unacceptably high. Uncertainty also remains about the duration of acquired immunity from past infections.”
https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)31927-9/fulltext
“Initially, some local authorities and journalists described this as the herd immunity strategy: Sweden would do its best to protect the most vulnerable, but otherwise aim to see sufficient numbers of citizens become infected with the goal of achieving true infection-based herd immunity. By late March 2020, Sweden abandoned this strategy in favor of active interventions; most universities and high schools were closed to students, travel restrictions were put in place, work from home was encouraged, and bans on groups of more than 50 individuals were enacted. Far from achieving herd immunity, the seroprevalence in Stockholm, Sweden, was reported to be less than 8% in April 2020,7 which is comparable to several other cities (ie, Geneva, Switzerland,8 and Barcelona, Spain9).”
https://jamanetwork.com/journals/jama/fullarticle/2772167
“In light of these findings, any proposed approach to achieve herd immunity through natural infection is not only highly unethical, but also unachievable.”
https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)31482-3/fulltext
“At present, herd immunity is difficult to achieve without accepting the collateral damage of many deaths in the susceptible population and overburdening of health systems.”
https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)31483-5/fulltext
“This quasi-equilibrium is maintained not because of herd immunity but because of changes in behavior. […]
The peaks occur at levels of infection far from that associated with herd immunity. Post-peak, shoulders and plateaus emerge because of the balance between relaxation of awareness-based distancing (which leads to increases in cases and deaths) and an increase in awareness in response to increases in cases and deaths.”
https://www.pnas.org/content/117/51/32764
“This allows us to “bend the curve” and predict temporary equilibrium states, far away from the equilibrium state of herd immunity, but stable under current conditions […]. Yet, these states can quickly become unstable again once the current regulations change.”
https://link.springer.com/article/10.1007/s00466-020-01880-8
“Transmission dynamics reveal the impracticality of COVID-19 herd immunity strategies”
https://www.pnas.org/content/117/41/25897
“Four months into the COVID-19 pandemic, Sweden’s prized herd immunity is nowhere in sight”
https://journals.sagepub.com/doi/10.1177/0141076820945282
“There are two possible approaches to build widespread SARS-CoV-2 immunity: (1) a mass vaccination campaign, which requires the development of an effective and safe vaccine, or (2) natural immunization of global populations with the virus over time. However, the consequences of the latter are serious and far-reaching—a large fraction of the human population would need to become infected with the virus, and millions would succumb to it.”
https://www.sciencedirect.com/science/article/pii/S1074761320301709
“Vaccination is the only acceptable path to herd immunity”
https://www.cell.com/med/fulltext/S2666-6340(20)30032-5
“Consequently, herd immunity may be the most important “short term” strategy to protect this portion of the population [97]. Until a safe vaccine is developed, research on specific new treatments (or an effective combination of existing treatments), together with action plans to contain the spread of the virus, seem to be the only alternatives for protecting at-risk populations [98].”
https://www.mdpi.com/2076-393X/8/2/236/htm
“Public health organizations condemn herd immunity scheme for controlling spread of SARS-CoV-2”
https://apha.org/news-and-media/news-releases/apha-news-releases/2020/public-health-orgs-condemn-sars-covid2-plan
“Attempts to reach ‘herd immunity’ through exposing people to a virus are scientifically problematic and unethical. Letting COVID-19 spread through populations, of any age or health status will lead to unnecessary infections, suffering and death.”
https://www.who.int/news-room/q-a-detail/herd-immunity-lockdowns-and-covid-19
“Early herd immunity against COVID-19: A dangerous misconception”
https://coronavirus.jhu.edu/from-our-experts/early-herd-immunity-against-covid-19-a-dangerous-misconception
“We do not endorse the idea of letting the epidemic a free hand in order to create sufficient herd immunity to end the epidemic; as it would entail an enormous burden on the healthcare system – United Kingdom, at first, considered a different approach – of unrestricted spread of disease without any brakes applied, but public health experts were able to convince the government to accept the more reasonable mitigation approach. We should clarify at the outset that an approach of uncontrolled spread, one that no country is following at this time, would be a terrible mistake.”
https://www.indianpediatrics.net/june2020/505.pdf
https://www.twitter.com/DrZoeHyde/status/1261645790142844933
https://twitter.com/AtomsksSanakan/status/1347618857607950338
Ugh, the politicized attacks on Lewis are getting more odd just as they seem to be getting angrier.
Let’s take one piece of “evidence” from one of the attackers:
“There are two possible approaches to build widespread SARS-CoV-2 immunity: (1) a mass vaccination campaign, which requires the development of an effective and safe vaccine, or (2) natural immunization of global populations with the virus over time. However, the consequences of the latter are serious and far-reaching—a large fraction of the human population would need to become infected with the virus, and millions would succumb to it.”
https://www.sciencedirect.com/science/article/pii/S1074761320301709”
This is supposed to be devastating to the claim that herd immunity can be reached and someone ought to attempt to figure out when. Guess what happens absent a vaccine? The only option is natural immunization. Guess how many times Nic Lewis argues against a vaccine? None.
Oh, but this is supposed to back up the claim that there is a third way! Our enlightened political leaders can force the virus away by strictly controlling the population. You’ll note they don’t actually say that anymore because it’s obviously not true. The places with the most strict lockdowns and most complete mask adoption are the places with the highest spread. This is true in and out of countries with someone named Trump.
These clear facts somehow have no impact on the willingness of the “pro-science” sect to bow to it daily.
There are a few things we know about Covid and immunity- we cannot stop the spread of a highly infectious virus with mandates (but those can destroy lives, economies and schools), we didn’t reach natural herd immunity, the American president’s work to rush a vaccine was critical and will save hundreds of thousands of lives thereby giving the American media a reason to praise Andrew Cuomo for no reason even more vigorously.
Ironically the one political leader most prescient about Covid spread was Angela Merkel, who told the press at the start of all this that 80% of Germans would get Covid.
Nic quibbled with the number 80 for science reasons and tried to come up with the right one.
VTG and his sidekick, meanwhile, still seem to be insisting that Los Angeles must be covid-free because the government ordered people not to spread it.
Why they think that makes them more accurate than Nic Lewis is beyond me.
Atomsk’s Sanakan: It is simply absurd for Lewis to act as if much of the scientific community was behind him on his obvious errors:
Good posts. However, Nic Lewis only claimed that “many” people disputed the possibility of a second wave, and were wrong; and he was wrong. What think you of the reasons he gave for explaining how he and they came to be so wrong?
Did anybody make accurate quantitative projections?
I have always found Mr Lewis’s papers interesting and that feeling is enhanced when he works his statistical magic through a real time event in which we are all participants. This paper is another thought provoking look at how we have continued to struggle to find a way out of the cul-de-sac of our own making.
My own experience of this novel virus’s ability to confound its chosen host in ways I have never before seen in the UK (or elsewhere) is unsettling. At one extreme my experience has been watching the fear of this virus from people who have little chance of contracting a serious episode of illness but could ostensibly asymptomatically transmit the disease to a more vulnerable person if such an event is possible. At another extreme has been the inadequacy of our public health response especially test, trace and track, where, at every stage, poor quality is and has been demonstrably present throughout while there has been little evidence of strong professional and experienced leadership at any level. Surely that has damaged much of the evidence that has been offered to us via these outsourced services both in blighted test results and inadequately finding and tracking those who have received a blighted test result be it false positive or false negative. Who is in charge of quality control and do they have the necessary experience, expertise and leadership ability?
Even the stance on variants of the virus, which virologists say can be frequent and numerous in all cases of a coronavirus, has been turned into an additional terrifying event which must, according to politicians, be stopped, not because the variant is known to be more or less pernicious but because our hospitals could be overwhelmed (almost exactly a year after they were previously about to be seriously overwhelmed so no change there then).
My take on this whole experience is that we have been badly let down by ruling classes that spend much of their time looking down and exercising thumbs and not enough of their time looking up and actually observing what is going on with people around them. They need to observe the fine detail of how we once accepted that the people who get sick and recover become a part of the crowd who have always in many senses eventually protected the more vulnerable. We can act to save lives and unnecessary deaths but the measures we have taken have simply failed to do that and the reasons are to be found in how we have attempted to manage this epidemic – forever murmuring OMG – instead of getting on with life as we know and accept it and riding the attack out.
Pingback: COVID-19: why did a second wave occur even in regions hit hard by the first wave? – Watts Up With That?
Seriously. Does anybody even have to ask this question?
I’ve been saying there would be second waves and probably even third waves and possibly even fourth waves.
As cases ebb from one wave, constraints on the spread start to loosen sometimes by deliberate government action but, if not that, then by the human tendency to think the disease has passed. As constraints loosen, new previously unexposed groups of people get ill and the next wave begins. Constraints tighten again through government action or fear so the cases begin to ebb again. The cycle repeats until there is herd immunity or enough unexposed people impose sufficient restraints on interaction that the virus no longer exists in the population. Wiping the virus out of the population is extremely difficult in a mobile free society.
Restrictions until enough people are vaccinated to reach herd immunity is the only option that will not result in significant numbers of deaths.
Pingback: COVID-19: why did a second wave occur even in regions hit hard by the first wave? |
> Since neither of the factors on which the HIT depends are fixed, the HIT is not fixed either.
Without any attempt at explication of reasonable parameters, that statement could mean that “HIT” could be almost anything and thus “HIT” seems pretty much meaningless.
I’m wondering if anyone can provide an example in history where “natural herd immunity” has been considered established only later to see the kind of surge in infectiousness on the order of what we’ve seen with COVID after there were claims that “herd immunity” had been established.
Keeping in mind, of course, that “herd immunity” is only really a meaningful term if it is being applied to a condition where a real society is engaging in at least close to normal behaviors.
I’m going to wager that no one can come up with one (absent something like a major new mutation coming into play).
So then the question would be why are people trying to say that the massive growth in infections we’ve seen is some how an outgrowth of “seasonality.” Is COVID considered somehow uniquely subject to seasonal variation? I don’t recall anyone making a scientific argument to support that view.
Or maybe having seasons is some kind of a new thing? I mean I know that 2020 was someone of a unique year in many respects and 2021 is off to an unusual start….
Without any attempt at explication of reasonable parameters, that statement could mean that “HIT” could be almost anything and thus “HIT” seems pretty much meaningless.
However, I have made such an attempt in my article. Why don’t you propose what you think are reasonable parameters if you disagree with my suggestions, rather than always just sniping unpleasantly?
“Keeping in mind, of course, that “herd immunity” is only really a meaningful term if it is being applied to a condition where a real society is engaging in at least close to normal behaviors.”
Which is exactly why I focussed on data from Sweden in previous articles.
The view of Sweden as a libertarian fantasy of freedom is not, and never has been true.
https://ourworldindata.org/grapher/covid-stringency-index?tab=chart&stackMode=absolute&time=2020-01-22..latest&country=SWE~GBR~NOR®ion=World
I am not sure about them being a Libertarian fantasy, but they are continued evidence that lock downs don’t seem to make much difference.
Posted that way upstairs. Along with some interesting Tweets from Tom Frieden…
> Which is exactly why I focussed on data from Sweden in previous articles.
Except Sweden wasn’t really that close to “normal” behavior even by Swedish standards – high schools and colleges closed, large groups not allowed, massive numbers of people working at home, people traveling less in ways that paralleled other similar countries, people not allowed to visit seniors in congregate housing, ill seniors basically just allowed to die rather than transported to hospitals, etc.
And even “normal” in Sweden is far from “normal” in countries which you prescribed should follow Sweden’s lead, based on your mistaken conclusion that they’d reached herd immunity. Probably age stratification is the most important predictor and many other countries are sort of “normal” when compared to Sweden in that aspect, but there are myriad other important predictors which, IIRC, in all you articles on this issue have NEVER even come remotely close to discussing. Even the most obvious ones such as household size or # of multi-generational households.
So it’s a bit hard to see how you took this issue of the generalizability of Sweden’s “normality” very seriously. Looks more likely to me that you simply made some convenient assumptions.
I do admire Nic for responding to the vague, pseudoscientific and largely rhetorical criticism here.
Kwok in the Nature piece I quoted above says that the HIT was highly variable on a country by country basis, ranging from 5% to 78%.
The problem with using R0 is that it is ill defined and not useful. It will depend strongly on the month in which the epidemic started to give just 1 example for any seasonal virus.
‘Normal conditions’ likewise is impossible to define rigorously. Just as trying to define a single weather ‘normal condition’ is impossible or meaningless for telling you what the weather will be tomorrow.
Re: “I’m wondering if anyone can provide an example in history where “natural herd immunity” has been considered established only later to see the kind of surge in infectiousness on the order of what we’ve seen with COVID after there were claims that “herd immunity” had been established.”
It doesn’t happen. As I pointed out months ago in July ( https://medium.com/@silentn2040/the-dangerous-myth-that-sweden-achieved-herd-immunity-fd2579526b8b ), multiple peaks during public health interventions and behavior changes that push one away from the baseline conditions of R0 that define herd immunity, are a clear sign herd immunity wasn’t achieved. It’s telling Lewis just got around to addressing this, yet still did so badly. Almost as if he didn’t understand the basics of the epidemiology, immunology, infectious disease, etc. he was distorting…
“If significant herd immunity developed following initial major water supply contamination, a multipeaked and/or prolonged epidemic would not be expected to occur [page 722].”
https://www.sciencedirect.com/science/article/abs/pii/S0168827894802297
https://mobile.twitter.com/AtomsksSanakan/status/1332738821998456841
A lot of vague non science in this thread. I still think the weight of evidence is pretty clear. There has been nothing substantive here from the usual nonscientists to even attempt to explain the course of the epidemic over the summer in the US. Especially in New York and California, where restrictions if anything have gotten more severe over time, there has been a strong surge this winter and a relative lull over the summer. The story is much the same in virtually every European country. The quite marked summer lull in the epidemic is a rather glaring feature which requires an explanation.
Nic has at least tried to outline the science and the math. And the response is a bunch of ankle biters raising baseless objections.
It is remarkable how challenging the coronavirus is proving to gain understanding of its epidemiology. What’s clear is that we’re not out of this, not by a long shot.
λ will always exceed 1 (its level in a homogeneous population); pre-epidemic λ may be ~4.
Small mathematical point: a parameter with range from 1 to infinity is poorly framed. Like degree of anisotropy for example (maximum over minimum eigenvalue). It would be better for it to range from 0 to 1 so it is well bounded and can be converted into a percentage. This is easy to do – just invert the ratio and precede it by “1 minus”.
The practical range of possible values of λ is not that large.
I don’t think that reparameterization would really help here.
Pingback: COVID-19: why did a second wave occur even in regions hit hard by the first wave? – Climate- Science.press
The vaccine will give us some hard data on HIT. In my state the count of those vaccinated plus those who caught the virus is ~7% and new case counts are still high.
It’s interesting watching all the commenters on this site who told us lockdowns and masks would stop any new wave of the virus attempt to pick on Nick for trying to estimate herd immunity.
I know a bunch of people in California. They are all locked down tighter than we ever were in my city and I think they all wear masks even when alone. Their hospitals are discussing how to ration ventilators right now. While CNN attacks the governor of Florida for no reason. Again.
A whole lot of people were wrong about this virus
“ I know a bunch of people in California. They are all locked down tighter than we ever were in my city and I think they all wear masks even when alone. Their hospitals are discussing how to ration ventilators right now. While CNN attacks the governor of Florida for no reason. Again.”
Something we have all gotten use to. There are no facts anymore. No truths. Just narrative and propaganda. The MSM have no idea how much they are despised.
And people are getting really angry and real damage is being done.
Our blue governor flip-flopped on school reopenings (and looks to be flipping back again). Before the election, anyone who wanted the schools open was a raving Trumpist. After the election, they tell us the “science” says reopening schools is safe and necessary. Which is true of course, in fact it always was, but because they denied it so vocally for so long we literally have petitions flying around my city with left, right and center angrily waving pre or post November “the science” at each other and calling each other evil liars. The only bi-partisan agreement is that “the science” is wrong. Great job everyone!
They better get the vaccine out faster. After the end of January they won’t have anyone to hate as a distraction from all the damage they’re doing.
Can you imagine if CNN had their brainwashing operations going in 1941?
https://pbs.twimg.com/media/EgjbdziXYAYlMUH?format=jpg&name=900×900
HIT from a vaccine will be different than HIT from those who get the infection. Especially, when we are vaccinating the least socially connected people. This makes some sense of course since they are also the most vulnerable.
jeffnsails850,
I live in LA County, which right now seems to be the COVID capitol of the USA. I do not feel “locked down” (shades of Brian White’s comment https://judithcurry.com/2021/01/10/covid-19-why-did-a-second-wave-occur-even-in-regions-hit-hard-by-the-first-wave/#comment-939548 below).
Yes, we have a lot of businesses shut down, but I still go to the supermarket, Walmart, Target, Costco, and so on (the bigger stores). Malls are open, and I can get take-out food from the Food Court (I miss Sbarro stuffed pizza). I even went to a Farmers’ Market this morning in La Canada. Yes, masks are required in stores, but the “quality” of the mask is unimportant. Just as long as you have your nose and mouth covered, and the nose part seems often optional. Checkers at Trader Joes can bag your groceries now as “fomites” have apparently been deprecated as a mode of transmission. More people seem to be allowed in stores as well. Rush hour traffic has been almost normal for awhile as cars are no longer social-distancing.
If I go hiking at the regional park in Montecito Heights or even walking down the street, my mask is under my chin. Some stay at home order! I’m not aware of any police enforcement. Maybe “large gatherings”, but I don’t go to them anyway.
Weather doesn’t seem to matter, even if Joe – the non-epidemiologist’s comment about latitude and seasonality is right. We have had summer-like weather here for awhile and very little cold weather. But cases are “surging” if the PCR test is to be believed. And all those scientific expectations of an autumn / winter surge don’t seem to have influenced hospital planning much.
Very crazy times.
LA Bob – “Weather doesn’t seem to matter, even if Joe – the non-epidemiologist’s comment about latitude and seasonality is right.”
Bob – I am not claiming to be right as much as pointing out the observation that the covid surges/trends seem to be following century old historical trends with prior influenza viruses.
There is an old saying that you cant know where you are going if you dont know where you have been. In the case of Covid, it seems the experts have thrown out all the science about prior influenza pandemics and substituted “the new Covid science” .
The other observation is that the Covid surges seem to be following historical trends (albeit on a larger scale due to covid novelty) irrespective of the level of mitigation instituted by the states governors and the level of compliance with those mitigation steps. For example colorado and Minnesota had very similar covid surge trends in the fall with most of the other northern states with vastly different levels of mitigation compliance.
Thanks for your first hand report from LA. I lived in LA for a few years, so I’m familiar with the excellent climate.
“Weather doesn’t seem to matter, even if Joe – the non-epidemiologist’s comment about latitude and seasonality is right. We have had summer-like weather here for awhile and very little cold weather.”
Yeah, I have a problem too with the seasonal hypothesis. LA barely even has seasons until you get inland some distance from the ocean. Even then, the temperature variation is pretty mild.
I’m in Phoenix. We are at the same latitude as LA, and have the same annual rainfall (yes, technically, LA is in a desert).
And we are also having a surge – almost as bad as that in LA county. You have about 140 new cases per day, and we have about 120.
Interestingly, both of our states had both a summer and a winter surge, but our summer surge was about double LA county’s. (graphs on us-covidtracker dot com).
So, I find that a hypothesis tying the seasons to our epidemic to be weak – at least if we define seasons in terms of weather. As I’ve mentioned here before, we have had a peak at our hottest time, and now we have one at our coldest time. We’ve had no precipitation at all during those peaks.
Manaus has had huge surges in opposite seasons.
Manaus Brazil which is near the equator , the trends for covid are following The hope Simpson surge trends for the regions 0 to 30s latitude – albeit on a larger scale. ( reasonably close with some variation , ie similar trend)
Hi, Joe,
I poked around a bit in order to understand Hope-Simpson, who at least had the temerity to question common views on influenza transmission. In light of the 30th parallel thing, I’m trying to understand Brazil’s Worldometer numbers. Brazil fits almost entirely above the 30th parallel in Southern Hemisphere terms, and they should be in their summer season now, correct? But their daily new cases are trending higher than last winter, and daily deaths are trending about the same.
I suppose one might suggest that Brazil got a later start on COVID-19 than Northern Hemisphere countries, so there’s not been enough time to establish a pattern.
I also found this 2008 review, which seeks to explain Hope-Simpson flu seasonality by vitamin D, which seems to be a hot topic these days.
https://virologyj.biomedcentral.com/articles/10.1186/1743-422X-5-29
LA Bob – my purpose in mentioning hopes simpson is to bring some historical insight into influenza.
Brazil is mostly north of 30S. I have attached a link to his study (page 39, figure 2) which shows 30s -0 to have the its surge in June/July, with a summer surge of Nov/Dec/Jan. with both surges to be rather flat vs the strong surges north of 30N and south of 30s.
The charts for brazil at 91-divoc . com shows a somehat similar trend for brazil . Unfortunately, I couldnt pull a chart for just manasus.
I also agree with your comment regarding Vitamin D,
I ask again: how does Hope Simpson apply to Arizona. Specifically, since the population is in a narrow latitude band (33N), Maricopa County, AZ.
Again, we had a large peak in the summer and are experiencing an even larger peak now in the winter.
Also, do you have a reference to Hope Simpson (I assume you mean Hope-Simpson)? I find a Wikipedia article that doesn’t address the issue:
“His career-long interest in the manner of transmission of the influenza virus was first stirred by the great epidemic of 1932–33, the year in which he entered general practice. It culminated in a book, published in 1992, which questioned the theory of person-to-person transmission being enough to explain the simultaneous appearance of influenza in places far apart. His initial hypothesis proposed that the cause of influenza epidemics during winter may be connected to a seasonal influence.[2] His later research suggested that the correlation may be due in part to a lack of vitamin D during the wintertime, after documenting that influenza A epidemics in temperate latitudes peak in the month following the winter solstice.[3][4] His findings were based not only on observation in his practice, but also on extensive historical research into past epidemics”
I note: “questioned the theory of person-to-person transmission being enough to explain the simultaneous appearance of influenza in places far apart.” sounds like quackery, given that it was published in 1992.
Hi LA Bob et al,
I have family and work associates in California, they tell me it’s crazy that I can go to the beach and workout at the gym in Virginia (a blue state, so nobody in the media complains).
In fact, we could read the Washington Post bashing Florida for their “insane” and “deadly” policy of having open beaches while we enjoyed open beaches in Virginia, Maryland and Delaware. All allegedly within the Washington Post’s local coverage area. The Washington Post even ignored or downplayed a mini-scandal where our governor was photographed, maskless of course, taking selfies with large crowds on the Virginia Beach oceanfront right after ordering people to wear masks and social distance.
Seasonality? Think in terms of “people gathering indoors” rather than outdoor temperature. I think we had a bump in cases here mid-summer because it was so hot people who had gathered with friends outside went indoors for the air-conditioning.
And then you can see the spikes after any traditional holiday.
And all the college kids gathered in the fall for a few months to trade viral loads from all over the country, freely gave it to each other, and then their brilliant college administrators ordered them to go home and give it to their parents.
The Covid Tracking Project presents and interesting graphic (https://covidtracking.com/data/charts/deaths-per-million-by-state) which shows that the states hit hardest by Covid-19 early (New York, New Jersey and Connecticut) not only had dramatic fall-off in death rates through the summer, but have not seen an anywhere near comparable rise in death rates associated with the apparent seasonality of the virus. States that experienced dramatic rises in death rates in November and early December (North and South Dakotas and Iowa) also saw death rates fall rapidly. Suggests perhaps a degree of herd immunity does come into play. The latter three states have tended to be much less restrictive regarding Covid-19.
“States that experienced dramatic rises in death rates in November and early December (North and South Dakotas and Iowa) also saw death rates fall rapidly. Suggests perhaps a degree of herd immunity does come into play. The latter three states have tended to be much less restrictive regarding Covid-19.”
Go back and look at Arizona in the summer. We had the highest rates in the country, and then we had the lowest rates.
Then look at it since then. Once again, we have the highest rates in the country.
See the last graph on the following page.
http://91-divoc.com/pages/covid-visualization/?chart=states-normalized&highlight=Arizona&show=highlight-only&y=both&scale=linear&data=cases-daily-7&data-source=jhu&xaxis=right&extra=North%20Dakota&extraData=deaths&extraDataScale=separately#states-normalized
That is a great link. But TX, CA, FL, NJ, and NY all had two rather high peaks.
With the very first death from the virus there is a “degree of ‘herd immunity’ com[ing] into play.”
This is not a particularly useful application of the term herd immunity.
Useful twitter thread showing (lack of) correlation of first wave infections/deaths to second wave, using US data
https://twitter.com/youyanggu/status/1347266544946929665
The UK is seeing higher numbers of deaths with a positive covid test than in April, but overall mortality is now back to normal for the time of year in England, below normal in Wales and N Ire, and way below normal in Scotland. I guess they must have Dominion machines counting all these new cases and covid deaths.
At the foot of the page:
https://www.euromomo.eu/graphs-and-maps
England has been adjusted back upward for the last week.
Where is the mention of the mandated PCR test cycle count (amplification rate), and it’s impact on positive testing results and influence on the 2nd wave data? IIRC, the CDC/FDA held the recommended level to a 40 Ct rate until recently.
“However, increased population immunity resulting from some combination of further spread of infections and vaccination programmes, the combination varying from one country and region to another, should bring COVID-19 epidemics under control within the next few months.”
You get a choice (1) Forgo the vaccine entirely and reach HIT naturally or (2) take the vaccine and reach HIT artificially. Standard practice has been to go with (2) as that is what the vast majority of MD’s imply when speaking of herd immunity. /: Just like every other infectious disease has been treated in the past since at least the polio vaccine. This is a prime example of the fallacy of moving the goal posts. As perhaps tens of millions would have died if we even tried to reach HIT via letting nature take its course and living your life as if there was nothing to see here please move along to your own more probable death.
Those who claim that lockdowns have no effect on the progress of the virus should study Australia’s management. One has to be careful to define what can be seen as a “gain” or “success”. Australia currently has varieties of lockdown in all of its 6 States, who control management of the virus. There are currently under 10 new cases a day reported nationwide. But is this success? It might be no more than an artificial adjustment of the time base, meaning that we are simply postponing an inevitable worsening, at the economic cost associated with lockdowns. The hope is that future vaccination, due to start in a month or so and end 6 months later, will make theories of acquired herd immunity somewhat irrelevant. That is a hope yet to be tested.
However, Australia and New Zealand are strong examples that lockdown can be used to recover from a growing infection rate and to hold that rate close to zero, for whatever the purpose might be.
BTW, I deplore the actions of those here who are critical of Nic Lewis in a nasty, personal way. Nic is attempting to clarify a complicated piece of medical science that has some math features best handled by a mathematician. The eventual understanding of the problem is not helped by criticism of those trying to help understanding. Modern science has attracted many commentators of unproven scientific ability who seem to enthuse about cancel culture processes developed, inter alia, in or around the climate science community, which will not end up well regarded in the fullness of time. Geoff S
You are wrong. Lift your game. This is a science blog.
Geoff S
I agree with Geoff…Joshua you are way off base with your attitude…It’s simple without getting your undies all bunched up… nick had a thought about transmission of the virus relating to heard immunity…. turns out that idea was probably not so good in hind sight….so what….I can’t find anyone, anywhere who has accurately predicted the path of the Pandemic. Would be nice if you took that little bit of “vengence” that pervades your on-line persona somewhere else. It’s a problem you have to work out on your own.
Josh is often illogical, very biased, often wrong, and talks way too much. Word count to substance ratio is very high even for anonymous nonscientist ankle biters. That he thinks he can show such disrespect to serious scientists is arrogant.
SteveS –
> turns out that idea was probably not so good in hind sight…
Except it’s not only in hindsight. I agree that there’s certainly nothing wrong with Nic using his talents to build toy models of COVID spread. But he made obvious errors that were pointed out quite a while ago and for.mknths he resisted accepting evidence that obviously ran in contrast to his theories.
Further, he used his science in support of policy advocacy, and again in his policy advocacy he selectively ignored uncertainties that ran counter to his preferred policies.
So that’s life. And and just as. I one present perfect analysis, there’s also nothing wrong with criticizing his analytical output nor with questioning whether his work reflects ideological biases.
Hopefully criticism can be offered without turning personal but that gets tricky when someone offers analysis that is presented as bias free yet reflects bias. I’m certainly not going to cry about it if someone feels I’ve done do unfairly and calls me on it and I’m more than happy to examine whether I’ve stepped over the line.
On the other hand. Nic is a big boy and I’m sure the he can easily get past it if I’ve said something that’s unfair.
Josh, You pointed to no specific scientific or mathematical error. You spent many hours of your precious life on sarcastic and very vague accusations with no specific point.
>You pointed to no specific scientific or mathematical error.
I certainly pointed to no mathematical error. I don’t question the validity of Nic’s math, as I wouldn’t be able to do so. Nor do I question the value of him doing mathematical explorations of the pandemic. He has the potential to either make an extremely valuable contribution through providing new insights into how the pandemic might play out, or to make less important but useful contributions by showing the ways that people can erroneously model the pandemic.
As for “specific scientific errors,” I think I have pointed to specific mistakes he made in his scientific processes – in his treatment of uncertainties and in his weaknesses in controlling for potential biases. Those are fundamental scientific errors. For example, it is a specific scientific error to extrapolate from unrepresentative samples, as I think he has done numerous times in his posts on COVID.
At any rate, it’s obvious that Nic made specific scientific or mathematical errors in his analyses even if I haven’t outlined exactly what they are. That’s obvious because the conclusions that he drew were falsified by reality, subsequently. I’d guess that for the most part, his errors where in misguided assumptions about human behavior or a lack of respect for the uncertainties involved in projecting mathematical calculations into complex social contexts – but that’s to be expected when someone attempts an endeavor such as Nic attempted. That doesn’t mean I’m criticizing him for making an attempt.
My point is that he should have been (1) more circumspect about his ability to model complex behaviors amid an incredibly complex social context (2), been more open to looking at how unfolding events conflicted with his theoretical analysis and most importantly, (3) been more careful to at least attempt to control for obvious confounding variables before attempting to project the impact of his modeling mixed with his personal feelings about optimal social policy, so as to prescribe that his preferred policies be followed. It’s certainly his right to do all of that. Just as it is certainly my right to criticize him for doing so. I don’t particularly care whether you like it or not if I do that.
Again, my primary criticism is that Nic looked at Sweden, mistakenly simplified how the behavioral context influenced the presumptions underlying his modeling, and then prescribed policies based on having done so. And also I think he should have been less resistant to integrating how events unfolded into his analytical processes.
But worse, in doing so he ignored the obviously confounding factors that would enter into play as he prescribed extrapolating from Sweden to advocate the for implementation of similar policies in other, vastly different contexts. Of course, he has the right to do all of that. But given that a US Senator pointed to his mistaken modeling as a partial basis for developing US policies, the implications became real in terms of how lives are affected.
Epidemiology is a very inexact science with huge uncertainties. Given the ill-posed nature of the problem and the models, Nic has done a reasonable job here of showing how uncertain the math shows things to be. He did that too with regard to the earlier posts on heterogeneous susceptibility and Sweden. Perhaps you are just too math illiterate to understand the implications of that math. This is why the scientific community as a whole has a bad track record with regard to predicting the course of this epidemic.
Do you have no idea what an uncertainty analysis is? A simple one would be to take any of the models Nic has dealt with and show outcomes as a function of the parameters.
You have no idea what constitutes “bias” and how to control for it since you don’t understand uncertainty analysis. Here’s some material that might help you understand selection and positive results bias.
https://royalsocietypublishing.org/doi/full/10.1098/rsos.160384
https://www.nature.com/news/registered-clinical-trials-make-positive-findings-vanish-1.18181
https://www.nature.com/news/beware-the-creeping-cracks-of-bias-1.10600
Your long and repetitious comment shows that you have no specific scientific criticism and misrepresent what Nic said and the implied uncertainty, which is large in this field.
Science is in real trouble right now and independent people like Nic play a vital role in reform.
https://onlinelibrary.wiley.com/doi/abs/10.1111/eci.13484
New paper on effectiveness of stay at home orders and school closures. Lots of references too.
Some genuine good news for the US. A home test kit which will tell you whether you have covid-19 within 15 mins of a nasal swab. It’s been approved by FDA and is in mass production phase. Mid 2021 is the expected release date:
https://youtu.be/cssC5HY6m0Q
well herd immunity is an interesting concept..somewhat true..is it useful?
On one end we must deal with contamination which is extremely complex with the consequence of high uncertainty on the other hand, we must assume that immunity is acquired.. the virus mutates..at a given pace and the population of the world changes a a given pace too…a virus epidemic may never “ends”
so herd immunity concept is useful in cas of ” yearly” vaccination or to compute the number of cases to expect..
it seems to me it only help to be able to be sure…this epidemic is over now..
as climate change it is an interesting scientific question.. but we probably will not be able to answer it before having observed it.
Nic: Looking at Tkachenko’s paper, I have serious questions about how accurately the heterogeneity parameter λ can be determined via plots of log Re(t)/R_0 vs S(t). Figures 3 and S3. S(t) is the susceptible fraction of the population at any time t. Unfortunately, testing misses essentially all asymptomatic infections and symptomatic infections that are mild enough that people aren’t tested. For NYC there is a linear relationship from 5% immune/95% susceptible to about 25% infected/75% susceptible. Do we have good seropositivity data at any point in the pandemic that allows us to correct for missed infections? It isn’t clear to me what period the data is from. Today’s data (NYT website for NYC) shows that a cumulative 5.8% of the population has tested positive, and probably only 3% if the data in Figure 3 is exclusively from this spring. If correct, it seems that this analysis is assuming that about 9 cases were missed for every one detected by PCR assay. If we are only missing 4 cases for every detected case, λ would be around 2 and HIT would be a more reasonable 40-50%. I’ve commented on different posts why testing can’t possibly be missing 90% of cases and don’t want to debate that subject now.
Instead I’d like to ask what happens if we look at the same plot where S(t) runs from 0 (everyone immune/no one susceptible) to 1 (no one immune/everyone susceptible). Perhaps up to half of the detected infections in NYC have come after the spring surge. In that case, we might be approaching 50% of the total population being immune. In that case, log Re(t)/R_0 must be very small. Today the rate of detection of new cases is 50% higher than it during the worst of the pandemic. It is hard to see how data points from today’s surge can possibly fit on this graph. (The most recent points (fewest susceptible fraction) were already bending upward. Intuitively, this graph is saying that today’s surge is “impossible” because we have already reached herd immunity. Indeed, for λ = 4, your table says herd immunity is reached between 70% and 75% susceptible for R_0 = 3-4 and the final size of the epidemic is 44% to 53%
Or it would be impossible if R_0 weren’t allow to change. Despite its subscript zero, R_0 isn’t a constant. It changes with social distancing and mask wearing. And Re(t) is calculated from S, which is uncertain. Figure S2 shows what happens if R_0 changes with Google’s mobility data. In this graph, the vast majority of the reduction in transmission is explained by a change in R_0 and not the decreasing fraction of the population that is susceptible. There no longer is a nice linear relationship, but the overall slope could be closer to λ = 2.
Finally, looking at Figure S3 for 11 states, we see that the data doesn’t fit theoretical expectations very well.
In summary, I have no doubt that the most gregarious people in a pandemic will become immune faster that the less gregarious people and lower herd immunity. The same if there are significant differences in susceptibility. I just don’t think that we have reliable enough data tell if λ is 4 or 1.4.
Respectfully, Frank
Nic: The other thing I struggle with is a tangible interpretation for λs or λb (dispersion in number of contact/day and dispersion is biological susceptibility). Let’s start with λs. Let’s say the average person has 5 potentially infectious contacts per day. If dispersion means the standard deviation of the number of contacts per day is 1 (20% of the mean), then normally distributed dispersion would mean about 97% of people have between 3 and 7 contacts per day. What would that make λs? See Equation 12. Now I can ask the same question about a standard deviation that is 2 (40% of the mean), but asking about a standard deviation that is 60% of the mean begins to get absurd because we can’t have a negative number of contacts. So perhaps my question needs to be reformulated in terms of some other distribution, perhaps a log normal distribution. (Moving to a log scale makes things less tangible to me.) So the question is how do λs of 1, 2, 3 or 4 translate into a more tangible measure of dispersion. If you tell me that λs = 2 means the average person has 5 potentially infectious contacts per day and the 10th percentile has 0.5 contact per day and the 90th percentile has 50 contacts per day, then my intuition says your model has an unreasonable amount of dispersion. If you tell me that λs = 2 means the average person has 5 contacts per day and the 10th percentile has 2 contacts per day and the 90th percentile has 12.5 contacts per day, then my intuition says your model has a reasonable amount of dispersion.
λb comes from a different formula. And λ = λs*λb. So one could perhaps get an intuitive understanding of λb = λs = 2 and a few other combinations.
Franktoo, I agree that Tkachenko’s estimates of λ for NYC and/or Chicago might be inaccurate, but they also obtain similar results from other data. Moreover, their estimate of the social CV being 1 and the biological CV being at least 1 were derived from data on respectively contact network properties and age-group differences, both of which are independent of the NYC and Chicago epidemic data.
You say “Despite its subscript zero, R_0 isn’t a constant. It changes with social distancing and mask wearing.” Yes; changes in R_0 are a focus of my article. But seasonality appears also to be an important determinant of R_0 – maybe the most important one. Infections stayed low in the summer when people reduced social distancing (e.g. in Sweden, but also in other countries) but took off in the autumn and winter even with stronger social distancing.
You might find this paper from the Max Planck Institute for the Physics of Complex Systems interesting: https://doi.org/10.1371/journal.pone.0239678 . They conclude that “quantifying the the effects of mitigation requires knowledge on the degree of heterogeneity in the population”. They suggest that the German data is more consistent with their heterogeneity parameter, alpha, being no more than 0.2, which corresponds to lamba = 6, if mitigation had been mild. However, they find that if mitigation had been strong then heterogeneity could have been much smaller. The confounding effect of mitigation is a key reason for my past concentration on Sweden, where mitigation has been less strong than in most other western countries.
Nic: Thanks for the thoughtful reply. You quote the Max Planck paper:
“quantifying the the effects of mitigation requires knowledge on the degree of heterogeneity in the population”. However, the converse is also true: quantifying the the effects of heterogeneity in the population requires knowledge of the effectiveness of mitigation. It also requires knowledge of the ratio of detected cases to total cases (resulting in immunity), because the slowing of the pandemic is mostly caused by depletion of susceptible people and more-susceptible people. If you don’t include heterogeneity, then slowing of any pandemic well short of herd immunity is attributed changes in R_0. When you add a heterogeneity parameter, you can explain any reduction in new cases/deaths by whatever combination of changes in R_0 and heterogeneity you choose. And when you can have anywhere from 2 to 10 total cases per detected case (let’s call this P for PCR), you really have no idea of how fast the pool of seronegative people is diminishing. (I don’t see where the Max Planck paper discusses P, but it is obvious critical to any analysis.) Many combinations of P, changing R_0, and heterogeneity parameter can fit the spring wave of this pandemic. Isn’t that the conclusion we should draw from the graphs in Figure 6 and the quote you cite?
The fall wave of this pandemic is now placing some limits on these parameters. So all papers that don’t include information from the fall surge are now obsolete! When we have a record number of current cases and cumulative detected infections approaching 15% in many places, it is difficult to believe that P might be above 3, except early in the pandemic when testing was limited. This is especially true if heterogeneity is an important factor in limiting the pandemic. As we have discussed, comparing the IFR in hard hit US states in the spring to the IFR the north central states hit hard only this fall suggests that P has decreased less than a factor of two. Germany’s CFR has only fallen a factor of slightly more than 2 and there must have been some improvement in treatment and shift to younger, more-survivable cases.
Germany now has 2 million cumulative cases of COVID. If you look Figures 6b, 6e and 6h, you’ll see that only the scenario (6b) with strong mitigation and no heterogeneity allows cumulative infections in the absence of mitigation to rise above 1 million cases. If you look at Figure 6i (high heterogeneity, low mitigation), R_0 had returned to nearly 2 in July, which meant that the pandemic was being held in check mostly by the depletion of the most susceptible population, while mitigation was having minimal effect. The total number of cases in Germany has risen 8-fold since July. Either R_0 has risen dramatically, or the high heterogeneity scenarios are absurd.
FWIW, There are some papers suggesting that seasonal flu occurs almost exclusively in the winter because infectious aerosols persist longer when the air is dry and more evaporation leaves lighter droplets. R_0 increases in the winter, a seasonal epidemic grows until it is limited by approaching herd immunity in March and soon after warmer temperature make indoor air more humid. If so, the same factor likely caused the fall surge in COVID which hit the coldest US states the earliest. I’m no longer satisfied with cloth mask that blocks droplets, but lets most aerosol through.
Frank –
I mentioned this before.
> And when you can have anywhere from 2 to 10 total cases per detected case
We have 23 million identified cases in the US. Let’s say P is at the low end of your range. That would mean a population infection rate of 14% and an IFR of 0.85% as it stands right now. Of course, we’d have to add more fatalities into that because if we kept the # of infections the same there would be another month or so of deaths, so let’s maybe say another 30 days of around 1500 deaths per day averaged over that entire period? Which would bring the IFR up to about 0.94%
So that’s way higher than some of our friends are claiming. Almost 3 x as much.
Let’s look at the other end. Let’s say that P is 8. That would bring the IFR more in line with what some of our friends are claiming – about 0.23%.
Hmmm, but that would mean that the population infection total would be about 57% right now. And the infection rate actually still increasing, but lets say the infection rate started dropping tomorrow and reached 0 in a couple of weeks….even then the population infection total would have to actually be a bit higher.
But our friends say that herd immunity should kick in at maybe 15%.
Seems to me that somethings gotta give.
Joshua: I suspect we agree, but are approaching the problem from different directions. I think it is absurd for COVID and influenza to have similarly low IFR’s, as many concluded after the early seropositivity surveys reported about 10 undetected cases for every one detected case. (That high ratio could have been an artifact from a false positive rate of a few percent in the field.) This led to the early conclusion that COVID and influenza had similar IFRs. Both are transmitted by the same combination of mechanism and now show clear signs of the same seasonality.
Given that COVID has been killing as many Americans in an average month as influenza does in the average year, an IFR similar to influenza seems absurd. And this is happening despite our best efforts to suppress the pandemic – efforts we rarely make for influenza. Those efforts appear to be suppressing the seasonal influenza pandemic that would normally becoming apparent by now. And most of the COVID pandemic has occurred in months when influenza kills few Americans. We are only beginning to experience the lethality of COVID during the months when most influenza deaths take place. My guessestimate is that COVID is likely to kill about 50 times as many Americans in a year as influenza does during the average year. (We are already about 12 times as many.)
Nevertheless, the above measures of deadliness do not translate directly into a high IFR: Deadliness is a function of transmissibility, lethality per infection and the effectiveness of measures to mitigate the pandemic. Deadliness merely suggests COVID should have an IFR significantly greater than influenza. I’ve commented previously about many places where the cumulative number of detected cases is approaching 15% and at least one where the pandemic is still raging. If there were more than 1-2 undetected cases for every detected one, approaching herd immunity would be playing a significant role in slowing transmission. So I don’t think the current CRF/IFR ratio is greater than 2-3. Early in the US pandemic (in the Northeast), the CFR was about two-fold greater than it is in Montana, and some of that 2-fold improvement comes from improve treatment and the lower age of those being infected (as the elderly living in the community protect themselves more effectively.)
Epidemiology papers discussing the importance of heterogeneity are all based on observations made before the fall surge in the pandemic. Those papers showed a wide range in heterogeneity was consistent with the spring surge in some locations, but they didn’t rule out many possibilities. Those skeptical of mitigation naturally emphasized the possibility that heterogeneity and approaching herd immunity were playing a role much sooner than expected. Assuming we are right, there will be definitive papers modeling both the spring and fall wavers showing that heterogeneity must be playing a less important role. However, given human nature, the papers refining how big (or small) a role heterogeneity plays will probably appear more slowly than the exciting papers offering hope of earlier herd immunity. In the meantime, it’s not surprising to me that our disorganized handwaving arguments are not as compelling as we think they should be to others invested in heterogeneity. The best I can do is draw attention to data I find important and see if others eventually agree or poke holes in my arguments.
Most of all, I don’t want to erect a straw man by saying our friends claimed herd immunity will kick in at 15%, when they probably said or meant to say, that X, Y and Z suggest that herd immunity could be as low as 15%.
Frank –
> Most of all, I don’t want to erect a straw man by saying our friends claimed herd immunity will kick in at 15%, when they probably said or meant to say, that X, Y and Z suggest that herd immunity could be as low as 15%.
That’s fair. I won’t pretend that I can sometimes be prone to stretching what people say into the shape of straw men. Still, people should be careful to highly uncertainties and to foreground caveats. And they should be quick to reinterpret their analyses given that facts change on the ground. Nic has done all of that poorly, IMO, in this situation. Consider that a US Senator referred to Nic’s work to support a firmly stated conclusion that NYC had reached “herd immunity” and that the country should implement policies accordingly.
https://twitter.com/randpaul/status/1283753633247563777?lang=en
That doesn’t mean that Nic shouldn’t be building toy models or that his doing so doesn’t make a contribution to the science. And it certainly doesn’t mean that he shouldn’t advocate for the policies he favors.
Frank –
I assume you’ve seen this…but as for seasonality.
https://twitter.com/B_Nelson_Manaus/status/1349705393249923072
I think there’s a reason that it seems many people who study this a lot are reluctant to weigh on on the strength of the signal of seasonality for COVID.
Josh: Yes, I know about Manaus and have mentioned it. However, I can find prisons in the US where the pandemic has infected more than 80% of the population and a ski/party town in Tyrol that was evacuated/isolated a few weeks into pandemic in Northern Italy. More than 40% of locals were seropositive in spring and returning skier spread the pandemic through Europe. Are the conditions in any of these places relevant to the US and Europe? Clearly there are some places where the disease in innately more transmissible and the HIT will be higher. To be rigorous, the first thing I like to know is whether R_0 (or just the doubling time) early in the pandemic in Manaus similar to what hard hit areas in the US experienced before mitigation began. Then I could be more confident that the lessons from Manaus should be applied elsewhere. (My instinct says yes.) However, I suspect Manaus probably lacked adequate testing and deaths will need to be used to determine the time course of the pandemic. I’d love to find a paper that discusses whether Manaus is an appropriate model for the US and Europe.
Frank –
I don’t know if you’ll see this…
I was just perusing Worldometers and looking at Denmark and Sweden.
So Sweden has considerably more cases per capita and far less testing per capita (suggesting that the gap in infections is even bigger than it would seem when you look at the difference in identified cases only). One would think that they might see the the “herd immunity” effects of having more people infected and recovered (or died) reflected in a lower infection rate, and maybe even see that differentially lower infection rate increase over time as the relatively higher per capita rate of infections in Sweden has increased over time.
Only problem with that idea is that they have a higher infection rate in Sweden – although it does seem that over time the infection rate in Denmark has been closing the gap.
So where is the signal of “herd immunity?”
Here is a piece on SARS-CoV-2 written by Dr Irina Metzler FRHistS, medical historian and former lecturer at the University of Swansea, as well as a Wellcome Trust University Award Fellow, which covers a lot of stuff you may not have read elsewhere especially focusing on what maintaining good health in children and adults means in the real world and why we may be seeing this epidemic run rife in the UK etc.
https://lockdownsceptics.org/covid-hyper-medicalisation-and-virus-interference/
From UK – Weather’s linked article –
“The hygiene hypothesis is well known and argues that the more ‘clean’ we have become, the less chance our immune systems have had to be ‘trained’ in how to ward off pathogens. Hyper-hygienic conditions, which have been advocated in most high-income countries, through things as basic as using anti-bacterial products for everything from chopping boards for food preparation to the now ubiquitous hand sanitisers, have in fact contributed to the lack of training in childhood for most Westerners’ immune systems.”
This points out one of the long term dangers from the lockdown/mask police / mitigation proponents. They are essentially advocating that the human race evolve to where humans can only survive in a sterile environment
FWIW, new orleans had the lowest positivity rate until just recently in the state of louisiana (during, this, our third wave)…
Thanks Nik,
I’m going to drop this here intending to circle back and comment in a reply.
Sadly Brazil has given us solid epidemiological data on SARS2 spread absent mitigation/suppression: Attack rate on the order of 75%
https://science.sciencemag.org/content/early/2020/12/07/science.abe9728
Covid-19 is an immunopathological disease (think AIDS). Seropositivity can not be factored into the SEIR model as presented. An adaptive immune response results in severe disease. Recall that Trump was hypoxic, in trouble, and seronegative at the time of administration of the Regeneron mAb.
Most infections are neutralized my mucosal immunoglobulin.
There’s an interesting critique of that convenience sampling methodology:
https://twitter.com/WesPegden/status/1336801382955933697?s=19
So typical of the scientifically illiterate. A peer reviewed published paper is criti
critiqued in a Twitter thread. What pseudo-science.
If you read my article, you would see that I cite the paper you mention and show why its claims are questionable, and that other data suggests its 76% figure is quite likely a gross overestimate.
So much for herd immunity in Manaus:
https://twitter.com/B_Nelson_Manaus/status/1349705393249923072?s=19
More for the “but lockdowns are killing people” crowd to think about as they rest on counterfactual assumptions about what would have been different absent interventions:
https://www.ft.com/content/9cd1d3dd-c9b5-424b-a84c-30b788a15ed1
This is a really interesting video about the UK variants. It is an exchange between a virologist and an epidemiologist about the possibility of the CoV-2 variants being more or less transmissible that the founder and what is needed to work the answer out. It covers work by Amy O’Toole in Edinburgh with official data and her observations of what is going on and the problems with other experts jumping the gun.
The video is produced by Vincent Racaniello who tirelessly brings to us good quality science (search for him on YouTube) and has done throughout the pandemic. Frankly speaking it is far more enlightening and heartening to watch (these guys give you no BS) than the daily Downing Street Saga which may be running short of believable anything.
https://www.cnbc.com/2021/01/13/moderna-ceo-says-the-world-will-have-to-live-with-the-coronavirus-forever.html
Those who fanticize about ‘crushing’ the virus need to read this. COVID-19 will join influenza as a yearly epidemic with new strains. It will require new vaccines at least every year.
Moderna sell vaccines – and he’s right for now but I hear the massive R&D they did for COVID has lead to incredible discoveries in bioscience that will extend far beyond the current pandemic. Yesterday Moderna announced is developing three new mRNA based vaccines for seasonal flu, hiv and nipah virus.
Check out AlphaFold and their breakthrough with molecular protein folding or these guys who seem close to isolating the cause of so many COVID symptoms.
“The Odd Structure of ORF8: Scientists Map the Coronavirus Protein Linked to Immune Evasion and Disease Severity”
https://newscenter.lbl.gov/2021/01/12/odd-structure-of-orf8/
Roomers are that a cure for
Scientific progress is always good of course.
Nic Lewis, I respect your attempts to model the disease, but I think the net effect of these essays is to show that the model is too simple to be accurate, and information on the necessary entities (cf Ockham’s Razor) is limited.
Yes Matt. Nic has done a good job of analysis but modeling of epidemics seems to me to be an ill-posed problem and most predictions will turn out to be wrong to some degree or another.
Matthew,
I don’t think complexity of the model is not the issue.
Others have used much simpler models with reliable results.
Here’s one from a climate scientist…
https://bskiesresearch.wordpress.com/2020/04/14/model-calibration-nowcasting-and-operational-prediction-of-the-covid-19-pandemic/
The issue, I think, is more generating an extraordinary result from a model and not questioning its veracity. Virtually nobody else believed herd immunity was in action. A model from anyone naive to the field generating such a result should always drive questioning the model first, rather than the whole of the rest of the field.
verytallguy: https://bskiesresearch.wordpress.com/2020/04/14/model-calibration-nowcasting-and-operational-prediction-of-the-covid-19-pandemic/
Thank you for the link. Where is the second wave?
No one did a good job with predicting the second wave because the models cannot predict things like new strains, or seasonal effects or mitigation on R. There is not a lot of good science yet. We can’t even predict influenza with any accuracy. Each year the severity is significantly different.
For an ill-posed problem modeling will provide only very limited predictive skill. This is well known to anyone with any mathematical training.
It is common for laymen to believe the biased literature which has a strong positive results bias.
Matthew,
2nd wave depends on lifting restrictions, which was not knowable.
The level of prevalence in the uk was nowhere near herd immunity at the time. Nor in Sweden, obviously.
VTG, You are so obviously wrong. The second wave is just as severe in California and New York as Texas and Florida. California has had a tightening of an already extreme lockdown as has New York. Texas and Florida have only very light restrictions. Lifting restrictions had nothing to do with it.
Most viruses are strongly seasonal including influenza.
dpy,
A link to the views of mine you’re quoting on seasonality would be lovely.
Thank you!
Compare and contrast:
> There is not a lot of good science yet.
and
> VTG, You are so obviously wrong.
The causality behind the efficacy of interventions is obviously, extremely complicated. As Frieden says:
https://twitter.com/DrTomFrieden/status/1347711472282046473
I”d love to see a comprehensive treatment of this issue.
verytallguy: 2nd wave depends on lifting restrictions, which was not knowable.
If you look at the experiences of New York, Florida, Texas, and California, I don’t think you can sustain a claim that the 2nd wave resulted from changes in the lockdown orders.
When a forecast depends on knowing something that has not occurred yet, a partial remedy is to forecast several evolutions of that something, and make the forecast for each such evolution. Climate modelers have done this plenty.
Matt, The problem with ensembles here is that the problem is ill-posed so the range of outcomes is huge.
I think the primary problem with modeling is that humans react to the virus and that changes the parameters in unpredictable and difficult to measure ways.
In late February I wrote my own SEIR model just to get an intuitive feel for what might be coming. The model produced the same results, roughly, as Ferguson’s, except mine has only 100-200 lines of Python code. A
nd, the results turned out to not be predictive at all, because of changes in behavior. Those changes are both the result of government, and very importantly, individuals acting on whatever information they get. In the age of the Internet, far more information, including good information and bad, is available than governments put out (I’m talking about intentional hiding), and people react to it.
That means that Rt (or Re as some use) deviates radically from R0, and does so in ways hard to predict. If Rt remained at R0, my model would have been pretty close – it’s basic epidemiological laws, and it showed a Farr’s Law shape, of course.
Here in Arizona, we were late to put in mitigations in the Spring, but well before mitigations, many of us were sharply curtailing interactions with those outside our household.
If we look at the whole history of our three waves (the third currently at record levels), we see:
First wave… no mitigation until late, then pretty strong mitigation, including stay-at-home orders (with significant exceptions) and closure of all but “essential” stores and jobs. The wave peaked and only subsided a bit.
Then, as the temperatures of our normally very hot summer season came, we had a record peak for the US. – even though it was too hot to have much in the way of outdoor activity, even after dark.
Not long before the peak, the governor finally allowed localities to impost mask mandates for the first time, resulting in 90% of the population being under those. He also closed or limited some high risk businesses. The wave subsided dramatically, with the Rt dropping to around 0.7.
Then, as our winter arrived, and with it, very comfortable temperatures for outdoor activities, our third wave came, and has yet to end, although it might have leveled off. Government measures hardly changed at all and still have not, even after repeated requests from medical officials.
The cause of the rise? It’s hard to say. Schools opened, but we keep being told that schools are not a major source of spread. The holidays arrived, around the same time as our “snowbirds” – people, mostly retired, who winter in Arizona and spend the rest of the year in northern climes.
It would be hard for a model to deal with all of this, because the behavior is so hard to anticipate. The transmission of the disease is far easier to model than the behavior of the people.
It would also be hard to use modeling to deduce much about the underlying factors – heterogeneity, mobility, immunity , given this complexity. And that’s my point – models work, but only if you know what to feed them, and variable human behavior defeats that here.
For the “lockdowns don’t work” folks. That’s nonsense, but they only work if they are enforced. And, of course, they imposed huge costs, but that’s a different issue from “don’t work.”
Personally, I would like to see some stronger measures here. At a minimum, something to reduce the mobility – I see way more cars on the road than at any previous time in the pandemic, even when our rate of cases is twice that of our previous (July) peak, and our hospitals are stressed to the point that the largest chain banned all non-emergency procedures starting the first of this year.
For me the problem here is that the evidence that stricter measures work is not there. There are studies showing a variety of results. Strict measures of course impact the most vulnerable most strongly. Upper middle class people can still pay the bills. Most people when they get laid off cannot. Giant deficit spending to subsidize most people is not possible without very serious long term economic harm.
“For me the problem here is that the evidence that stricter measures work is not there.”
It depends on what you mean. They work well in China. They work in Australia – so far. And based on what is known about the mechanics of respiratory virus spread, they should work>
This issue is how strict, which measures, how much compliance you get, and very importantly, whether stricter measures actually increase compliance, especially given that a lot of people are acting based on information other than the little that comes from government.
So yes, they work. You raise the issue of the cost of them, and that’s a valid issue, but it doesn’t affect whether they work, but rather whether they are worth it.
“The problem here” is that people reach simplistic conclusions about the differential impact of interventions based on lazy assumptions about counterfactuals related to what would have been had things been different.
> Strict measures of course impact the most vulnerable most strongly.
The pandemic of course impacts the most vulnerable most strongly. And a lack of interventions would affect the vulnerable most strongly.
The beautiful irony here is that when I posted about how Covid was having a differential impact on minorities you say that was “irrelevant.”
> Most people when they get laid off cannot.
Without the interventions, it is quite possible that people would have been laid off anyway without extended unemployment benefits and without stimulus checks. People would have been forced to decide in many cases whether to protect their health or stay home from work and get fired.
There is no free lunch.
> Giant deficit spending to subsidize most people is not possible without very serious long term economic harm.
There’s plenty of evidence to suggest that in the long run, the best way to mitigate the economic impact to the poor and everyone else would have been to have a more robust federal assistance approach. And the return of investment on something like a massive rollout of antigen tests could have been a massive net positive.
There is no free lunch.
messo comment – ‘It depends on what you mean. They work well in China. They work in Australia – so far. And based on what is known about the mechanics of respiratory virus spread, they should work”
The mitigation steps worked well in china, Australia , NZ , SK, etc because they were able to stop it early before the virus infiltrated the general population, Once the virus became embetted in the general population, mitigation steps become far less effective
“he mitigation steps worked well in china, Australia , NZ , SK, etc because they were able to stop it early before the virus infiltrated the general population, Once the virus became embetted in the general population, mitigation steps become far less effective”
What becomes less effective is test and trace mitigation. The other mitigations become more, not less important, and their effectiveness doesn’t change.
Also, in China, they had a very high case rate in Wuhan, and they stopped it quickly and thoroughly through very strong lockdowns.
I had hoped that in the US, the spring lockdowns would bring it down enough to do test and trace. Unfortunately, we didn’t have adequate testing capacity when needed, and also tracing doesn’t work as well here due to resistance, and to privacy concerns.
I cited earlier a study in European countries finding no benefit from stronger lockdown measures. Europe is probably a better match to the US than Australia. China is not very comparable as they have a strongly authoritarian government and a population that has a culture of compliance.
> due to resistance, and to privacy concerns.
And cynical exploitation of the testing situation for the sake of political expediency. As in “Everyone who wants a test can get a test.” and out testing program is “beautiful.” and “The only reason we have soany cases is that we’re testing so much.” and arguments by political leaders that we’d be better off with less testing.
Of course, political leaders who supposedly deserve so much credit because they supposedly accomplished so much are entirely devoid of any responsibility for our failures in testing despite that they are the head of the government.
On R0:
The more thought I give to the concept of R0 the less useful it appears.
In fact, the number of individuals who’re on average infected by someone with the disease is virtually impossible to determine.
1. In the early stages of this epidemic, testing is scarce and so case numbers were dramatic underestimates.
2. Different countries or even different counties within a State will have dramatically different transmission rates.
3. Most viruses are strongly seasonal. Thus do we want to say that R0 can vary by a factor of 4 depending on which month the epidemic started in? Of course, its impossible to determine exactly when an epidemic started because initial growth is very small.
So, you can average over all seasons, all countries, etc. and get a number that will be incredibly difficult or impossible to compute accurately. This number is useless for predicting anything at a local or even country level where it’s actually important.
To be fair most epidemiologists know this and don’t really care much about R0. It seems to be mostly used as a communication tool to convince people epidemiology is a mature science that can predict things. The papers Nic has discussed here show that much more sophisticated models offer better results.
So, now you’ve had time to think, how many times has the uk reached herd immunity so far?
Once, twice, three times, more, or never?
This is a question I will leave as an exercise for you to answer using simple math software.
Well, at least you’ve realised just how ridiculous your argument is.
I suggest you do a refresher on reading comprehension. My argument is correct. R0 is not a magic universal constant for each new virus. In fact it is virtually impossible to compute and useless for predicting anything. R varies hugely depending on many things.
“I suggest you do a refresher on reading comprehension. My argument is correct. R0 is not a magic universal constant for each new virus. In fact it is virtually impossible to compute and useless for predicting anything. R varies hugely depending on many things.”
R0 is the closest simple approximation we have for the inherent transmissibility of a new virus in humans. It is a constant, not exactly a universal one.
R0 is what we would expect to see if an epidemic happened in an immunity immunilogically naive population without mitigation. Of course, there can be some different R0 values for different life styles of a population, but that’s about it. For a given population, R0 will only change through genome change in the virus (mutation, recombination).
R (Rt seems to be the most common notation to this non-expert) is a time varying transmissibility value, with the same meaning (of transmissibility) as R0. Rt varies based on many things, including: ratio of susceptible to total population; mitigation; travel changes. Rt is *not* R0. You can use it in the same formulas and models as R0, but it is a time varying function of R0.
R0 has a 0 in it for obvious reasons: the 0 time of the epidemic. That’s normal scientific notation for all sorts of parameters in many fields.
Dpy,
How many times has the uk been in and out of herd immunity so far?
Don’t be embarrassed, just inform us. We wait your wisdom.
I told you how to do the analysis yourself. I generally don’t waste time on unimportant questions that are being used as a disingenuous rhetorical device. Nonscientist anonymous activists often do that on blogs.
Many places saw R drop below 1 this summer. The UK was one of them. You must use smoothed data to determine anything. The data is very noisy.
It’s not a difficult question dpy.
How many times, according to you, has the uk been in herd immunity this pandemic?
It’s almost as though you’re not answering because your positing is indefensible.
You didn’t read what I wrote or your reading skills need a refresh. I said the UK crossed a herd immunity threshold this summer. As winter set in R increased a lot and the herd immunity threshold increased.
Kwok ( who I linked above) said that Hit varied from 5% to I think 78% or so on a country by country basis. On a regional or city basis it’s bound to be greater.
The question for you is why you continue with rhetorical and silly questions to which the answer is pretty clear given a moments thought. You don’t seem to be willing to try to advance the discussion.
Meso, Some people do define herd immunity as 1 – 1/R0. Others use other definitions. My argument is that using R0 is worthless for doing anything. We know that R will vary a lot depending on the season. So this definition using R0 gives a number that is strongly dependent on which month an epidemic starts in!!! At the very least you need to average R over the coarse of a year, but there is the effect of the response and its effect on R.
It’s just a silly definition. It makes nonscientists think they have learned something when what they have learned has little scientific value.
” My argument is that using R0 is worthless for doing anything.”
That’s simply nonsense. Scientists understand R0 and its limitations. I understand R0 and its limitations. You appear to believe that we don’t – that we treat it as a magic talisman or something. COVID19 is hardly the first epidemic for which this stuff has been used.
And, for the population for which R0 is derived, the HIT without heterogeneity is in fact 1 – 1/R0. That’s a simple fact and can be proven mathematically. Heterogenity(lambda) changes that, as has been discussed, but also in a deterministic way that can be calculated.
But when people say that a population reached herd immunity at some level, and then sometime later the same disease reaches higher levels, they were simply wrong (unless the pathogen changes). It’s that simple.
Herd immunity is when the herd won’t sustain an outbreak, under conditions that are the same (other than immunity) as when the R0 value (and lama, and any other decorations) was derived. That’s a very useful definition.
Mathematically, one models epidemics using R0 and factors that modify it. Those models work – if no mitigations take place. If you can predict the impact of mitigations and the level of them, the models will again work. If the weather changes and that modifies transmissibility (for example, increases aerosolization due to low humidity), that can again be factored in.
You end up with an R(t) (usually written Rt) – R0* [set of factors].
It is true that if you hold all of that constant, you can calculate that the outbreak *under those conditions* will peak at 1 – 1/Rt (and with lambda tossed in if you know it). But that is not herd immunity, and it confuses people to refer to it as such.
That false definition has been used to argue against mitigations. While there are arguments to be made against mitigations, the “we reached herd immunity at only 10% of the population” argument (when R0 and lambda make it much larger), are bogus. Likewise, arguments that “we must have reached herd immunity because the epidemic is receding” are bogus, but often used. People look at the curve, notice that it fits the shape of a herd immunity induced curve, and draw the wrong conclusion (unless there was no change in conditions from the R0 condition, in which case herd immunity was reached).
Of course, in an emergent disease situation, we don’t know a lot of this. We don’t know R0 accurately. We don’t have numerical values for the effects of specific mitigations. Even more difficult, is that we cannot predict behavior changes as a result of either government mandates, or independent decisions resulting from knowledge of the epidemic.
Yes, You just made the case that defining herd immunity using R0 has no practical implications. It is very possible to have reached the R0 herd immunity threshold but for the epidemic to keep growing for a very long time. That can happen if R keeps increasing as it might in winter time for example.
BTW, all viruses will mutate over time and develop new strains with different infectiousness. So which one should be used to define R0?
It’s a simple minded definition that tells us nothing either about how an epidemic will progress or for comparing various pathogens.
“Yes, You just made the case that defining herd immunity using R0 has no practical implications. It is very possible to have reached the R0 herd immunity threshold but for the epidemic to keep growing for a very long time. That can happen if R keeps increasing as it might in winter time for example.
BTW, all viruses will mutate over time and develop new strains with different infectiousness. So which one should be used to define R0?
It’s a simple minded definition that tells us nothing either about how an epidemic will progress or for comparing various pathogens.”
Your approach is to throw out the baby with the bathwater. You keep asserting R0 is worthless, and yet I have yet to see an epidemiologist say that.
You are correct that R0 might be different for winter than summer, although it probably doesn’t change much. Can you explain the mechanisms that would cause it to be different?
When viruses mutate in a way that significantly alters transmissibility, then R0 has to be re-estimated for that variant. So far, in this epidemic, that has only occurred once (more than one strain with the trait, both appeared recently). So, you adjust R0 up, of course – it’s a different pathogen.
But it is not a “simple minded definition” and it certainly does not “tells us nothing.” That’s an extreme statement.
Science needs simplifications. It relies on them. But they have to be used knowing that.
So, in epidemiology, it makes sense to look for a constant that defines transmissibility. That’s R0 (or for slightly more complex, R0 and a heterogeneity constant or two). It doesn’t mean that you stop at that point or that you assume you know everything. But it also doesn’t make it meaningless.
Your argument like saying that Newtonian physics is useless because relativistic physics is more accurate.
Kwok says in theNature piece that the hit varies from 5% to 78%.
The Newtonian physics example is very wrong. For velocities below say a million miles per hour, Newtonian physics is extremely accurate. The relativistic terms are less than machine precision on all computers.
If you believe Kwok, using a single R0 will give predictions that fill the possibilities space. That’s a useless model because it can’t predict anything.
It’s like trying to define with a single number the weather at a given place. You might get an average temperature over the year, but that is useless for predicting anything. Spokane and Seattle have similar average temps but totally different weather and climate. One is a maritime climate and the other is continental.
Try to think of a quantitative prediction using R0 that is skillful.
Any model that has a chance must try to model how R might change.
David –
> Kwok says in theNature piece that the hit varies from 5% to 78%.
And by your argument, with 0.1% (or less) of a population infected it could reach a “herd immunity threshold” as I illustrated in my example. And by definition Nic couldn’t be wrong no matter when he said they’d reached a HIT in Stockholm – no matter that they have an active pandemic now even though he said they’d crossed a HIT 8.minths ago.
Once again, based on his calculations, and based on HIS DEFINITION of herd immunity, Nic projected another 1,000 deaths or so from COVID in Sweden. Since that point there at another 5,000 or so deaths (and about 450K of their 530K identified cases). IOW, his definition of herd immunity and associated projection was off by 500% on the more favorable for him metric of deaths. But according to your logic he wasn’t in error.
You’re entitled to such a belief even if it is ridiculously. I prefer to think, however, that you know that it’s ridiculous but you’re just holding on desperately and trying to convince someone who isn’t actually evaluating your argument.
This is another meaningless response from you Josh. You can disagree with Kwok and deny science if you want. Nic was right about Sweden crossing a herd immunity threshold. Purely rhetorical sniping is childish and wastes your time.
Mr. Lewis,
If we use the formula you write about – what then is the theoretical ideal lockdown period?
Or put another way: are there criteria which falsify the premise of the equation?
Otherwise we’re all looking at a medical equivalent of the Drake’s equation – unprovable and thus useless.
In my view the ideal period for maximum restrictions (not necessarily a lockdown) is when herd immunity is being approached, so as to minimise the overshoot (the excess of the final epidemic size over the HIT). See my article https://judithcurry.com/2020/05/29/when-does-government-intervention-make-sense-for-covid-19/
niclewis: when herd immunity is being approached, so as to minimise the overshoot
Sounds good, except that there is no way to tell when herd immunity is being approached. Not that you have told us so far.
“Herd immunity by infection is not an option.”
https://science.sciencemag.org/content/371/6526/230
This article is based on the very questionable results of the Buss et al. study that estimated a 76% attack rate in Manaus.
I linked to a critique of their survey methodology (and findings) elsewhere in this thread. There’s also this:
-snip-
All the experts who spoke with The Scientist agree that the main limitation of this new study is that blood bank donors are not representative of the total population. The researchers adjusted for some of the demographic skew in the sample by comparing it to the population census, but the approach is far from perfect.
-snip-
Also questions about the assertion of “unmitigated.”
https://www.google.com/amp/s/www.the-scientist.com/news-opinion/study-estimates-76-percent-of-brazilian-city-exposed-to-sars-cov-2-68272/amp
Just goes to show – extrapolating from unrepresentative (particularly convenience) sampling – as with the Diamond Princess or the Santa Clara study or from Sweden to prescribe policy for other countries – is a recipe for confirmation bias
Texas is leading the country in number of shots administered.
https://www.nbcnews.com/health/health-news/map-covid-19-vaccination-tracker-across-u-s-n1252085
jim2 https://www.nbcnews.com/health/health-news/map-covid-19-vaccination-tracker-across-u-s-n1252085
Thank you for the link.
Seems that per capita would be a much more meaningful measures.
Amazing that West Virginia, of all places, leads in that regard (doing considerably better than Texas, for example). I wonder why.
Texas is notable because it is a large state which makes logistics more complex. Yet the governor there has been able to get the vaccine injected rather than rejected.
“Seems per capita would be a much more meaningful measure.”
Only to a fool.
You have a limited number of vaccine doses provided to each State. The States responsibly is to get those doses administered as quickly as possible to appropriate individuals. Tracking per capita hides the poor performance of States.
Lol.
> You have a limited number of vaccine doses provided to each State.
So that’s a reason to go by the total number of vaccines given as opposed to the per capita rates? Which was my point – as a comparison of the two, not which single metric in and of itself would be best. Obviously, all these factors plus more factors should be considered.
Looks like Texas does rank 6th in % of vaccines distributed. But not that much higher than a lot of other states. All things considered, that seems like a rather weak basis on which to make substantive comparisons.
Given how so many conditions would vary by state, such as infrastructure available or the level of demand in the various states or the cost efficiencies of the rollout, or the rate of the distribution to the states – all of which I’m guessing you haven’t a clue about, it looks like yet another example of where people look at life and death matters and care most about scoring partisan points.
Sad. But knock yourself out if it helps you to feel good about yourself.
Tracking percentage of unallocated vaccine doses hides the actual number of doses available but unadministered. States should be getting the vaccine administered asap. They should also track number of doses that go bad due to being beyond expiration date.
Think. What is the purpose of metrics.
lol. So do you feel the same way about NY, which is also a large state and where the per capita vaccination is roughly on par with Texas? ‘Cause personally, living in NY, I think the vaccination rollout has been terrible here.
Or how about Alaska, a large state where the per capita vaccination is quite a bit higher?
It’s really sad that you’re so focused on politicizing everything.
Latest on lockdown orders:
https://onlinelibrary.wiley.com/doi/10.1111/eci.13484
A little dose of NPI is good; more NPI is no further improvement.
snippet: This analysis ties together observations about the possible effectiveness of NPIs with COVID-19 epidemic case growth changes that appear surprisingly similar despite wide variation in national policies.31–33 Our behavioral model of NPIs – that their effectiveness depends on individual behavior for which policies provide a noisy nudge – help explain why the degree of NPI restrictiveness does not seem to explain the decline in case growth rate. Data on individual behaviors such as visits to businesses, walking, or driving show dramatic declines days to weeks prior to the implementation of business closures and mandatory stay-at-home orders in our study countries, consistent with the behavioral mechanisms noted above.34–36 These observations are consistent with a model where the
severity of the risk perceived by individuals was a stronger driver of anti-contagion behaviors than the specific nature of the NPIs. In other words, reductions in social activities that led to reduction in case growth were happening prior to implementation of mrNPIs because populations in affected countries were internalizing the impact of the pandemic in China, Italy, and New York, and noting a growing set of recommendations to reduce social contacts, all of which happened before mrNPIs. This may also explain the highly variable effect sizes of the same NPI in different countries. For example the effects of international travel bans were positive (unhelpful) in Germany and negative (beneficial) in the Netherlands (Figure 2).
Yes that makes common sense to me too. Degree of compliance I expect to decrease as measures get more strict and start to damage people’s livelihood. That may be what we are seeing in California. Strict measures are forcing large numbers of people to move out of “paradises” like California and New York to Texas, Florida, Tennessee, etc. For New York the permanent loses could be large. This can become a death spiral where declining tax revenue and growing expenses force ever more confiscatory taxation.
Matthew,
As pointed out above, the paper equates Sweden and South Korea in their response (!)
As with much of Ioannidis’ work on COVID, its interest lies more in why it is so eagerly shared outside the expert community than its actual content.
Should you be interested in further analysis of the content, you could do worse than go here.
https://mobile.twitter.com/GidMK/status/1349164532627693570
verytallguy: As pointed out above, the paper equates Sweden and South Korea in their response
“Equate” is the wrong word here. There are no perfect comparisons, but those are nations that did not enforce strict lockdowns. Obviously the paper should not be “believed” before there are other supporting publications, but is there any evidence that the stricter lockdowns improve upon the benefits of the voluntary social distancing and less strict lockdowns?
I do hope that the authors follow up this work with parallel analyses of data up through, say, Dec 2020.
“There are no perfect comparisons, but those are nations that did not enforce strict lockdowns. Obviously the paper should not be “believed” before there are other supporting publications, but is there any evidence that the stricter lockdowns improve upon the benefits of the voluntary social distancing and less strict lockdowns?”
Any analysis that treats Korea and Sweden as similar in response is wrong.
South Korea has had an intensive test and trace capability, and a citizenry willing to participate, to the extent that their cell phone mobile data is available in detail. I have not heard the same about Sweden.
Also, South Korea had strong measures when outbreaks happened. Sweden had weaker measures.
verytallguy: https://mobile.twitter.com/GidMK/status/1349164532627693570
Thank you for the link.
Matthew,
Reply intended here is at the foot of the thread.
The problem here meso, is a deep confusion about statistics and the scientific method. Statistics is used to draw conclusions about general categories of things by randomly sampling that category. Thus you want to get as many dissimilar samples as you can and make them as random as possible.
Absolutely nothing wrong with using as many countries as possible in a study of whether interventions work.
“The problem here meso, is a deep confusion about statistics and the scientific method. Statistics is used to draw conclusions about general categories of things by randomly sampling that category. Thus you want to get as many dissimilar samples as you can and make them as random as possible.”
What does that have to do with my comment? If they treated the interventions in SK and Sweden as similar, they made a big mistake. If they didn’t, then the characterization I was responding to was incorrect. I pointed out the substantial differences between the two country.
Ioannidis et al absolutely did NOT treat them as “similar” what ever that means. They are two different data points in a dataset that is analyzed using statistics.
VTG’s criticism is based on a fundamental confusion about statistics and the scientific method. Generally, you want as large a dataset as possible and one randomly includes a wide variety of cases.
Further –
Getting the vaccines out for COVID is an enormous and unprecedented task that was inevitably going to encounter massive obstacles. I see no particular reason why Pub politicians should earn more credit as group for getting vaccines out that Demz, but if after controlling for important variables it looks like they did to better I’ll happily give them credit for that.
That said, it’s hard for me to believe we wouldn’t be able to do much better and save many lives and significantly limit economic damage had (1) we not starved our public health systems for decades and (2) Trump actually made vaccine distribution the highest of national priorities.
Hopefully Biden will use National Defense Authorization. What should it be used for if not to save American lives?
Joshua: Hopefully Biden will use National Defense Authorization.
Boy I hope not. The Constitution gives most power and authority for health and safety issues to the states, a feature of our governance respected by Trump.
OTOH: https://www.wsj.com/articles/covid-19-vaccine-leaders-waited-months-to-approve-distribution-plans-11610737935
There’s also nothing mutually exclusive about using the national defense act to address logistical issues and economic issues, and states having most authority and power over health and safety issues.
Good info on why some states performing better than others and sure, state-level autonomy looks important:
https://www.cnn.com/2021/01/15/health/states-vaccinating-faster-west-virginia/index.html
>Biden’s plan, while relying in part on federal resources, ultimately gives more power to states to accelerate their vaccine disbursement as quickly as they can.
“We’ll fix the problem by encouraging states to allow more people to get vaccinated beyond health care workers and move through these groups as quickly as states think they can. That includes anyone 65 and older,” Biden said Friday.
https://www.nbcnews.com/politics/white-house/biden-speak-covid-vaccinations-security-concerns-delay-inaugural-rehearsal-n1254381
Matthew:
This is hilarious
> The Constitution gives most power and authority for health and safety issues to the states, a feature of our governance respected by Trump.
So under Trump its a good sign that he respects the Constitution. But under Biden it would be tyranny. Really, you boyz are hilarious.
>> To help Pfizer, the deal calls for the government to invoke the Defense Production Act to give the company better access to roughly nine specialized products it needs to make the vaccine. One person familiar with the list said it included lipids, the oily molecules in which the genetic material that is used in both the Moderna and Pfizer vaccines is encased.
https://www.nytimes.com/2020/12/22/us/politics/pfizer-vaccine-doses.html
>>> Under the leadership of President Trump, the Department of Health and Human Services leveraged the Defense Production Act (DPA) to apply priority rated orders for contracts with Becton Dickinson (BD) and Quidel Corporation through September
https://www.hhs.gov/about/news/2020/08/20/trump-administration-uses-defense-production-act-to-aid-our-most-vulnerable.html
>>>> In the early days of the COVID-19 pandemic, the Trump administration was criticized for not using the Defense Production Act quickly enough or in sufficient quantity to counter the spread of coronavirus. Now, after using the act 33 times since March 18, there is no denying that the administration is utilizing the act to fight this public health crisis.
https://www.heritage.org/defense/commentary/administration-ramps-use-defense-production-act
Matthew:
This is hilarious
> The Constitution gives most power and authority for health and safety issues to the states, a feature of our governance respected by Trump.
So under Trump its a good sign that he respects the Constitution. But under Biden it would be tyranny. Really, you boyz are hilarious.
From the NYTimes
>> To help Pfizer, the deal calls for the government to invoke the Defense Production Act to give the company better access to roughly nine specialized products it needs to make the vaccine. One person familiar with the list said it included lipids, the oily molecules in which the genetic material that is used in both the Moderna and Pfizer vaccines is encased.
From HHS
>>> Under the leadership of President Trump, the Department of Health and Human Services leveraged the Defense Production Act (DPA) to apply priority rated orders for contracts with Becton Dickinson (BD) and Quidel Corporation through September
From HERITAGE.
>>>> In the early days of the COVID-19 pandemic, the Trump administration was criticized for not using the Defense Production Act quickly enough or in sufficient quantity to counter the spread of coronavirus. Now, after using the act 33 times since March 18, there is no denying that the administration is utilizing the act to fight this public health crisis.
From Heritage – used it 33 times since March 18:
https://www.heritage.org/defense/commentary/administration-ramps-use-defense-production-act
Joshua: From Heritage – used it 33 times since March 18:
I stand corrected. Good comments
Matthew –
Good response. So, what’s the conclusion – that Trump is a tyrant or that there’s good reason to use the Defense Protection Act?
There is, of course, risk involved. But then again, there’s no free lunch.
Joshua: So, what’s the conclusion – that Trump is a tyrant or that there’s good reason to use the Defense Protection Act?
I had been thinking of some of the harsher proposals that I had read, such as imposing and enforcing uniform lockdown standards across all the states, or enforcing uniform distribution standards for the vaccine. But his own speaking emphasized recognizing the authority of the states.
Trump was not a tyrant, and the DPA can be used well or badly.
And besides, if you just pass it off to the states and some of them screw it up, it makes great partisan fodder. What’s tens of thousands of lives if you can score some cheap political points?
Not merely the intention. A more effective rollout of vaccines would no doubt save many lives and reduce massive damage to the economy.
There are some risks involved – but there’s no free lunch. The question is whether it’s less sub-optimal than the alternative.
Joshua: God forbid he used it to save lives.
Surely the announced intentions will be good.
Joshua: >Biden’s plan, while relying in part on federal resources, ultimately gives more power to states to accelerate their vaccine disbursement as quickly as they can.
That’s yet to be tested, but hope springs eternal.
South Korea and Sweden had approaches almost as different as it’s possible to be.
To treat them as equivalent in a comparison is positively perverse.
As far as the benefits and costs of more compulsory and voluntary measures is concerned I think it’s very hard to compare between cultures in different parts of the world.
And until we are genuinely out of the pandemic we can’t know who actually had the best approaches.
But so far, Sweden has done much worse, actually both in terms of more restrictions *and* more deaths than neighbours who locked down hard to start. South Korea has done far better than any Western country, again in both respects.
VTG –
Re the latest Ioannidis paper (watch David duck).
There’s an unsupported implicit assumption about direction of causality in all of this. There is a distinct possibility that the most stringent interventions take place where precipitating conditions are the worst. You obviously NEED to control for that to evaluate efficacy of interventions.
https://twitter.com/GidMK/status/1349164532627693570
Oops –
Didn’t see that you linked previously.
Now we can watch David duck in two sub-threads!
Joshua,
I read the paper before I saw the critique; it’s doing the rounds of the usual suspects.
Putting SK and Sweden together is just risible.
VTG –
But he’s a lion.
This is all word salads with no substantive criticisms. The tweet just quotes from the paper as far as I can see. You of course can find support for any wrong headed point on Twitter.
Just as an example claiming that Korea and Sweden are treated as equivalent is wrong and simple minded. They are 2 instances in a statistical analysis and not equivalent.
Just one substantive criticism is that there is a fundamental flaw in the analytical concept.
They don’t control for the distinct possibility that the strongest interventions were implemented in the countries where the pandemic was the most virulent. Thus, there would be reasons to expect the outcomes to be worse in those countries independently of the degree of efficacy of those interventions. A lack of control for direction of cauality should explicitly be elaborated upon as a limitation. Was it?
The problem here is that VTG’s criticism and indeed yours Josh shows a deep ignorance of statistics and the scientific method.
To show that “statins improve lipid profiles” you do a random sample with as many different people as possible and do a double blind experiment. Limiting your sample to people who are for example over 70 will lower the statistical power of your experiment and also make your conclusions much less useful in clinical practice. Further, there is a massive cost to doing a large number of studies of different “types” of people. There are an infinite number of possible “types” of people.
What Ioannidis and company did is perfectly normal rigorous science.
Not controlling for, or even discussing the limitations of not controlling for, direction of cauality is not normal and rigorous epidemiological analysis to determine the efficacy of interventions.
This last comment takes the prize for drivel Josh. I doubt you have even read the paper. Normal science looks for associations. These then have implications for causation.
“Implementing any NPIs was associated with significant reductions in case growth in 9 out of 10 study countries, including South Korea and Sweden that implemented only lrNPIs (Spain had a non-significant effect). After subtracting the epidemic and lrNPI effects, we find no clear, significant beneficial effect of mrNPIs on case growth in any country. In France, e.g., the effect of mrNPIs was +7% (95CI -5%-19%) when compared with Sweden, and +13% (-12%-38%) when compared with South Korea (positive means pro-contagion). The 95% confidence intervals excluded 30% declines in all 16 comparisons and 15% declines in 11/16 comparisons.”
If the measures didn’t cause changes in the number of new cases as you seem to think should be “controlled for” (you have no idea even what that would look like), then they had no effect and the analysis is correct.
Texas became the first state to administer 1 million COVID-19 vaccine doses. Gov. Greg Abbott (R-Texas) announced the milestone on Thursday, saying the Lone Star State is leading the nation in its vaccine program.
Abbott said the state opened up vaccination hubs and expanded its vaccine eligibility.
“As you know, we’ve now opened up an additional category and that is anybody over the age of 65 or older, or people under the age of 65 who face or deal with certain chronic healthcare conditions that have been pre-identified,” Abbott announced.
https://www.oann.com/texas-becomes-first-state-to-administer-1m-vaccine-doses/
an ivermectin story:
https://buffalonews.com/news/local/after-judge-orders-hospital-to-use-experimental-covid-19-treatment-woman-recovers/article_a9eb315c-5694-11eb-aac5-53b541448755.html
In Italy 30,000 restaurants opened on the same day in defiance of covid lockdown rules which is destroying their businesses:
https://youtu.be/7YZGXNwri0o
dpy,
So, when cases reduced in November, was that herd immunity threshold again, or something else?
I think we’re done here.
dpy,
So, when cases reduced in November, was that herd immunity threshold again, or something else?
I think we’re done here.
See my summary below.
I’ve been taking an interest in the UK data since December, before I was aware of Nics ideas on this, plotting my own graphs for some locations around the UK. I’m no expert, just a beginner.
It feels like there is potentially a lot of very interesting, very useful information in there that needs to be thoroughly and urgently looked at. But it seems that every time I think I might see some conclusions newer data seems to throw that out of the window, and the vaccine is probably now going to further confuse any analysis or conclusions. I’m probably doing it all wrong. Presumably the modelers etc are doing all this anyway?
I thought just before Christmas I was seeing the worst 2 or 3 places in south east England starting to fall due to a kind of herd immunity in the non adequately shielded group (because they appeared out of control despite lockdown) at around 11 to 13% cumulative on my adjusted numbers (see below, which are probably wrong and different from anyone elses, but maybe better than no adjustment), but then they took off again. Then I thought I was seeing them peaking due a kind of herd immunity at around 16 to 18% cumulative on my adjusted numbers. But they all peaked on 30th or 31st December which seems rather strange to me, also considering some other places are now also showing they peaked on those same dates, but much lower % and other UK places are still rising.
Of the locations I’ve plotted, I think 4 have so far peaked at 16-18%, 3 at 12-13%, one at 7%, my numbers. Maybe some places are peaking due to a kind of herd immunity in the non adequately shielded group(?) and the value depending on whether more or less Christmas mixing occurred(?) in other places lockdowns are now effective? Did Christmas shopping etc and 25th cause a lot of mixing then immediately everyone was much more careful, plus being at home, not at work, etc so lockdown was finally effective(?), but wouldn’t we expect to see the result more like 10 days later not 5? On other hand my data could be out plus or minus several days. And wouldn’t we expect a lot more places to peak about that same date if it’s the lockdown?
I think Liverpool is particularly interesting. I am perhaps wrongly assuming it has now peaked but it is showing a slight drop both in daily cases and the % of positive results for a few days now. Below are the 4 that I wonder if they have reached a sort of herd immunity; first number is cumulative % (my badly adjusted numbers) just before entering the November/December wave, 2nd is % at peak, then date of peak (UK format) –
Swale (Kent, south east) – 3.6%, 15.7%, 31/12/20
Medway (Kent) – 5.1%, 17.8%, 30/12/20
Dartford (Kent) – 8.1%, 17.8%, 30/12/20
Liverpool (North west) – 13.6%, 17.1%, 6/1/21 tbc
I think Liverpool was in the news a lot in October and Swale in the news a lot in December as one of the first badly affected by the new variant and out of control. No doubt Liverpool will prove me wrong in a few days by not peaking yet.
Rather than Nics lambda, or heterogeneity I’m just splitting the local population into a group A and B in my mind. I’m assuming group A are those not adequately shielded and assuming they represent most of the infections and are x% of the population and x changes value with lockdowns and behaviour, maybe e.g. lockdown1 x= 0.3 (30%), 0.8 over the summer and 0.5 now? If z is the % of A who need to have been infected to bring their R down to 1, (and changes with season) then I’m simplistically assuming the number of daily cases should start dropping when C = zx. Where C is cumulative infections as % of local population. So e.g. if x is 0.5 (50%) and z 0.4 I might see a peak in daily cases when cumulative value is 20%, say?
My next thought was how do you adjust the data for cases to get to an actual value. I used a very simple, probably wrong method taking into account recorded cases, number of tests and percent of tests positive. I’m sure my simple equation is way off and needs a better method, it was just a quick 1st attempt to see if there was potentially anything interesting in the graphs. After adjusting, the first wave has gone up massively, I’m concerned its too much, but 1st and current waves now look better than the raw values in relation to the hospital admissions.
Paul: The number of cumulative cases detected by testing is a fraction of those becoming immune, because testing missed asymptomatic cases and mildly-symptomatic cases that don’t get tested. Surveys in the spring suggested that 6-15 people became immune for every infection that was detected by PCR. That ratio appears absurd today. Given what we know, 1 missed infection per detection infection seems like a reasonable, but “soft”, lower limit. The absence of unambiguous signs of a slowdown in due to approaching herd immune suggests to me that 2 missed cases for every detected case is reasonable upper limit.
Around Thanksgiving, North and South Dakota were the hardest hit US states, with cumulative cases (essentially all in the fall) rising above 10%. ND is 12.6% today, but has the fewest current cases in the US (20/100,000). Some combination of restrictions, behavior changed by fear (the hospital are overflowing) and POSSIBLY approaching herd immunity have caused a dramatic reversal. In Burleigh County with the capitol of Bismarck, cumulative cases are nearly 15%. However, the conclusion that approaching herd immunity MUST BE playing an important role is suspect.
Arizona is currently the worst state (119 cases/100,000/day, but only 9.2% cumulative testing positive). Two of its counties are proof that the pandemic can rage (134 cases/100,000/day) with 15+% cumulative cases. Things are bad along the Rio Grande where the pandemic hit hard in early summer: Webb county, TX: 12.4% cumulative cases, 203/100,000/day! Los Angeles and San Bernardino counties: 10.0% and 11.4%, 139 and 157 cases/100,000/day. These places are approaching herd immunity at a rate of about 1% of population testing positive per week.
Many thanks Franktoo, I think your numbers give some support to my adjustment results with my first wave going up massively and recent results being about twice the raw values. Also my daily numbers, adjusted seemed roughly in line with ONS (office for national statistics) estimates of 1 in 30 being currently infected a few weeks ago if I assume each person is infected for 7 to 10 days. And perhaps also not far out for cumulative values over the summer when I think I heard they thought about 1 in 20 had so far been infected.
Although I suspect my equation (adjusted cases from recorded cases, number of tests, % of tests positive) is going way out at the extremes; considerably over adjusting when testing capacity very low and % of positive results very high and under adjusting when testing very high and % positive very low. Ideally I’d rewrite my equation and more accurately check/calibrate against estimates from other sources/methods on how many people have been infected etc.
I was thinking along the lines of there should be some statistical/mathematical method to adjust the recorded cases taking into account also number of tests and % of tests positive. I’m thinking this is important to do so we can hopefully get useful and important information out of the data: Enable better comparison of old and recent waves. But even within waves I suspect changes in number of tests gives false trends, false peaks and troughs and with suitable adjustment of each days cases based on total recorded cases, number of tests, % of tests positive we can better study actual size of peaks and their dates?
I was hoping my method might be useful in looking at whether a kind of herd immunity is forming in some places, effect of lockdowns, if immunity lasts from first wave, if that immunity works for the new variant, if the new variant is more transmissible.
The first place I started watching was Swale, Kent as it was in the news as the highest numbers in the country for a long time and probably the first badly affected by the new variant. Numbers kept going up despite lockdown early November and it has been in lockdown ever since. It’s numbers stayed high and slowly rose for a long time, got to about 950/100k/week (unadjusted). My feeling at the time was it was out of control, government or local council not taking it seriously enough and hospitals coping because they could send people out to other areas which were yet to be affected. So Swale first sparked my idea that unless lockdown was considerably tightened up the only way it would stop would be when the non adequately isolated group reached a sort of herd immunity.
Of the places I’ve plotted so far, Medway, next door to Swale was initially behind then overtook and got to 1168/100k/week, Brentwood in Essex got to 1562 (223/100k/day, I presume), unadjusted. So it felt like some of these places, looking out of control must only be peaking due to a kind of herd immunity as I couldn’t see enough in lockdown2 then tier 3 then 4 to bring it under control. But I have low confidence in that conclusion now looking at other places: Just been looking at a few more UK locations and so many places seem to have peaked on the same two days, 30 and 31 December and at lower cumulative % implying a lockdown effect? Or is it because the lower ones have managed to so far avoid the new variant?
Just to recap since this “interaction” is replicated many many times on the internet including many of times here with the same suspects. There are two of these types who have a massive comment footprint on this post.
Anonymous nonscientist sees something by a real scientist that is an interesting contribution but might harm his political narrative. Activist joins conversation.
1. He quibbles about definitions, provides no evidence, and denies the citations of the scientist from other scientists.
2. Since the activist is incompetent to do real science or math he is reduced to proof texting various pseudo-scientific sources such as that leading scientific journal, Twitter.
3. He then goes on the attack a lion of the medical field with vague falsehoods and other rhetorical devices to cast doubt on an excellent body of work.
4. Then he shows his utter ignorance of statistics by stating that doing a statistical analysis involving data from many different counties means the authors are saying they are equivalent.
5. Activist ignores real scientific analysis of HIT, R0, R and whether these are useful concepts or not.
6. Activist slinks away to read the Guardian and other top flight scientific journals.
The sad part is that these activists are terrible merchants of doubt. Some people I guess don’t have more interesting things to pursue.
The pandemic is winning:
Biden goal: 100 million vaccinations (50 M people) in 3 months. That’s 0.55 million/day. Currently 0.24M/day test positive. We probably miss at least 1 and possibly 2 asymptomatic or mildly-symptomatic cases for every case we detect.
Conclusion: The pandemic is currently immunizing Americans as fast or faster than the vaccine.
What is the value in Operation Warp Speed when it is going to take longer to simply vaccinate Americans than it took to develop a vaccine!
Very good point. Many realists have been saying that herd immunity was our destination and anything beyond mild mitigation would not change the outcome much and be devastating in its human toll.
The Humana CEO says covid 19 will be with humanity for a long time.
Reports out of Norway indicate that quite a few older folks have died from complications after vaccination.
Sometimes tragedy strikes and there is little we can do. The misleading idea that science can conquer anything (often propagated by scientists themselves) is quite wrong. The political narrative that if we followed the science we would have prevented hundreds of thousands of deaths is worse. It’s about as valid as burning “witches.”
“Many realists have been saying that herd immunity was our destination and anything beyond mild mitigation would not change the outcome much and be devastating in its human toll.”
Okay, that is just plain nuts!
First, we have vaccines. Those vaccines will save a lot of lives over, at this point, trying to achieve herd immunity through vaccination.
Second, assume we did not have vaccines. We would have three choices:
1) Let the infection run its course. Let’s assume that everyone just goes on and lives life as normal – which is which most herd immunity through infection advocates want. The rate of infection would rapidly get beyond the ability of the medical system to handle. The infection fatality rate would soar, plus even those not infected would die because of the destruction of the health care system (loss of capacity, and death and disability of providers). This happened in Tijuana, for example, and in other places. It would be catastrophic.
But, it wouldn’t work that way. A lot of people would cut back their interactions with others as much as possible. They would stop traveling, stop eating in restaurants, stop going into stores, and wear masks. The result would be pretty much the same economic wreckage we see with lockdowns. Except, this wouldn’t be enough to keep the health care system working, so we’d still have a catastrophic failure, but at a lower level.
This would continue until either stronger measures were used, or vaccines became available. And it would be a disaster.
If you disagree, please respond to my points.
2)Control the infection the way the Chinese and Taiwanese have, while waiting for better medical treatment, and beyond. It works, it does not require a vaccine, but it does require extremely good surveillance, and it requires *short term* lockdowns to stamp out outbreaks. It also has major impact on travel. It does not require totalitarianism – Taiwan is a democracy.
3) Hold the course with mitigations until vaccines provide herd immunity.
Please explain why, at this point, that point 3 is wrong.
“The Humana CEO says covid 19 will be with humanity for a long time.”
Your point?
“Reports out of Norway indicate that quite a few older folks have died from complications after vaccination.”
Evidence. We knew all along that people would die after getting vaccinated, because people die all the time. Some people, a small number in the US so far, have dangerous reactions. None so far have died. A few might over time. So what?
“Sometimes tragedy strikes and there is little we can do. The misleading idea that science can conquer anything (often propagated by scientists themselves) is quite wrong.”
That’s a complete straw man argument and has nothing to do with this debate.
“The political narrative that if we followed the science we would have prevented hundreds of thousands of deaths is worse. It’s about as valid as burning “witches.”
What BS!
Your evidence on vaccinations is out of date. Norway has attributed as I recall 7 deaths to vaccine reactions and is investigating 24. I will get vaccinated but it’s not risk free.
There are by now a host of papers on lockdowns. One I cited earlier finds no statistically significant effect of harsh measures. At best there is not a conclusive consensus.
You really didn’t respond to much else. You believe that we can suppress an epidemic with strong measures. As California, New York, Florida, Texas are very convincing evidence of the futility of strong lockdowns. They also show how strong measures can kill your economy. People are fleeing New York and California in droves. Businesses are leaving. Tesla is a big one.
You cherry pick two countries as evidence. That’s not convincing. Virtually every European country is experiencing a strong second wave regardless of wide ranging levels of mitigations and cultures. Ditto in the US.
Even Cuomo says they must open soon or there will be nothing to reopen.
Meso’s point #2
“2)Control the infection the way the Chinese and Taiwanese have, while waiting for better medical treatment, and beyond. It works, it does not require a vaccine, but it does require extremely good surveillance, and it requires *short term* lockdowns to stamp out outbreaks. It also has major impact on travel. It does not require totalitarianism – Taiwan is a democracy.”
The problem at this point, is that the virus has become to deeply embedded into the general population to make that approach even remotely feasible. The only time that approach could have been implemented effectively was in Jan/Feb which the political powers on both sides of the Aisle opposed.
“The problem at this point, is that the virus has become to deeply embedded into the general population to make that approach even remotely feasible. The only time that approach could have been implemented effectively was in Jan/Feb which the political powers on both sides of the Aisle opposed.”
The Wuhan experience demonstrates that this is false. The virus was “deeply embedded” there, too. They whipped it.
meso – china whipped as you say because they had an extremely hard lockdown – they have a repressive country which allowed for an extremely hard lockdown. The other countries that did it taiwan , SK were able to do it before it became embedded.
That type lockdown that china imposed would never have worked in the US
“That type lockdown that china imposed would never have worked in the US”
Sadly, you are probably right – to our detriment. Their economy is far more open than ours and has been since they beat the virus down. They get outbreaks – there’s one right now near Beijing, but they beat them down too.
The relatives and friends of the nearly 400,000 dead Americans probably wish we did have that lockdown. Done right, which is, as you say, harsh, the lockdown only lasts one to two months. After that, you have your life back with some exceptions: mask wear required, some social distancing, etc.
But some free countries have done far better than the US. Taiwan, of course, did very well, but that’s because they caught it very early, and were ready because of SARS.
South Korea has done pretty well too, although they have had bigger outbreaks, as has Japan. Both were, I think, more prepared because of SARS and MERS (which hit South Korea). These are free countries.
http://91-divoc.com/pages/covid-visualization/?chart=countries-normalized&highlight=South%20Korea&show=highlight-only&y=both&scale=linear&data=cases-daily-7&data-source=jhu&xaxis=right&extra=Japan#countries-normalized
Also bear in mind that China’s data may be unreliable. Their press is constantly pushing propaganda.
“Also bear in mind that China’s data may be unreliable. Their press is constantly pushing propaganda.”
China’s data is unreliable, but not so unreliable as to undermine the fact that they have done far better than the US, and are continuing to do far better.
They have propaganda and are totalitarian, but a lot more information gets out of there than from the old USSR, or from China in Mao’s day.
https://virologyj.biomedcentral.com/articles/10.1186/1743-422X-5-29
I think epidemiology I’d too immature for us to know if mitigation measures work. In any case the empirical evidence is that they don’t. We can’t get people to quit smoking and lose weight. Compliance in a free society is always going to be spotty
dpy
I think epidemiology I’d too immature for us to know if mitigation measures work. In any case the empirical evidence is that they don’t.
You see to be arguing that social distancing doesn’t mitigate flu outbreaks.
Please tell me that I’ve misunderstood, otherwise that’s seriously insane, given what’s happened to flu prevalence through the COVID outbreak.
https://www.who.int/teams/global-influenza-programme/surveillance-and-monitoring/influenza-updates/current-influenza-update
We don’t know the cause of flu ‘disappearing.’ We know very little about flu. Yet you show up here again with no science and spread falsehoods about science you are ignorant of.
https://virologyj.biomedcentral.com/articles/10.1186/1743-422X-5-29
dpy
We don’t know the cause of flu ‘disappearing.’
Wow. Just wow.
We institute worldwide social distancing policies to prevent the spread of an infectious respiratory disease.
Influenza disappears at the same time, in a way which has never been seen before.
Yet we don’t know the two are causally related.
First of all herd immunity doesn’t require much, if any actual immunity, now this.
Alright, I’m done. You’re way beyond any kind of rational discussion here.
I understand VTG who you don’t want to further embarrass yourself and expose your ignorance. Association does not mean causation.
I’ve seen many possibilities discussed. Testing for flu may have gone to zero as people focus on covid 19 is one example.
Your response to a link I posted to a peer reviewed paper on the flu and our deep ignorance is? Reading real science I guess takes you away from that leading scientific journal Twitter.
For Josh and VTG, our resident nonscientists, I have been doing a little research on lockdowns and their effects. There are a score of papers all saying pretty much the same thing. There is no effect attributable to lockdowns particularly in Europe during the first wave.
I also found a devastating debunking of Flaxman’s paper on lockdown in England. Flaxman was one of the “experts” Gelman cited in his blog post on Ioannidis’ latest study of NPI. After reading the critique of Flaxman, I would give him no credence on this subject. Josh why can’t you spend even a little time on real research using Google Scholar?
https://www.medrxiv.org/content/10.1101/2020.04.24.20078717v1
Looks like Flaxman’s simplistic model has come in for a lot of criticism in the literature.
https://www.medrxiv.org/content/10.1101/2020.09.26.20202267v1
And Nic has joined the chorus with a devastating critique of Flaxman’s paper.
https://www.nicholaslewis.org/did-lockdowns-really-save-3-million-covid-19-deaths-as-flaxman-et-al-claim/
These look to me to be pretty serious errors by Flaxman and indeed pretty obvious errors.
And one more discussing excess deaths from covid vs. excess deaths from something else. It seems in March and April in the US, 50% of excess deaths were attributed to something other than covid.
https://jamanetwork.com/journals/jama/fullarticle/2768086
One more especially for VTG and Josh. How many papers with the same conclusion will it take to convince them of anything?
https://www.medrxiv.org/content/10.1101/2020.12.28.20248936v1
” The data suggest that efficient infection surveillance and voluntary compliance make full lockdowns unnecessary at least in some circumstances.”
dpy
Re: “We don’t know the cause of flu ‘disappearing.’”
No, this is basic biology / epidemiology. You don’t know this because it isn’t your field and you haven’t done the reading on it. Those of us who do know this field understand what’s going on.
Seasonal influenza has a lower basic reproduction number (a.k.a. R0) than SARS-CoV-2, which means seasonal influenza is less contagious at baseline; i.e. under the types of conditions we would have had at the same time of year in 2019, in the absence of additional public health interventions + behavior changes beyond that:
https://www.thelancet.com/journals/laninf/article/PIIS1473-3099(20)30484-9/fulltext
https://link.springer.com/article/10.1186/1471-2334-14-480
https://journalofinfection.com/article/S0163-4453(20)30154-7/fulltext
Both SARS-CoV-2 and seasonal influenza are respiratory viruses with similar routes of transmission. So the same behavior changes and public health interventions that mitigate the spread of SARS-CoV-2, also do the same for seasonal influenza. With additional behavior changes and public health interventions, R0 gets forced downwards to Re, a.k.a. the effective reproduction number. Since seasonal influenza’s R0 was already lower than SARS-CoV-2’s, influenza’s Re gets pushed below 1 before SARS-CoV-2’s Re does. That means influenza spends less time as an accelerating outbreak, causing it to infect less people overall. We’ve known this for awhile now.
So verytallguy was right, dpy. Try actually listening to those who know more than you, for once.
“How coronavirus lockdowns stopped flu in its tracks”
https://www.nature.com/articles/d41586-020-01538-8
“COVID-19 measures also suppress flu—for now”
https://www.sciencemag.org/news/2021/01/covid-19-measures-also-suppress-flu-now
“Nonpharmaceutical public health measures can also be effective in reducing transmission, allowing suppression or mitigation of influenza epidemics and pandemics.”
http://perspectivesinmedicine.cshlp.org/content/early/2020/09/28/cshperspect.a038356.full.pdf+html
https://twitter.com/AtomsksSanakan/status/1332738821998456841
A question on notation:L I notice you used Re for the effective reproduction number. I’ve seen others use Rt for the same thing, and others use Re. I assume the “t” is for the time varying nature, the “e” for effective as you say, but is it correct that they are the same meaning, just different conventions for some reason? Presumably “0” is for time zero, or initial conditions.
More deception here. The Nature piece also talks about other causes like people not showing up for treatment and testing. There is no real science there. It’s all speculation. That’s because we know very little for sure about the flu.
https://virologyj.biomedcentral.com/articles/10.1186/1743-422X-5-29
In my state (Virginia) alone 5.3 million people with “flu-like” symptoms got tested for Covid and, so far, 4.8 million of them had something other than Covid. Something unidentified.
Why are we assuming here that they didn’t have the flu?
Even if you assume only 10% of those testing negative actually had flu-like symptoms, Virginia alone identified 480,000 flu cases in 2020. If fairly normal, that would translate into 20 million cases of the flu in the US in 2020. Which is not “disappeared.”
> Why are we assuming here that they didn’t have the flu?
Assuming your numbers are right (previous history suggests they are suspect), do you have any idea how many of them were subsequently tested for the flu?
Jeff,
Virginia is not the world, and your speculation is not informative.
Most countries run flu surveillance programs.. The WHO runs a programme to coordinate.
Here’s (again) the latest from WHO
https://www.who.int/influenza/surveillance_monitoring/updates/latest_update_GIP_surveillance/en/
And CDC.
https://www.cdc.gov/flu/weekly/index.htm
If you are actually interested in Virginia, I daresay you can find that too.
From a nurse practioner managing a COVID hotline in Boston, that receives hundreds (thousands?) of calls a day.
-snip-
We have capability of testing simultaneously for Covid, flu and strep in any combination so for people with classic flu symptoms we usually test for flu and Covid, e.g for someone with headache and diarrhea would test only for Covid but someone with HA, fever, body aches, chills, diarrhea would test for both. If severe sore throat and swollen glands would also test for strep. Also depends on if they have had a flu shot this season.
“Assuming your numbers are right (previous history suggests they are suspect), do you have any idea how many of them were subsequently tested for the flu?”
The pot calls the kettle. But here are the numbers.
https://www.vdh.virginia.gov/coronavirus/coronavirus/covid-19-in-virginia-testing/
Who would have tested for flu? In my area, the GP offices were closed for a couple months and then wouldn’t allow you in the building if you had any “flu-like symptoms.” You could go to an emergency center, where they would test you for Covid (and only for Covid) outside and tell you to go home and self isolate until the results came in. If positive they told you to monitor your oxygen saturation, if not they told you to drink lots of fluids. When testing became even more prevalent, you could get a COVID test if you were sick (and only if you were sick according the rules, which most ignored because they wanted a test). I received one of these tests when I really did have something going on. They told me I tested negative for Covid and thats it, no other test performed (and why would they, they’re slammed in those labs). So… what did I have? Any other year we’d all say: “the flu.”
I’ve been diagnosed with “the flu” dozens of times and never tested for it.
The WHO monitors respiratory disease admissions.
Finish the following sentence: Millions of people in the United States reported coming down with something that we typically call “the flu” and subsequently tested negative for COVID, meaning that they most likely had….”
Jeffn: “…the flu.”
Joshua and VTG: “…nothing at all, they must have been faking because lockdowns completely, if temporarily, eradicated all contagious respiratory viruses except for the one they were instituted to prevent. Oh, and we doubt anyone actually tested negative for Covid, show us the numbers, or maybe that only happened in Virginia.”
You guys are off base here. A coworker of mine told me about talking with an acquaintance who works on this area for the CDC. Flu testing is always very limited and they use models based on very limited data to generate the numbers of flu “cases.” These numbers are very uncertain. You have no expertise to judge how accurate they are and you have done little research on it either. Just repeating speculation on the internet.
https://wwwnc.cdc.gov/eid/article/26/5/19-0994_article
finds:”Although mechanistic studies support the potential effect of hand hygiene or face masks, evidence from 14 randomized controlled trials of these measures did not support a substantial effect on transmission of laboratory-confirmed influenza. We similarly found limited evidence on the effectiveness of improved hygiene and environmental cleaning.”
rather casting doubt on VTG’s layman’s opinion.
Jeff –
I spoke to a NP who deals with hundreds of people calling about covid in a week. They test a lot for the flu.
Speaking of week, your weak speculation and rhetoric questions are worthless.
Joshua,
I got a test because I had flu-like symptoms. They tested for Covid and didn’t bother to test for anything else. They’re busy.
One more time, what do you think all those sick people who test negative for Covid have? If your politics weren’t tied up in it, you’d say “flu.”
At the high end, there are 1.2 million flu tests in the entire nation. https://www.pharmacytimes.com/resource-centers/flu/cdc-20182019-flu-season-the-longest-in-a-decade#:~:text=During%20the%202018%2D2019%20influenza,and%205.0%25%20for%20influenza%20B.
That compares to 299 million Covid tests performed in the US so far. Note that one of your political claims is that COVID testing in the US was awful at 299 million, but now you say 1.2 million is “all the time.”
They do not test for flu routinely during any year. 1 million tests is nothing.
Nobody is seeking flu tests during this pandemic – nobody who believes in lock downs is going to any location that tests so they can hang out with Covid patients to find out if they have influenza A. If they want to be tested, they seek a covid test and that’s what they get. That’s all they get.
Doctors offices and pharmacies are not testing for flu- they are telling people not to enter their building if they have flu like symptoms and they’re offering drive-thru testing- only for COVID and, allegedly, only if you are sick. I got my test at a CVS, drive through only, special lane only, appointed time only. CVS did not test me for anything other than the flu.
Hospitals may be doing some influenza tests (probably on people who also have COVID) but way less than before. I got bitten by either a snake or spider in August, my GP and I hesitated to send me to the ER even though it was bad, eventually deciding to do so. GPs and everyone else were telling people to stay far away from hospitals unless they absolutely had to go. Nobody with the sniffles was going to the ER. The one I went to had a triage tent outside for “sick” that was sending people home unless their ox levels were bad.
Jeff –
I talked to someone who actually knows what she’s talking about.
I suggest you try doing the same.
@dpy
Re:“You guys are off base here. A coworker of mine told me about talking with an acquaintance who works on this area for the CDC.”
[s]
That sounds like an incredibly reliable chain of evidence. Maybe we should also play 6 degrees of Kevin Bacon or telephone (‘Chinese whispers’)?
https://en.wikipedia.org/wiki/Chinese_whispers
[/s]
Anyway, the CDC’s website affirms what I said about mitigation measures limiting influenza transmission. That includes seasonal influenza and pandemic influenza strains. Relevant mitigation measures include stuff we’ve done more of this flu season than in previous flu seasons, such as:
– school closures / less in-person schooling
– mask-wearing
– social distancing
– postponing / cancelling non-essential public gatherings
etc.
“Nonpharmaceutical interventions (NPIs) are readily available actions and response measures people and communities can take to help slow the spread of respiratory illnesses like influenza. NPIs that should be practiced by all people at all times are particularly important during a pandemic. They are called “everyday preventive actions” and include staying home when sick; covering coughs and sneezes; frequent handwashing; and routine cleaning of frequently touched surfaces and objects. Community-level NPIs may be added during pandemics to help reduce social contacts between people in schools, workplaces, and other community settings (e.g., dismissing schools temporarily, providing telework options, and postponing large gatherings).”
https://www.cdc.gov/flu/pandemic-resources/planning-preparedness/community-mitigation.html
“Interim Pre-pandemic Planning Guidance: Community Strategy for Pandemic Influenza Mitigation in the United States
[…]
All these interventions should be used in combination with other infection control measures, including hand hygiene, cough etiquette, and personal protective equipment such as face masks.”
https://pbs.twimg.com/media/EdjpM2_XoAApnub?format=png&name=small
https://web.archive.org/web/20161219230924/https://www.cdc.gov/flu/pandemic-resources/pdf/community_mitigation-sm.pdf
@mesocyclone
Re:“A question on notation:L I notice you used Re for the effective reproduction number. I’ve seen others use Rt for the same thing, and others use Re. I assume the “t” is for the time varying nature, the “e” for effective as you say, but is it correct that they are the same meaning, just different conventions for some reason? Presumably “0” is for time zero, or initial conditions.”
Yes, it’s just different conventions. Some people use Rt to refer to what R0 changes into once there’s a non-zero amount of people immune to infection (R0 assumes no one is initially immune). Others use Rt to refer to what R is once there are additional public health interventions and/or behavior changes. Same for Re. It doesn’t really matter which definition one uses, or if one combines both definitions into a single one, as long as one is clear on what one means and doesn’t conflate it with R0.
And yes, the “t” refers to it being time-varying, though technically R0 varies with time as well, based on what baseline conditions are in that region for a given time of year. That doesn’t help Lewis at all, of course, since Sweden hasn’t been at baseline conditions since at least March 2020, if not earlier, given their decreased physical interaction with those outside their home (as reflected in mobility data and surveys of the public), closing of large events, closing of high schools, restrictions on air travel, etc. They even had to make the restrictions stricter near the end of 2020, as their level of distancing got too close to baseline to keep Re below 1, contrary to what Lewis predicted. If Sweden actually had been at the baseline R0 conditions of herd immunity during the time Lewis falsely claimed they reached HIT, then they would have a crippling high number COVID-19 deaths per capita, instead of just the negligently high number they have now in comparison to their Nordic neighbors.
“Sweden on Monday announced a ban on public events of more than eight people at a press conference where ministers urged the population to “do the right thing”.
[…]
The new limit is part of the Public Order Act and therefore is a law, not a recommendation like many of Sweden’s coronavirus measures. People who violate the ban by organising larger events could face fines or even imprisonment of up to six months.
[…]
“There should not be social situations with more than eight people even if they are not formally affected by the law. This is the new norm for the whole society, for all of Sweden. Don’t go to the gym. Don’t go to the library. Don’t have dinners. Don’t have parties. Cancel,” he said.”
https://swedishchamber.nl/news/sweden-bans-public-events-of-more-than-eight-people/
https://twitter.com/DrKatrin_Rabiei/status/1339988276229206017
Another comment that misrepresents the science and deflects from the science papers I cited above.
1. The literature on mitigation measures is rather meager. I cited several papers above that call into question mitigations such as lockdowns and stay at home orders as well as masks and hand sanitization.
2. The CDC guidelines you cite do not appear to be peer reviewed or to review the more rigorous science.
3. These guidelines fall into the category of what I characterized as crude mechanistic explanations that lack rigorous scientific support or quantification.
4. The paper I linked about the flu shows conclusively that the science on flu is weak in virtually all areas.
https://virologyj.biomedcentral.com/articles/10.1186/1743-422X-5-29
If Josh read Jeff’s reference he would see that:
“Since the 2010–11 season, the CDC estimates influenza virus infection has caused 9.3 million–49 million symptomatic illnesses, 4.3 million–23 million medical visits, 140,000–960,000 hospitalizations, and 12,000–79,000 deaths during each influenza season. It is noted that preliminary estimates for the 2018-2019 season fall within these ranges.”
In short, the CDC can’t determine within a factor of 5 how many cases of flu there have been. That can only happen when testing is very spotty supporting Jeff’s point.
Not that I would EVER question Jeff’s expertise in these matters, but still, Ima gonna go with this guy:
-snip-
“There’s no question this year is an extraordinary year,” said Kaiser Permanente flu expert Dr. Randy Bergen, who told ABC7 News that by early January, California would normally see a serious uptick in flu cases, but not this year.
“We are testing still thousands of people in our emergency room settings and in our hospitals for a combination of COVID and flu tests and we’re essentially seeing no flu. Some weeks we’ll have no cases, other we’ll have maybe one or two cases.
Maybe so, maybe not Josh. Anecdotal evidence is not really very good. The simple fact of the matter is that even Sanakan’s proof texted Nature reference is purely speculation as to why with no real scientific evidence. What I said remains true. I don’t know what the cause is and neither do you or VTG. The best you can do is vague rhetorical verbal formulations.
jeffnsails850
“Finish the following sentence: Millions of people in the United States reported coming down with something that we typically call “the flu” and subsequently tested negative for COVID, meaning that they most likely had….”
Well all those millions of people you claimed tested negative for COVID-19 but had flu-like symptoms must be using home remedies because cold and flu OTC sales are down biggly.
You might provide a link to support your claim of millions of people with flu-like symptoms that also tested negative.
“Sales of cold, flu and cough medicine in the U.S. plummeted 46% in the five weeks ended Dec. 26 from the same period last year, according to data from market research firm Nielsen.”
https://www.ibj.com/articles/cold-medicine-sales-plummet-amid-pandemic
“Well all those millions of people you claimed tested negative for COVID-19….”
Really Jack, you deny anyone tested negative? Or are you claiming that every single person who endangered themselves by going to a testing place where people with COVID were gathering (hence the positive tests) were feeling great and just thought it would be fun and safe? Got any evidence for that? I mean any evidence at all?
Because I have evidence to the contrary. Testing sites here are limited to those who are sick or who have been in close contact with the sick. The CDC site says you should get tested only if you’re sick or have been in close contact with someone who is sick, or who “has traveled” which makes all those claims that “airlines are safe’ interesting.
https://www.cdc.gov/coronavirus/2019-ncov/symptoms-testing/testing.html
According to the rules (if you believe in lockdowns, you believe people follow the rules) you can’t get a test unless you’re sick. My numbers assume only 10% of those testing negative are actually sick. But, again, you believe in the effectiveness of government orders so you believe that number must be much more than 10%. Yet here you are saying it’s zero. Without evidence. And I have evidence- questionaires before testing appointments are allowed and CDC orders.
https://www.cdc.gov/coronavirus/2019-ncov/symptoms-testing/testing.html
Flu stuff sales: One more time, every single business in my state, which has “reopened,” has a sign on the door ordering people not to enter if they have any flu-like symptoms. Period. Every GP, pharmacist, hospital is telling people if they feel sick to stay at home, stay away from their family members or anyone else, to drink fluids and, if necessary, take any pain relievers you have lying around. A buddy of mine just posted on Facebook that he tested positive for Covid. They told him to stay at home, get someone to bring him no-sugar Gatorade, stand up frequently (every 2-3 hours even at night), take vitamin D3, and sleep on his stomach. No “flu remedies” on the list. One said baby aspirin to avoid clots. You will note that he is not being told to go wandering about town looking for a test to see if he also has Influenza A or B. Because everyone except lockdown activists knows better.
OK. So on the one hand we have people who are intimately involved in the testing processes “anecdotally” telling us that people are getting tested for the flu. On the other hand, we have politically committed blog commenters speculating about whether people are being tested for the flu and deciding they aren’t.
Got it.
Josh, You don’t know how many people are getting tested and neither do I. In any case, what I said on this subject is true. No one knows why the flu “disappeared.” You have not helped us make progress here at all.
dpy
“The paper I linked about the flu shows conclusively that the science on flu is weak in virtually all areas”
The COVID19 virus was first transmitted to humans in 2019. The clue is in the name.
Your 13 year old paper, rather unsurprisingly, says nothing about how the evidence on suppression of COVID 19 (nineteen, note) also suppressed flu.
I don’t expect you’ll catch up with your own advice to read more widely any time soon.
So you haven’t read it and have no intention of reading it.
Meaningful discussion is impossible with someone like you who relies on rhetorical devices and refuses to actually go to the literature.
The paper is relevant because it points out how ignorant we are about flu transmission and its epidemiology. Yet you, with no qualifications, treat as obvious a largely political narrative that mitigation is causing flu to disappear when there is essentially no scientific support. This is based on crude mechanistic explanations that have little quantification with real science.
David –
> Josh, You don’t know how many people are getting tested and neither do I.
I gave you information from people who are families with how many people are getting tested about elevant to how many people are being tested for the flu.
You can instead choose to rely on the anecdotal reporting of an blog commenter who has no idea what he’s talking about.
You’re entirely within your rights to do so.
Josh, What you are doing here is very disrespectful. I am relying on peer reviewed science. I am not backing what Jeff says here. He doesn’t know either and you don’t know. Stop misrepresenting what I say.
“…speculating about whether people are being tested for the flu…”
Stunning. Are you unable to read? I gave links to the paucity of flu testing. I provided several independent examples- including the CDC – ordering people not to get tested unless they’re sick, I gave you links to the dashboard that shows 4,800,052 people in Virginia went to get these tests and did not have Covid. I cited my personal experience getting one of these tests.
Ohhh, but your friend is in Boston. Where they test everyone for the flu. You heard it yourself. Except, once again, the CDC says 500k flu tests nationwide. And you may be surprised to discover that Massachusetts reports on their activity, just like Virginia!
https://www.mass.gov/info-details/covid-19-response-reporting#covid-19-interactive-data-dashboard-
That’s the link for more “speculation.”
Your chosen state performed 13,874,158 Covid tests. Seems like quite a discrepancy to me 500,000 v 14 million.
Oh and 13,369,594 of those tests were negative.
Maybe you think all those tests were given routinely to health care workers, not to “sick” people. We’re in luck- Mass. tracks “first time” tests- those given to people who think they have COVID. There were 4,625,449 of those tests. Even if you make the strange assumption that every single COVID case in your state was identified by a first-time test, Mass. has identified 4,120,885 people in the state who were sick with “flu-like” symptoms who did not have COVID and, as we’ve already proven, were not actually tested for the flu according to the CDC- which we all know tracks the flu carefully.
So. Let’s see if, for the first time in your life, you will answer a question posed to you repeatedly: what did those 4 million people in Massachusetts have that they thought, incorrectly, was COVID because of their “flu-like symptoms?”
David –
> I am not backing what Jeff says here.
I’m glad that you agree with me that Jeff’s anecdotal reports are worthless.
On the other hand, I don’t dismiss reporters from experts and front line healthcare workers who are making decisions about testing for flu for hundreds of people weekly to be worthless.
You are, of course, entitled to your opinions.
You are bring disrespectful again Josh. I don’t know what the quality of Jeff’s evidence is. It may be correct. I don’t have time to look into it. Stop reading my mind.
If we had listened to doctors on the front line, we would have used hydroxychloroquine with zinc and saved a lot of lives. Now ivermectin is proving its worth at any stage of covid. Herd immunity isn’t necessary.
https://judithcurry.com/2021/01/10/covid-19-why-did-a-second-wave-occur-even-in-regions-hit-hard-by-the-first-wave/#comment-942462
and
https://judithcurry.com/2021/01/10/covid-19-why-did-a-second-wave-occur-even-in-regions-hit-hard-by-the-first-wave/#comment-942461
Yes dpy, I read it.
Follow your own advice. Read more widely.
There is nothing to read since you haven’t cited anything in the scientific literature. You have repeated over and over again your narrative which is based on crude mechanistic explanation that lack quantification. My peer reviewed paper supports that last statement in addition to being a good read.
Jeff –
I just took about 1 minute to look at the CDC website on flu surveillance They do flu surveillance very years. It’s one of the things they are expert at. Believe it or not, they actually have people there who really know how to do that kind of thing.
They are saying that flu occurrences are low this year. I’m sure that among the ways that they measure something like that is by comparing flu hospitalizations and deaths year over year. I doubt there are many people in need of hospitalization due to the flu, or who are dying from the flu this year that they’re not capturing.
Now on the other hand, despite your confidence in your own abilities to conduct flu surveillance, I’m rather sure that you actually know nothing about the topic, and I”m quite sure that you don’t have the resources available to conduct that surveillance that the CDC has. I hate to break it to you, but you have zero credibility here and the CDC’s credibility is quite high on something like this.
The reasons why the rate of flu is low this year is undoubtedly complicated – but people who actually know about this say that quite likely a very significant factor are the interventions undertaken to limit the spread of COVID. Certainly, simply the fact that a lot of children haven’t been in school – particularly high schools – has to be a significant factor. I think you’re way off base if you think that the reduced in-school attendance this year wouldn’t necessarily have to reduce the prevalence of flu infections. So if you’ve convinced yourself that flu infections this year aren’t unusually low, the surveillance system you just invented for yourself is almost undoubtedly highly flawd.
dpy
“Josh, What you are doing here is very disrespectful. I am relying on peer reviewed science.”
You have quoted precisely zero peer reviewed science from the relevant period.
You rely on your own assertion. Nothing else.
Your purely rhetorical comments, VTG, and continued evasions look like bad faith to me.
If you read the paper, you know the important points. If you didn’t read it, you are dissembling. There is a 3rd alternative: You “read” it but didn’t understand it. That last one seems most likely to me.
The paper points to our deep ignorance about flu transmission. If we don’t understand transmission, we can’t attribute its disappearance this year to mitigation. The first sentence of the abstract says: “The epidemiology of influenza swarms with incongruities, incongruities exhaustively detailed by the late British epidemiologist, Edgar Hope-Simpson.” Swarming with incongruities is a good description of most of the comments here from you, Josh, and Atomsk here.
VTG, You moved the goal posts from peer reviewed science to “the relevant period.” The flu has been around for decades and the past science is still valid. This is such an obvious rhetorical tactic to hide the fact that I have real evidence and you have none.
David –
> I don’t know what the quality of Jeff’s evidence is. It may be correct.
Yes, it may be correct The CDC, which hires teams to conduce flu surveillance year after year, with all kinds of expertise and years of experience, and all kinds of resources to apply to the task, and communication networks specifically for accomplishing that task, says that the occurrences of flu is low this year. But some dude on the Interwebs disagrees. But he may be correct.
Yes, that’s true.
I have to say – each and every time I think that my much beloved climate Etc. “skeptics” have exhausted their ability to amaze and amuse me, they prove that I underestimated them.
So, still no peer reviewed science from the relevant period.
Still the insistence others must do what you can’t.
Still no explanation for the strange coincidence of unprecedented flu suppression and covid suppression.
Still plenty of assertion and bombast.
Still I urge you dpy: take your own advice.
You know what’s really weird? Why haven’t people in NZ and other places checked in with Jeffnsails850 before figuring out what’s going on with the flu in their countries?:
-snip-
A tracking system that monitors a cohort of at least 30,000 people for influenza-like symptoms shows as few as 0.3% of New Zealanders reported coughs or fevers some weeks during their winter, a tenfold decrease on some previous years.
-snip-
https://www.theguardian.com/society/2020/sep/17/falling-flu-rates-in-southern-hemisphere-offers-hope-as-winter-approaches-coronavirus
Now this MIGHT seem informative – but it’s actually just prolly ’cause they’re testing less for the flu – just like Jeff says.
-snip-
Fewer than 40 Australians have died from influenza this year, compared to more than 950 last year, “and there haven’t been any deaths for the past three to four months”, Barr added.
Above I posted a link to a story about the huge drop in sales of cold and flu medications. It’s one of many supporting lines of evidence that show the precautions taken to prevent COVID-19 have had a measurable effect on the spread of other respiratory diseases.
Here is another new analysis of the 2020-2021 flu season (02/04/2021):
https://fivethirtyeight.com/features/how-covid-19-ended-flu-season-before-it-started/
“In the third week of 2021, clinical laboratories nationwide tested 23,549 specimens for influenza. Of those, just 0.3 percent (65 tests) turned up positive — a number that is, to put it mildly, absolutely wild.
“Normally, this time of year, we’d be running 20 to 30 percent positive,” said Lynnette Brammer, the head of the Centers for Disease Control and Prevention’s Domestic Influenza Surveillance team.”
Jack –
But Jeff disagrees, ’cause he…. saw something or other at some hospital.
Or something like that..
So as a peace offering, I’ll note that the 538 article that Jack linked did mention the issues that Jeff referred to as well as the notable uncertainties – even as it noted the relative certainty that (1) the flu is dramatically down this year in the US and around the world and that (2) signs point to COVID mitigation measures playing a significant role in the drop.
There, that wasn’t so hard now was it?
dpy
“I have real evidence and you have none”
Once again. Take your own advice. Read more widely.
https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)32166-8/fulltext
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7588551/
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7462755/
https://www.eurosurveillance.org/content/10.2807/1560-7917.ES.2020.25.42.2001816
etc etc etc
And the second reference from VTG is equally inconclusive. It points out merely that the Hong Kong Flu season in 2019-2020 was actually more severe than the 2017-2018 season despite covid mitigations.
I’m done VTG since you can’t even follow simple evidentiary rules and have frequently misrepresented both my references and indeed your own. This level of bad faith makes actual conversations impossible. This is a new low even for you.
dpy
And the second reference from VTG is equally inconclusive. It points out merely that the Hong Kong Flu season in 2019-2020 was actually more severe than the 2017-2018 season despite covid mitigations.
Incredible.
Title of that paper
Abrupt Subsidence of Seasonal Influenza after COVID-19 Outbreak, Hong Kong, China
VTG this is a new low even for you. Your thesis is that mitigation caused the flu to disappear. The paper you cited doesn’t really support that. You must not have read it.
You’re first reference also doesn’t support your thesis. Whether … Could affect other infections is conditional and does not confirm your thesis.
‘As restrictions are reinstated in Europe to control increasing COVID-19 case numbers, the southern hemisphere experience suggests consideration must be given to whether these NPIs could affect other transmissible infections—particularly influenza, with its high morbidity, mortality, and health-care costs—and how this off-target effect on viruses other than SARS-CoV-2 could protect health system capacity. As evidence on both the benefits and costs of NPIs in the COVID-19 pandemic accrues, their role in the management of future influenza pandemics can be carefully considered.’
dpy,
It’s literally in the title of the paper.
This is surreal.
This is classic goal post shifting.
1. You repeated many times that covid mitigation caused the flu to disappear.
2. I said we don’t know why. There are several possibilities.
3. You post some links that point out that flu disappeared but do not really attribute it to mitigation. They discuss many possibilities.
4. VTG then employs the classic proof by repetition.
Your own links prove you wrong. It’s a new low even for you in bad faith.
dpy,
The papers all discuss the unprecedented reduction in flu caused by the covid restrictions.
Indeed, the quotes you make discuss the potential to use NPIs for future flu suppression based on this evidence.
dpy6629 | February 6, 2021 at 12:25 pm |
This is classic goal post shifting.
1. You repeated many times that covid mitigation caused the flu to disappear.
2. I said we don’t know why. There are several possibilities.
3. You post some links that point out that flu disappeared but do not really attribute it to mitigation. They discuss many possibilities.
4. VTG then employs the classic proof by repetition.
Your own links prove you wrong. It’s a new low even for you in bad faith.
verytallguy | February 6, 2021 at 12:32 pm |
dpy,
The papers all discuss the unprecedented reduction in flu caused by the covid restrictions.
Indeed, the quotes you make discuss the potential to use NPIs for future flu suppression based on this evidence.”
Above is just two of the long running argument between Very tall guy/ josh / DPY
Yet has anyone noticed the contradiction –
The Flu Infection is near non existent due to the Covid mitigation steps
vs
Covid continues to spread rapidly because people arent taking proper mitigation steps .
Joe –
> Yet has anyone noticed the contradiction –
The Flu Infection is near non existent due to the Covid mitigation steps
vs
Covid continues to spread rapidly because people arent taking proper mitigation steps .
There’s no “contradiction.” Both can easily be true at the same time.
There are at least two relevant considerations.
(1) Just because COVID is spreading rapidly (although at a decreasing rate recently) doesn’t mean that the NPIs aren’t reducing the rate of spread compared to what it would be absent the NPIs – just like it is mostly likely doing with the flu that can see by looking at historical precedent. There is no historical precedent for comparison with COVID.
(2) the mechanics of viral spread are different with the flu than with COVID. For example, the lack of in-person school attendance might affect the spread of the flu more than it affects the spread of COVID.
Think.
VTG, can you read? Your second reference says:
‘the southern hemisphere experience suggests consideration must be given to whether these NPIs could affect other transmissible infections’
‘Whether these could’. This in English means that it’s a possibility to be considered. That supports my position that no one knows for sure. It’s one of many possibilities. Could be a combination of several factors.
Continuing to deny what your own references clearly say is well not acting in good faith.
Joe cuts to the point:
VTG and Joshua want us all to know that lockdowns made the flu “disappear” (as defined as not disappearing) because they “work” against every virus except the one they were intended.
Furthermore, despite the fact that the CDC, all state health departments, and all hospitals say they are testing millions of people who have “flu like symptoms” for COVID, finding them negative, and per the CDC are not testing millions for other flu, we must conclude that none of them were sick.
Because if we admit that sick people are sick then the lockdowns didn’t even stop the viruses we didn’t need to stop. Hence they are not sick, if they ever existed in the first place.
“VTG and Joshua want us all to know that lockdowns made the flu “disappear” (as defined as not disappearing) because they “work” against every virus except the one they were intended.”
As has been explained, seasonal influenza has a much lower R0 than SARS-CoV-2, and thus mitigations which only slow down the latter are enough to prevent sustained outbreaks of the former.
Jeff – that is close to the point I was trying to make.
The covid mitigation has reduced covid transmission very little, maybe 10%-15% with a max of maybe 25%.
Whereas, covid mitigation has reduced flu transmission by as much as 90%-95%. . That is too large of a discrepancy to reconcile. If mitigation steps were the primary factor, the delta between between the two viruses should be much smaller. Possibly data collection error,
“The covid mitigation has reduced covid transmission very little, maybe 10%-15% with a max of maybe 25%.”
Where did you get that number?
“Whereas, covid mitigation has reduced flu transmission by as much as 90%-95%. . That is too large of a discrepancy to reconcile. If mitigation steps were the primary factor, the delta between between the two viruses should be much smaller. Possibly data collection error,”
If it reduces it by as little as 25%, it can reduce the number of cases by 99.9%. That’s how exponentials work, especially since influenza basic rate of transmission (R0) is already close to the threshold below which outbreaks die out rather than grow.
Joe –
> The covid mitigation has reduced covid transmission very little, maybe 10%-15% with a max of maybe 25%.
Whereas, covid mitigation has reduced flu transmission by as much as 90%-95%. .
People who know about this stuff, you know the ones who actually study this stuff year after year, prolly know more about this stuff than people who don’t know anything about this stuff and think that they can just figure it out.
The people who know about this stuff point to the simple fact that the seasonal flu is less transmissable than COVID as a partial explanation for why the reduction in the flu this year would be so dramatic in comparison to previous years. For that reason as well as others, a given NPI could have a greater suppressing impact on the transmission of flu than it would on COVID. For example, the closing of in-person schooling would have a noticeably greater impact on the transmission of flu because than on COVID because, it appears, kids transmit COVID at a lower rate than the rate at which they transmit the seasonal flu.
And of course, you actually have no idea how much the NPIs have reduced COVID transmission.
And of course, as Jeff has said, and as people who know about this stuff agree, a lack of people showing up at clinics and prolly less testing than typical, have reduced the rate at which the flu has been detected in comparison to other years. So the actual drop in flu transmission might not be as great as it appears if you simply look at identified cases.
> If mitigation steps were the primary factor, the delta between between the two viruses should be much smaller. Possibly data collection error,
But you did use the word “delta,” so that does convince me that you prolly know more about this issue than, say, the professionals at the CDC who have expertise in these matters and study the data and conduct surveillance year after year, and have communication networks related to these matters.
Here’s a report from a while back from the CDC. But hey, what do they know, right?
https://www.cdc.gov/mmwr/volumes/69/wr/mm6937a6.htm
Did you read your reference Josh? It actually says that
“These changes were attributed to both artifactual changes related to declines in routine health seeking for respiratory illness as well as real changes in influenza virus circulation because of widespread implementation of measures to mitigate transmission of SARS-CoV-2.”
I saw no attempt to quantify these effects either.
Yet another inconclusive non contribution to science.
Lol. No David. I didn’t read it. I linked it but I didn’t read it.
Frank –
> What is the value in Operation Warp Speed when it is going to take longer to simply vaccinate Americans than it took to develop a vaccine!
I don’t understand your point. The value is that in the end many fewer get infected and/or die than otherwise.
“Conclusion: The pandemic is currently immunizing Americans as fast or faster than the vaccine.
What is the value in Operation Warp Speed when it is going to take longer to simply vaccinate Americans than it took to develop a vaccine!”
It’s only winning because people are not taking adequate precautions – especially in the LA area and here in Arizona. How to get them to behave better? I’m not sure, but I don’t think we try hard enough.
I’d have a big propaganda campaign to wake people up to it – like the one we had against cigarette smoking. Images of people on ventilators on signs all over the place. Frequent strong talks by popular politicians and media figures. That sort of thing.
That antismoking campaign was only very partially successful. With legalized weed no one seems to care much anymore. Scaring people about COVID-19 is unethical. People need the best evidence to make their own decisions.
Monthly suicide rates in Japan increased by 16 percent from July to October 2020 compared to the same period the previous year, Japanese researchers have found. The three-month interval coincided with Japan’s second wave of Chinese coronavirus outbreaks, Reuters reported Sunday.
https://www.breitbart.com/asia/2021/01/18/study-japan-suicides-jumped-16-percent-coronavirus-second-wave/
David –
> I did not mention cases in this thread.
1l I listed a series of things that you said that are incorrect. In that list I included when you made incorrect statements about testing in Florida relative to other states.
2) In response to my list you said you didn’t make any statements about testing.
3) I showed that you had, and that your statement about testing was wrong (as was your statement about cases in Florida relative to other states) – exactly as I had said.
4) You responded by calling me names.
5) And on and on
> You keep denying that my main point was correct in July.
Your “main point” was as I quoted you. And you were wrong in July. You wrongly stated the number of tests (and cases) in Florida relative to the other states. Again, you said:
——
dpy6629 | July 28, 2020 at 11:44 am |
Of course testing in Florida and Texas testing is vastly more extensive than in New York.
—–
That was wrong. Spectacularly wrong. And I showed you at the time that you were wrong. Spectacularly wrong.
> You are the one who constantly was citing cases as meaning something. Your main points were stupidly wrong then and you are stupidly wrong now.
Cases matter. I can only assume that you’re trying to make some kind of point of comparing true cases to infections, where not all infections are identified as true cases. If that was your point, the fact that there are more infections than there are true cases (or identified cases) doesn’t mean that the number of cases isn’t important (in conjunction with the number of tests conducted).
Once again. All over the world. Everywhere. The number of illnesses, hospitalizatons, and deaths all track with the number of cases, even if the ratio varies depending on context. Everywhere.
You made this error in the summer, where you declared that the spike in cases wouldn’t lead to a spike in hospitalizatons and deaths. You were told you were wrong at the time, but you wouldn’t listen then. It seems that you are maybe making the same mistake now?
If so, you shouldn’t feel too bad as you are not alone. A lot of people have been trying to decouple the number of cases from the level of illness, hospitalizatons, an deaths from early on in the pandemic. I mean it’s bad that so many people are so wrong, but maybe you can take some solice in that you aren’t the only one?
SteveS | January 11, 2021 at 9:56 pm | Reply
I agree with Geoff…Joshua you are way off base with your attitude…It’s simple without getting your undies all bunched up… nick had a thought about transmission of the virus relating to heard immunity…. turns out that idea was probably not so good in hind sight….so what….I can’t find anyone, anywhere who has accurately predicted the path of the Pandemic. Would be nice if you took that little bit of “vengence” that pervades your on-line persona somewhere else. It’s a problem you have to work out on your own.
dpy6629 | January 11, 2021 at 10:38 pm | Reply
Josh is often illogical, very biased, often wrong, and talks way too much. Word count to substance ratio is very high even for anonymous nonscientist ankle biters. That he thinks he can show such disrespect to serious scientists is arrogant.
I guess It’s obvious to many commenters here Josh that you have personality issues.
niclewis | January 11, 2021 at 11:30 am | Reply
Without any attempt at explication of reasonable parameters, that statement could mean that “HIT” could be almost anything and thus “HIT” seems pretty much meaningless.
However, I have made such an attempt in my article. Why don’t you propose what you think are reasonable parameters if you disagree with my suggestions, rather than always just sniping unpleasantly?
That was in response to Josh
Florida, which has been more open and less restricted than its west coast counterpart, has recorded just above 1.5 million cases. Yet California, despite being among the strictest in the nation when it comes to COVID lockdowns, has recorded nearly twice that figure since the start of the pandemic.
https://www.foxnews.com/politics/closed-california-versus-open-florida
Jim – California total cases per 100k population is approx 7600 vs florida with approximately 7300.
For all practical purposes, nearly identical rates. Though your point of vastly different levels of mitigation efforts and compliance with mitigation efforts is valid.
At the individual level/micro level, mitigation efforts should greatly reduce the spread of Covid. However on the macro level, the positive effects of mitigation dont seem to be having any substantive effect.
CDC says that disparity between California and Florida is already big and getting much worse very quickly.
In California the number of new cases in the last 7 days was 100.9 per 100k in California v 62 per 100k in Florida.
California recorded 279,000 new cases in the last seven days. Florida 93,000. Covid is spreading three times as fast in California as Florida.
https://covid.cdc.gov/covid-data-tracker/#cases_casesper100klast7days
Plus Florida is actually administering the vaccine to people. The state has vaccinated more of its population than California and is still vaccinating at a higher rate.
Thanks for that Jeff. There is a growing body of such comparisons that make it harder and harder to argue that lockdowns add benefit over mild and mostly voluntary policies.
https://virologyj.biomedcentral.com/articles/10.1186/1743-422X-5-29
was pointed to by someone earlier. It’s really an eye opener on just how ignorant we are about viral epidemics.
Hi dpy6629,
I think that when this is all said and done the verdict on lockdowns will be- 1. You can’t impose them in time. A highly contagious virus will spread a long way before anyone can act.
2. You can’t actually lockdown a place like New York City. The city was 8 million people all living on top of each other coughing on each other. Then half of them fled.
3. You can’t actually keep the entire young person population locked in their rooms for 10 months. Which means they’re going to catch it and give it to their parents, neighbors, local store clerks, etc etc. Especially once they realize the virus is harmless to them.
4. It’s a bad idea to politicize a virus. The number of new cases in the last 7 days (per capita) is worse in New York than Florida just as it is worse in California than Florida. Absent storm troopers, any “lockdown” requires voluntary compliance and accurate information about how effective the effort is. The LA Times, The New York Times, The Miami Herald, and CNN/ABC/NBC/CBS are all “reporting” the opposite of the truth. They report that the Florida government and Floridians are doing it all wrong. And that New York officials and residents deserve awards for their extraordinary effort. In which state do you think people will be more careful tonight?
Good summary Jeff. It is indeed true that over the last 4 years (perhaps 12) American corporate media has descended to yellow journalism and is worse even than the Gilded Age. They are very sensationalistic. They are openly highly partisan and ideological. Their ideology is woke ideology which is based on lies such as the 1619 project. They demonize anyone who disagrees and try to cancel them. At least in the 19th Century there were few barriers to entry in the newspaper business and there were literally tens of thousands of papers to choose from with every possible point of view.
Covid19 has indeed been weaponized by the media. They constantly focused on totally wrong case fatality rates up to 10% while ignoring more sober (and correct) scientists. They have portrayed Cuomo as a compassionate very competent leader and Trump and De Santos as causing death and suffering. They have stirred up unwarranted fear spread by TV “doctors” who are often unqualified. I don’t see how this can be fixed other than by breaking up into little pieces the media and tech titans.
Good summary Jeff. The corporate media have for at least 4 years descended into a level of corruption not seen since the Gilded Age. Extreme sensationalistic, open and extreme partisanship, and shaping every story to support ideologies that are based on lies. It’s actually worse now because in the 19th Century there were papers with a huge range of points of view. With the tech titans censoring what information most people see, it is similar to the Chinese government’s information operations.
The only solution is to break them up into little bitty pieces.
> .It’s a bad idea to politicize a virus. The number of new cases in the last 7 days (per capita) is worse in New York than Florida just as it is worse in California than Florida.
I’ll give you credit for the unintentiinal irony post if the day – where you completely ignore the relevance of rates of testing when you compare rated of identified infections across communities to draw conclusions a out the efficacy of interventions.
But I will certainly agree that it’s a bad idea to politicize the virus. Comparing across communities is the easiest way to do that, however – because controlling for confounding variables, timing of interventions, levels of “compliance” etc. is nearly impossible and just opens the door for confirmation bias.
Another example of why you shouldn’t make facile comparisons:
-snip-
New California Variant May Be Driving Virus Surge There, Study Suggests https://nyti.ms/39Khpqp
Josh, Another nonsense comment from you that is completely beside the point being made.
Jeff was making the completely true observation that the media has been doing a politically biased and fraudulent hatchet job on De Santos and portraying Cuomo as a hero of the people.
His comparison is not cherry picked. There is no obvious correlation between the political orientation of a state’s governor and their deaths per capita. Nor is there any obvious correlation with mitigation measures. Perhaps at some point, there will be a more rigorous analysis. For Europe, the best evidence shows pretty. much what I just said in that case.
David –
Your opinion is noted. Along with your arguments that a vaccine wouldn’t alter the course of the pandemic. And your argument that the pandemic was over this summer. And that the summer spike wouldn’t lead to a spike in deaths. And your opinion that Nic wasn’t wrong when he sad that Stockholm crossed a HIT 8 months ago and that they would only have around 7400 deaths in Sweden and that 0.085% population fatality was likely a hard limit. And your silly errors about testing rates on Florida relative to NY. And your silly characterization of Andrew Gelman as a know-nothing. And your silly defense of the Santa Clara study… because Ioannidis is a “lion.”
And your silly appeals to authority.
Drivel again. We’ve been over every one of your fraudulent out of context proof texting droppings. Hit was almost certainly reached in Sweden. I newer mentioned testing, that was Jeff. You are such a motivated ankle biter you can’t get any detail right. Grow up. Most people here feel the same way in case you are unable to read clearly.
Oh my, Joshua acknowledges that if you test more, you find more cases. The opposite of what he was saying before when it was just good fun to claim the massive ramp-up in testing under Trump meant that Trump was doing a poor job of “controlling the spread.”
But wait a minute. Why would testing in places so perfectly locked down – hermetically sealed by the magical powers of Democratic Party politicians – result in the discovery of any cases? How could there be any spread of a virus in areas where they are “controlling the spread?”
The answer, of course, is because in New York and California they aren’t now, never were, and never really could be “controlling the spread” and by pretending otherwise they accidentally encouraged people to go out and get sick.
Meanwhile in Florida, where for political reasons the entire media establishment is not pretending the spread is under control, people are more cautious.
Politicization is hurting Californians and New Yorkers- not that anyone in the D party cares, they’re still coming up with new ways to give awards to Cuomo for the worst infection rate and death rate on the planet.
Maybe that’s why people are moving out of New York and California and heading to Florida.
“Politicization is hurting Californians and New Yorkers- not that anyone in the D party cares, they’re still coming up with new ways to give awards to Cuomo for the worst infection rate and death rate on the planet.”
Nothing like commending coumo for doing a great job for putting the fire out after the building has burned to the ground
David –
> .I newer mentioned testing, that was Jeff. You are such a motivated ankle biter you can’t get any detail right. Grow up.
Why do you insist on having me use the Google to go back and show you your claims are false? We’ve been through multiple times. Do you think I’ll suddenly forget how to use the Google?
———
dpy6629 | July 28, 2020 at 11:44 am |
Of course testing in Florida and Texas testing is vastly more extensive than in New York. Actual infections would track deaths meaning that by any meaningful measure New York is 5 – 10 times worse per capita. Joshua typically can’t acknowledge any truth but nitpicks with largely meaningless statistics. He never sees the point so why would anyone expect him to behave like an adult.
Joshua | July 28, 2020 at 11:53 am |
> Of course testing in Florida and Texas testing is vastly more extensive than in New York.
LOL.
Tests per million:
NY: 289,000
Florida: 162,000
Texas: 128,000
Josh, The main point I was making that you quote from was correct then and correct now. In terms of deaths which is what counts, New York held then and continues to hold the top spot among the biggest states. As I recall I was saying testing and cases are less informative than hospitalizations and deaths. Someone else was citing cases as evidence that Florida and Texas were doing a poor job. Was that you Joshy? It sounds like you since you typically cite random and often inconclusive facts.
Current deaths per hundred thousand:
New York: 206
Texas: 107
Florida: 108
California: 79 and catching up rapidly.
You Josh are like some left wing canceller who combs the public statements of someone they don’t like and viola find that in a sterling record, there are a few things that can be interpreted to be wrong or offensive.
Just stop. It’s childish and immoral behavior. The equivalent of online stalking.
https://judithcurry.com/2021/01/10/covid-19-why-did-a-second-wave-occur-even-in-regions-hit-hard-by-the-first-wave/#comment-941043
Geoff Sherrington | January 11, 2021 at 8:25 pm | Reply
BTW, I deplore the actions of those here who are critical of Nic Lewis in a nasty, personal way. Nic is attempting to clarify a complicated piece of medical science that has some math features best handled by a mathematician.
He’s talking about you Josh.
David –
Which was, of course, after this:
———-
dpy6629 | July 27, 2020 at 9:34 pm |
Bear in mind that total cases in Florida and Texas are still vastly lower than in New York. Deaths per capita are also 5-10 times lower. But Florida did such a terrible job compared to saint Andrew according to the corrupt media. Florida dnd Texas have more uninfected people left. I have also hard the speculation about hot weather and air conditioning.
Joshua | July 27, 2020 at 10:39 pm |
> Bear in mind that total cases in Florida and Texas are still vastly lower than in New York.
LOL. Yeah. Bear that in mind: vastly lower
Worldometers:
New York 440,472
Florida 432,747
Texas 404,179
Check back in two days for Florida, maybe two weeks for Texas.
Well so what? Cases are irrelevant. What I probably meant was that infections were vastly higher in New York since deaths were vastly higher. Someone else continues to cite case numbers, looks like that was you.
David –
> Cases are irrelevant.
Cases, in conjunction with the rate of testing, are relevant as hospitalizatons and deaths very in parallel with cases. One of the problems with those who are downplay the pandemic is that they keep thinking they can decouple cases and harmful outcomes – which is clearly wrong.
That is the same mistake you made during the summer when you said the pandemic was over.
What you said about both testing and cases was wrong. Ridiculously wrong. As I pointed out at the time.
Then you said you didn’t say anything about cases, and that I couldn’t get the details right- both of which were false.
Josh you are like a mosquito. You infect any venue where you comment. You buzz around searching for any prey who will gratify your need to confirm your importance when in fact you are really totally irrelevant. Your comments are uninteresting and obfuscatory. Please go elsewhere with your unpleasant sniping.
I meant of course that in this thread I had not mentioned cases.
The basic point I was making in July was correct of course. Cases are largely irrelevant. It’s infections that are important. That you seem to dispute that (but in your usually obfuscatory and vague way) fits your pattern of online stalking. Just stop. By now everyone who reads this thread will agree with me.
David –
> I meant of course that in this thread I had not mentioned cases.
I listed a few of a long list of errors you made. Included was an error regarding cases. To which you said you never mentioned cases. So I show where not only did you mention cases – as I said – but where you were obviously wrong about cases. Then you called me names, as you often due, because I oij Ted out your original weir and your subsequent error about whether you made an error originally. I can show where you made each of those other errors I listed also. As in fact I’ve done before when you called me a “liar”
The basic point I was making in July was correct of course. Cases are largely irrelevant. It’s infections that are important. That you seem to dispute that (but in your usually obfuscatory and vague way) fits your pattern of online stalking. Just stop. By now everyone who reads this thread will agree with me.
> The basic point I was making in July was correct of course.
No. You incorrectly quantified the number of cases in different states, relatively. I printed out your error. Don’t move the goalposts.
> Cases are largely irrelevant.
Cases, in relation to the number of tests, is are absolutely relevant. Severe illness, hospitalizations, and deaths move in parallel with the number of cases, and with the number of identified cases. Not all cases result in such outcomes. Not all infections result in such outcomes. But everywhere, #’s of illsses, hospitalizatons, and deaths very in parallel with the number of cases even if not always in exactly the same proportion.
>It’s infections that are important.
Identified cases, and infections, are important. That you declare cases as unimportant doesn’t make it so. But you’re changing the subject anyway.
You made erroneous comments about the numbers of cases in different states. So I pointed out your error
Repeating a falsehood becomes an ad hominem.
I did not mention cases in this thread. Others were discussing it. In addition to your status as immoral you now add logic chopping and an inability to read. Or perhaps you can’t think clearly about what he is saying or can’t remember what others say.
You keep denying that my main point was correct in July. You are the one who constantly was citing cases as meaning something. Your main points were stupidly wrong then and you are stupidly wrong now.
You Josh are like some left wing canceller who combs the public statements of someone they don’t like and viola find that in a sterling record, there are a few things that can be interpreted to be wrong or offensive.
Just stop. It’s childish and immoral behavior. The equivalent of online stalking.
You have been banned at numerous blogs and put on moderation by Judith in the past for what you are doing now. They did this for a reason. Your behavior is unethical and makes comments unreadable.
Now you’re going off the deep end. Anyone claiming that case count for New York is even in the same zip code as “accurate” is trading in political fairy tales.
Nobody has the slightest idea how many cases there actually were in New York in March and April, intelligent people know it was several times the reported number. All you have to do is extrapolate from the fatality rate even after accounting for the outrageous, highly lethal, incompetence of the local and state government.
Which means he was right then and is now- case counts are pretty much irrelevant because anyone with a measurable IQ knows they’re wrong (but useful as Mosher would note).
Of course, intelligent people also know at this point that you cannot really “contact trace” a virus that has no symptoms in almost half of those infected. So Joshua applauds massive testing in New York, while saying there cannot be any cases to discover because “lockdowns work” and claiming we should pay no attention to case counts in New York because they are an artifact of massive testing, and mocking people for claiming case counts are irrelevant (except when he needs them to be irrelevant) or that test counts are irrelevant (except when it suits him politically to say test counts are irrelevant).
In other words, Joshua is a true modern partisan of the left- anything they say today they reserve the right to say the opposite of tomorrow and, if you dare try to address his points using the standards he himself used just yesterday, he will claim you are disingenuous. Oh, and he seeks “unity” with you. On his terms, of course.
> Anyone claiming that case count for New York is even in the same zip code as “accurate” is trading in political fairy tales.
Not sure where I described the early feeding as “accurate.”. While it prolly was relatively close in capturing true cases, obviously a lot more of the mild infections weren’t identified. Of course I wouldn’t doubt that.
The cross state comparisons are basically fodder for confirmation bias. Obviously, there are a lot of varying influences across states – like Florida’s older mean age as well as structural advantages like much less dense population centers, like Miami-Dade vs. NYC, and a “seasonal” type advantage with climate. Also they had a huge advantage w/r/t lead time to prepare versus NYC getting slammed early on, particularly with (I would assume – haven’t looked) international travel.
All the states and governors made mistakes. It isn’t a partisan thing. Dems on the whole prolly over-valued the return from NPIs, such as closing schools. And pubz on the whole prolly under-valued NPIs, and could have had better results had they stressed targeted NPIs more, encouraged social distancing more, mask wearing, etc.
And partisan loyalists on both sides have fallacious leveraged outcomes as proxy weapons in political warfare. So what else is new?
Bottom line, however, is thst if you’re going to play partisan warfare here you have a major problem if your coming from the place of supporting our federal government. As a country we have had outcomes about 4 x worse than the world average. Waaay higher % of cases (despite mediocre levels of testing) and deaths than % of the world population. Inexcusable, given that we have so many resources available.
Over 4,000 deaths again yesterday. Seems that the # of infections may be leveling off or even declining. That’s a good thing. Should hopefully be reflected in leveling off or declining deaths in a couple of weeks.
Maybe with a more focused effort federally to aid states with vaccine distribution we’ll get the inoculation rate increasing and start turning the corner. All this partisan bickering and name-calling is irrelevant in that regard.
It’s a collective failure. No need to try to parse put why one partisan ideology is more or less responsible than the other.
> Not sure where I described the early feeding as “accurate.”. While it prolly was relatively close in capturing true cases,
Should be …testing as “accurate.” And “…close in capturing severe cases..”
So which is it today Joshua? Are comparisons OK or are they not OK? You seem to argue all sides of every issue depending on your momentary rhetorical needs.
> Are comparisons OK or are they not OK?
Comparisons are pretty inevitable – they’re a basic and intrinsic part of how we reason. They are neither “OK” or “not OK.” Why be so binary.
They have limited value, and are breeding grounds for confirmation bias. To the extent that one can’t control for important confounding variables, their value is diminished.
When I consider the enormous resources we have to utilize to fight something like public health problems, I tend to expect that we should have very good results on a relative basis. Of course, I recognize that there may well be mitigating factors. For example, the fact that an island nation like NZ has done so spectacularly compared to the US could well be attributed to factors that have absolutely nothing to do with anything that humans can control beyond that accident of the geographical features of where they live.
But yah, when I look at the enormous level of resources that we can bring to bear, and the terrible results that we’ve had on the broad range of outcomes across all countries – I have come to expect better from this country and I think it is unacceptable that we’ve done as poorly as we have
If you’re content to think that we’ve done as well as we should have been expected to have done, that’s certainly your right. You may be correct about that. I’m not sure, exactly, how that’s really meaningful beyond just a simple fact that you have much, much lower expectations of this country than I do.
The best scientific evidence shows little benefit of strong mitigation.
https://onlinelibrary.wiley.com/doi/abs/10.1111/eci.13484
Your evidence free rambling is noted.
David –
Since you’re always telling me that I don’t meet your standard of authority, I’ll offer you an interesting question related to the article you linked from someone who presumably would meet your standard of authority (although, with your inane and insulting criticism of Gelman, a highly respected statistics expert, one never knows).
It’s an interesting question. Maybe you know the answer?
> Did anyone see a convincing explanation for why it was appropriate for the new Bendavid et al. paper to select only the countries they did? e.g., why they included Sweden but not its immediately comparable neighbors like Denmark, Norway, and Finland which implemented lockdowns and had covid death rates a fraction of Sweden’s. Or why they would include South Korea, a virtual-island nation (its one land border is impenetrable) with massive tracing and isolation, for use in comparisons involving the US and The Netherlands.
Contrast the selection in Bendavid et al. to that in the Brauner et al. study in Science last month:
https://science.sciencemag.org/content/early/2020/12/15/science.abd9338
Any other aspects of a study seem secondary if the data have been curated to produce the conclusion.
https://statmodeling.stat.columbia.edu/2021/01/20/what-about-that-new-paper-estimating-the-effects-of-lockdowns-etc/#comment-1669812
“I’m just babbling now . . . Ok, my point is that, like Flaxman, I’m super-skeptical of the claim that lockdowns etc. have no effect. It’s fine to be against lockdowns, and there are lots of good anti-lockdown arguments that don’t require the claim that they have no effect.”
I don’t see much substantial in this blog post. If you want to find some other peer reviewed evidence I’d consider it, but what was presented by Gellman is not real evidence and its certainly not peer reviewed. In any case, he misrepresented what the paper said. It never said lockdowns had no effects. They said it was not statistically significant but consistent with some effect.
You can do better than this I’m sure with your sterling credentials as a merchant of doubt.
“As a country we have had outcomes about 4 x worse than the world average. Waaay higher % of cases (despite mediocre levels of testing) and deaths than % of the world population. ”
Literally nothing in that sentence is true. The US is middle of the pack for wealthy western nations for deaths, tested more per capita than Europe (hence the higher case count- it’s the only way you ID asymptomatic cases) and even the media is coming around to noticing that Europe undercounted deaths. The Wall Street Journal had a good analysis of excess mortality recently.
On Jan. 20 3:41 p.m. you claimed case counts in FL were not “vastly lower” than NY, citing the ridiculously inaccurate official NY count. Hint, the spread in FL came after NY (and was, as the Washington Post found caused by fleeing New Yorkers and New Jerseyites who were, according to politicians, all safely locked down at home and not roaming Florida).
Death rate per capita in New Jersey and New York (the former due mostly to NYC) is just shy of twice that in Florida and Texas. Unless you want to confess that NJ, NY, MA, MI, IL are twice as incompetent as the governments of TX and FL, then you need to acknowledge that the real case counts in those states are about twice the official total.
And you still aren’t addressing the point: If, for political reasons, your media and politicians say people are doing it right in NY and CA but are doing it wrong in FL- guess where people are going to be more cautious? Florida, obviously. Next question- if you insist for political reasons that “lockdowns work” and “your state of California is doing great” will people think they are safer (and therefore be less cautious) in California? Yes, obviously.
One more time- the rank, partisan dishonesty about the virus is killing people in blue states and, ironically, saving lives in red states.
Yes Jeff, When Josh’s comments have actual content they are usually wrong and embarrassingly so.
Missed this earlier:
> “As a country we have had outcomes about 4 x worse than the world average. Waaay higher % of cases (despite mediocre levels of testing) and deaths than % of the world population. ”
>> Literally nothing in that sentence is true. The US is middle of the pack for wealthy western nations for deaths,
Notice the insertion of “wealthy” and “western” nations and you just have to laugh.
Here, let’s add some more adjectives and qualifiers.
The US is at the top of the pack for wealth, western nations in North America if you exclude Canada.
Once again, as a country we have outcomes about 4 X worse than the world average. Waaaay higher % of cases (despite mediocre levels of testing) and deaths than % of the world population.
What % of the world’s population do we have, Jeff? And what % of the world’s infections do we have? Compare and contrast.
Missed this earlier.
“Notice the insertion of “wealthy” and “western” nations and you just have to laugh.”
Why shouldn’t we compare like countries? Surely, based on your assertion that “it was all Trump’s fault” you’re eager to compare fatality and spread between the US and the enlightened democratic socialist state of Belgium- home of the EU.
No doubt you support the ouster of any politician in any country that did worse than us, right?
Also, about 21% of Florida population is over 65 vs about 15% for Cali.
It’s worth quoting the abstract of the paper I linked above for VTG for the rest of the denizens, I know VTG will not read it.
“The epidemiology of influenza swarms with incongruities, incongruities exhaustively detailed by the late British epidemiologist, Edgar Hope-Simpson. He was the first to propose a parsimonious theory explaining why influenza is, as Gregg said, “seemingly unmindful of traditional infectious disease behavioral patterns.” Recent discoveries indicate vitamin D upregulates the endogenous antibiotics of innate immunity and suggest that the incongruities explored by Hope-Simpson may be secondary to the epidemiology of vitamin D deficiency. We identify – and attempt to explain – nine influenza conundrums: (1) Why is influenza both seasonal and ubiquitous and where is the virus between epidemics? (2) Why are the epidemics so explosive? (3) Why do they end so abruptly? (4) What explains the frequent coincidental timing of epidemics in countries of similar latitude? (5) Why is the serial interval obscure? (6) Why is the secondary attack rate so low? (7) Why did epidemics in previous ages spread so rapidly, despite the lack of modern transport? (8) Why does experimental inoculation of seronegative humans fail to cause illness in all the volunteers? (9) Why has influenza mortality of the aged not declined as their vaccination rates increased? We review recent discoveries about vitamin D’s effects on innate immunity, human studies attempting sick-to-well transmission, naturalistic reports of human transmission, studies of serial interval, secondary attack rates, and relevant animal studies. We hypothesize that two factors explain the nine conundrums: vitamin D’s seasonal and population effects on innate immunity, and the presence of a subpopulation of “good infectors.” If true, our revision of Edgar Hope-Simpson’s theory has profound implications for the prevention of influenza.”
Article on new study: “College Campuses Are COVID-19 Superspreaders, Study Suggests”
“Senior author, Ellen Kuhl, adds: “Strikingly, these local campus outbreaks rapidly spread across the entire county and triggered a peak in new infections in neighboring communities in more than half of the cases.”
https://www.labmanager.com/news/college-campuses-are-covid-19-superspreaders-study-suggests-24876
As Sweden has been so much in the news and has been the subject of 2 articles here, I provide this link which gives comprehensive data from Swedish sources about 2020
https://softwaredevelopmentperestroika.wordpress.com/
The take home is that the excess deaths in 2020 in Sweden were the worst since 2012 when taking into account population size. In the UK every year prior to 2008 to 1900 had worse excess death rates than 2020 when taking the same factors into account. There are many factors to take into account in these bald figures.
Both death rates in Sweden and the UK made worse by the decision to send infected or non tested patients from hospitals into care homes and not protecting the hospitals which in themselves became vectors of infection.
In Britain up to 30% of patients being infected who came in for other reasons. It is currently 12% and rising again, around 40% of all UK covid deaths occurred in our care homes and although it dropped late last year is rising again as the same policies are provided, as the NHs try to clear the decks to provide more beds
It would be interesting to see this data picked apart by NIc as no one wants to pretend covid is a hoax and that it is not a serious ailment, with those affected ranging from not knowing they had it to death and every condition in between.
However that is one part of the balance sheet, the ongoing consequences of ill health through not getting treatments,(some 4 million treatments are in abeyance) mental health, the economy, societal damage, all have to be included into the bigger picture when considering if the West generally have followed the right actions
Tonyb
Tony –
> Both death rates in Sweden and the UK made worse by the decision to send infected or non tested patients from hospitals into care homes and not protecting the hospitals which in themselves became vectors of infection.
I don’t knew where your getting this from. I’ll look at your link. But Sweden is noted for not even treating COVID infected older people in congregate living with oxygen, let alone transporting them to hospitals – because they were focused on sacrificing those lives for the sake of not allowing thise people to infect others.
Also:
https://twitter.com/zorinaq/status/1351599368013537285?s=19
I remember posting a link last year about the similarity between Sweden the UK and I believe Italy, in not protecting their care homes properly with a resulting high mortality.
tonyb
Tony –
> I remember posting a link last year about the similarity between Sweden the UK and I believe Italy, in not protecting their care homes properly with a resulting high mortality.
Yes. The Swedish policies for how they handled infected older people in congregate living undoubtedly contributed in some ways to their high mortality rate. On the other hand, they enacted those policies in order to limit spread – mitigating their mortality rate.
It’s complicated. That’s just one reason why people should be circumspect about trying to use Sweden as an example for how other countries should develop and implement their own policies. There are many, many other reasons.
Trying to extrapolate across countries, imo, is most likely about confirmation bias more so than anything else – as we don’t as of yet have sufficient ability to control for confounding variables.
Counterfactual thinking about what would have been if things were different (in a particular country) is hard.
But Sweden enacted their policies specifically to help limit the spread in care homes. Sonrhres a problem with your logic.
Tony –
The other thing about the higher mortality in Sweden is that it reverse a fairly consistent trend of a drop in the mortality rate – so the increse this year should be viewed in that context. Just saying something like it isn’t unprecedented misses some of thst context.
Tony –
Finally, excess mortality is a very noisy statistic for evaluating the impact of the pandemic. Again, there are a lot of confounding variables – for example deaths that didn’t occur because of less travel, or fewer deaths from the seasonal flu. Of course the uncertainties can run in either direction. Sure, it’s a statistic that has meaning but it’s certainly not sufficient for measuring the impact of the pandemic – of only because focusing on deaths doesn’t account for the enormous amount of serous illness and hospitalizations, and the enormous economic costs.
Tony’s link makes the case that age and population adjusted mortality is the best measure to use. The CDC reports this metric too.
Yes Tony. CDC data shows the same thing for the US. Age adjusted mortality has been falling since 1900 and 2020 will come in at about the 2000 levels. This data is never shown because it’s not as scary as absolute fatality numbers.
Bevard’s tweet is misleading. First he truncated the scale of the second graph. The first graph supports Tony’s assertions.
Here’s an interesting commentary completely demolishing a Flaxman paper that claimed lockdown worked in England. Flaxman is one of those Gelman asked about the Ioannidis latest paper on NPI’s. Flaxman found it wanting. Sounds like Flaxman does pretty obviously bad science.
https://www.frontiersin.org/articles/10.3389/fmed.2020.580361/full?&utm_source=Email_to_authors_&utm_medium=Email&utm_content=T1_11.5e1_author&utm_campaign=Email_publication&field=&journalName=Frontiers_in_Medicine&id=580361
There are lots of other papers also pointing to the same thing: Less strict measures do seem to have an effect. Lockdowns and stay at home orders do not really.
Re: “As Sweden has been so much in the news and has been the subject of 2 articles here, I provide this link which gives comprehensive data from Swedish sources about 2020”
https://softwaredevelopmentperestroika.wordpress.com/
“The take home is that the excess deaths in 2020 in Sweden were the worst since 2012 when taking into account population size. In the UK every year prior to 2008 to 1900 had worse excess death rates than 2020 when taking the same factors into account. There are many factors to take into account in these bald figures.”
Your source is a crackpot blog, and your claim on excess deaths is wrong. You’re likely conflating ‘total deaths’ with ‘excess deaths’. Sweden’s excess deaths are the highest they’ve been in decades, as Joshua showed in the tweet he cited to you. Contrarians (like those at your blog link) evade that by ignoring that Sweden’s total number of deaths have been decreasing for years, and that needs to be factored in when figuring out if Sweden had an abnormally high number of deaths in 2020 due to COVID-19. If that’s hard to hard to grasp, then here are two layman level introductions:
https://archive.is/ndkNq#selection-2745.0-3031.22
https://archive.is/nW05m#selection-2755.0-3281.2
And here’s an analogy:
Imagine your start off with 1000 deaths, and each year you have 10 less deaths than the previous year. So you end up with this sequence:
1000, 990, 980, 970, 960, 950, 940, 930, 920, 910
So if everything is going normally, what would be next number of deaths in the sequence? It would be 900, obviously. But suppose the next number was instead:
945
Any competent adult can tell that 945 value is abnormally high, *given the preceding trend.* It represents 45 excess deaths. But what crackpot COVID-19 trolls do is pretend that 945 value is normal, since 945 is close to a 940 value from 4 years earlier. That’s their under-handed trick. Competent experts don’t engage in those sorts of disingenuous tricks, which is why they note that Sweden had high excess mortality in 2020; i.e. Sweden’s number of deaths in 2020 was abnormally high due to COVID-19, to a level not seen in decades. For example:
“Preliminary statistics on deaths (published 2021-01-18)” in:
https://scb.se/en/finding-statistics/statistics-by-subject-area/population/population-composition/population-statistics/
“Magnitude, demographics and dynamics of the effect of the first wave of the COVID-19 pandemic on all-cause mortality in 21 industrialized countries” (extended data table 1)
“COVID-19 and excess all-cause mortality in the US and 18 comparison countries”
“Estimating total excess mortality during a COVID-19 outbreak in Stockholm, Sweden”
“Excess mortality: the gold standard in measuring the impact of COVID-19 worldwide?”
“Statistics from Statistics Sweden show that it is likely that around 95,000 people will die, compared with an average of 90,000 per year.
– This provided that the death toll does not start to rise sharply, says Tomas Johansson at Statistics Sweden.”
https://web.archive.org/web/20201109203118/https://www.dn.se/sverige/scb-sverige-pa-vag-mot-hogsta-antalet-avlidna-sedan-2015/
[with: https://pbs.twimg.com/media/EnR7DerW4AcDH-Z?format=png&name=small (from: https://www.scb.se/en/About-us/news-and-press-releases/excess-mortality-in-sweden-is-followed-by-mortality-deficit/ )]
“Sweden looks set to move to an excess mortality rate of between 6,000 and 7,000 people by 2020.
[…]
We have had a peak this spring where many died for a limited time and so now in November, December where we also see a peak. The rest of the year it has been at fairly normal levels, says Tomas Johansson, investigator at Statistics Sweden.
[…]
Based on the level we now see, we can expect to end up with around 97,000-98,000 deaths when all have been reported. In absolute numbers, 2020 will be a year among those with the highest number of deaths since 1749. Since then, we have had three years with more than 100,000 deaths, one during the 18th century (1773), one in the middle of the 19th century (1857) and so during the Spanish flu in 1918, says Tomas Johansson.”
https://web.archive.org/web/20210112175701/https://www.dn.se/sverige/sverige-pa-vag-mot-7-000-i-overdodlighet-under-2020/
[with: https://pbs.twimg.com/media/ErjO0F8XIAA-Obc?format=png&name=small (from: https://www.scb.se/en/About-us/news-and-press-releases/excess-mortality-in-sweden-is-followed-by-mortality-deficit/ )]
Even layman’s level sources note this as well:
https://ourworldindata.org/excess-mortality-covid
https://www.economist.com/graphic-detail/2020/07/15/tracking-covid-19-excess-deaths-across-countries
https://www.nytimes.com/interactive/2020/04/21/world/coronavirus-missing-deaths.html [ http://archive.is/ZCoO3 ]
https://www.ft.com/content/a26fbf7e-48f8-11ea-aeb3-955839e06441
https://www.bbc.com/news/world-53073046
https://twitter.com/zorinaq/status/1353825821941784579
Like so much of Sanakan’s work this comment is based on a misrepresentation.
1. The blog post correctly shows total deaths which in 2020 are at historical highs. This is what Sanakan’s very long winded proof texting is about.
2. The main point is that VTG is is not very meaningful. Most professionals use age and population adjusted mortality. The blog post correctly says this statistic is at levels comparable to 2013 and lower than virtually every year before that.
3. Cdc statistics show that for the US this number for 2020 is comparable to 2003 roughly.
This Sanakan gem relies on misrepresenting the blog post.
1. It does show absolute mortality numbers and the graph shows that 2020 is higher than most of the record.
2. This number is very misleading and professionals use age and population adjusted mortality.
3. Using this statistic, 2020 in Sweden will be about at the 2013 level and lower than all preceding years back to the 1960’s.
4. The CDC also reports this statistic and in the US, 2020 will come in at 2003 levels roughly and will be lower than all recent years prior to 2003.
Sanakan’s endless proof texting on absolute annual mortality is known by professionals to be misleading.
Atomsk’s Sanakan
I link to another very detailed study using data from Office of National Statistics and Public Health England
https://architectsforsocialhousing.co.uk/2021/01/27/lies-damned-lies-and-statistics-manufacturing-the-crisis/
They reference ‘mortality’ as well as “Excess deaths.” and take into account the ageing profile and population numbers.
It comes to pretty much the same conclusions as previously, so I would be interested in your take on this.
One of your references refers to ‘absolute numbers”. Sweden had approx 5.5 million people in 1920 and around 10.5 million in 2020 so that is not a fair comparison.
tonyb
@dpy
I’ve already corrected your errors on excess mortality, as have others, including Andrew Gelman and those at his website. So your claims on the topic are moot, and your predictions failed:
https://archive.is/vtF6j#selection-36765.0-37101.16
@climatereason
Re: “I link to another very detailed study using data from Office of National Statistics and Public Health England”
No, you linked another crackpot blog. I strongly suggest you stop relying on such sources, stop citing them as if they’re credible, and instead go to reputable scientific sources. To do otherwise is like saying you want to learn about the shape of the Earth, while just reading and frequenting flat Earther blogs that claim to do “very detailed study using data from NASA” to claim Earth is flat. That’s simply move from garbage source to garbage source, as each of them are debunked.
If you want to know what the Office of National Statistics (ONS) and Public Health England (PHE) actually show, then go their websites. PHE shows a larger number of excess deaths due to COVID-19, as does ONS and EuroMOMO. I’m not interested in what random conspiracist websites make up to avoid that, anymore than I’m interested in what flat Earther blogs make up to avoid NASA showing Earth is round.
– PHE:
https://fingertips.phe.org.uk/static-reports/mortality-surveillance/excess-mortality-in-england-latest.html
– ONS:
https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/deaths/bulletins/changesinallcauseandcovid19mortalityovertimeenglandandwales/deathsoccurringbetween28december2019and4september2020
– EuroMOMO:
https://www.euromomo.eu/graphs-and-maps
And there are already published papers confirming what this data shows, including:
“Excess all-cause mortality during the COVID-19 pandemic in Europe – preliminary pooled estimates from the EuroMOMO network, March to April 2020”
“All-cause excess mortality observed by age group and regions in the first wave of the COVID-19 pandemic in England”
“Excess mortality in England and Wales during the first wave of the COVID-19 pandemic”
Re: “One of your references refers to ‘absolute numbers”. Sweden had approx 5.5 million people in 1920 and around 10.5 million in 2020 so that is not a fair comparison.”
Population size is already accounted for, and does not change the reality that Sweden had more excess mortality than it has in decades. I’m also going to ignore any further citations of random, crackpot blogs.
So the blog post Tony cited appears to be correct and Sanakan’s focus on absolute mortality numbers a poor statistic to use. This is a new low for even Sanakan.
In terms of age and population adjusted mortality 2020 in Sweden is lower than 2013 and all previous years back to the mid 20th century.
Joshua – Bevand statement is very misleading. Observed mortality in Sweden for 2020 with covid is below every year prior to 2013 with the exception of the 1950’s .
There seems to be a lot of talk about vaccine certificates, e.g. before being allowed to travel. Not sure I understand the logic; I thought we still don’t know if the vaccine stops you passing the virus on, so I’m assuming (perhaps wrongly?) that with the vaccine you can still get the virus but 90% of the time fight it off before you get bad symptoms, in other words we still have the pre or asymptomatic problem, potentially one of its great strengths, and the virus may suddenly be spread even faster and no one knows it’s happening, but it may or may not be a problem?
If so why would a plane say, be any more dangerous if a non vaccinated person got on board? If we need travel bans or quarantine requirements shouldn’t it still be the same for vaccinated people because they may have picked the virus up and pass it on? If we have restrictions to prevent new variants and we don’t know if they are vaccine resistant or not, presumably then the same restrictions need to apply to those vaccinated?
There are a couple of studies on airplane travel. Bottom line the risk of being infected on an airplane is no higher than pursuing your normal activities for an equal period of time. There seems to me to no scientific support for these certificates. These vaccines are coming online a lot faster than previous ones and the numbers on effectiveness could change a lot. We (the general public) are going to be participating in the trial that would normally come before approval of any new vaccine.
Paul –
> so I’m assuming (perhaps wrongly?) that with the vaccine you can still get the virus but 90% of the time fight it off before you get bad symptoms,
Yah. I’m not sure why you’re making that assumption.
> If so why would a plane say, be any more dangerous if a non vaccinated person got on board?
There aren’t guarantees until there are guarantees. The idea would be to minimize risk to the extent possible, when reasonable. If vaccines may prevent people from being infectious – and from what I’ve seen a fair amount of people who know about this stuff seem to think the probability is reasonably high – then you might be helping to shift the probabilities in a positive direction by requiring people to have vaccines to get onto a plane. It may turn out to be for naught. But you do what you can. Of course, if most people who know about this think that the probability is low, then the probabilities would be shifting.
Of course, the evaluation of “when reasonable” is a subjective one – but I know that I”d prefer being on a plane full of people who have been vaccinated than on a plane full of people haven’t – because there’s a chance that there would be fewer infectious people on board. And that’s even though there’s not a ton of evidence of infections being transmitted on board airplanes
My guess is I’m not the only one who might feel that way, and if so it would prolly make sense for airlines to require vaccinations as a way increase air travel. I would imagine that at the point that they determined that requiring vaccines is more an impediment than an encouragement, they would not want to require vaccines.
It’s rather like wearing a mask when people go to the supermarket. It’s possible that maybe one other person in the supermarket at a given time could be infectious, and it’s possible that if that person wears a mask they would be less likely to infect anyone else in the market. And then anyone who might have not gotten infected because that person wore a mask when passed next to them in the market won’t risk passing on the infection they got to their family members when they get home. And then all those family members won’t pass on an infection to anyone else they might have in their pods. Such is the nature of exponential growth in the context of the uncertainties.
Earlier this year, airlines made a friend produce evidence of having been tested. They didn’t need the results of the test, mind you, just that a test had taken place. Another one who flew said they had to have the negative results of a test, but the test could have been taken days ago and, of course, they were by definition traveling and therefore not locked down.
Our Covid protocols are political/scientific theater.
You will note that the “science” claims that sitting elbow to elbow with 200 strangers for hours indoors in an airplane is “safe.” This at the same time that “science” told my state to mandate 6-feet social distancing outdoors. (yes, they do, my state has signs up at parks and on beaches mandating 6-foot distance in the sun, outdoors).
Apparently the former is in part because there are “filters” on airplanes. Citing “science,” California rejects the notion that the same filters produce any safety in restaurants with 10-foot ceilings and six feet between people, but accepts the life-saving miracle of filters in airliners with 6 foot ceilings and six inches between people. Also masks, which inexplicably don’t prevent spread in grocery stores (15-foot ceilings and 10-foot social distance) but must in airplanes… because. Also masks because, naturally, nobody would think to eat or drink on a 10-hour flight, therefore mask wearing protocol is perfect.
Consider the damage from this. Restaurant and gym owners – already suffering – paid for and installed these filters because the experts said they work on airplanes. That’s when they discovered that both the “experts” and the politicians they advise reserve the right to use their own advice or not in a purely arbitrary fashion (except when it’s nakedly partisan such as during the protests over the summer).
You could not design a system of “expert” advice more perfectly organized to support autocracy, sow division, and build anger.
It would be a terrible business decision to require vaccination to board a flight. A large percentage in the US say they don’t want the vaccine. Limiting your customer pool that way is foolish.
Flying is safe now and it was safe before the pandemic. Chances of catching covid19 on a flight are not zero but comparable to an equal amount of time spent in your ordinary activities.
Why on earth would Judith delete my comment where I make fun of David, the king of appealing to authority, for acting as if he has more authority regarding the business decisions of airlines that airline executive do?
https://www.healthline.com/health-news/will-airlines-require-vaccine-passports-in-2021
Because evidence free and fact free drivel makes the comment threads unreadable. You are the king of that genre.
Vaccine passports are a virtue signaling requirement only. Air travel is already quite safe and additional requirements are unneeded.
Well if a lot of people “feel” as you do, they are not thinking clearly.
Flying would not become significantly safer if only the vaccinated could fly. That’s because its already very safe. Vaccination effectiveness is still uncertain because the trials are still continuing.
Narrowing your pool of customers to perhaps 50% of the population is a very foolish business decision. In the US a very significant percentage of people say they don’t want the vaccine.
Try thinking and using real inference and deduction. The facts don’t care about your feelings. No one here really cares about your feelings either as you are anonymous and exhibit disrespectful behavior toward Nic Lewis and myself among others.
Geoff Sherrington | January 11, 2021 at 8:25 pm | Reply
BTW, I deplore the actions of those here who are critical of Nic Lewis in a nasty, personal way. Nic is attempting to clarify a complicated piece of medical science that has some math features best handled by a mathematician.
niclewis | January 11, 2021 at 11:30 am | Reply
Without any attempt at explication of reasonable parameters, that statement could mean that “HIT” could be almost anything and thus “HIT” seems pretty much meaningless.
However, I have made such an attempt in my article. Why don’t you propose what you think are reasonable parameters if you disagree with my suggestions, rather than always just sniping unpleasantly?
SteveS | January 11, 2021 at 9:56 pm | Reply
I agree with Geoff…Joshua you are way off base with your attitude…It’s simple without getting your undies all bunched up… nick had a thought about transmission of the virus relating to heard immunity…. turns out that idea was probably not so good in hind sight….so what….I can’t find anyone, anywhere who has accurately predicted the path of the Pandemic. Would be nice if you took that little bit of “vengence” that pervades your on-line persona somewhere else. It’s a problem you have to work out on your own.
Intended as a response to Josh’s deleted comment. Suffice it to say Josh’s comment well deserved deletion.
Change in NPI during the pandemic in the US: https://jamanetwork.com/journals/jama/fullarticle/2775686?guestAccessKey=a604c991-4ba6-4afd-8723-19b1980f4cb3&utm_source=silverchair&utm_medium=email&utm_campaign=article_alert-jama&utm_content=olf&utm_term=012221
Overall decline but increase in mask wearing.
This is quite interesting. I think people are unconsciously or consciously treating mask wear as being highly protective. That was an attitude that was a concern when mask wear was first debated.
It’s easy to do. I’ve caught myself at it just a teeny bit, and I’m very careful, and well informed on the way this thing spreads (obviously, within the limits of knowledge available).
Ironic that after 9 months and many trillions of dollars spent on this pandemic and the economic fallout I still can’t find a genuine FDA certified N95 (w/behind the head straps) available over the counter. I can go online or drive 12 miles to a dental supply company but I still can’t pick one up at local stores like Walmart, CVS or Kroger. Come on man, it’s germ theory 101 for a respiratory diseases. I assume it’s the same world wide?
“Ironic that after 9 months and many trillions of dollars spent on this pandemic and the economic fallout I still can’t find a genuine FDA certified N95 (w/behind the head straps) available over the counter.”
I think that’s because those things are a pain to wear, and also because they are being made only for the medical profession. A friend who is a hospitalist posted a picture of his face after he took his off – big red marks from where the seal was against his skin. And, of course, he had to shave his beard off.
My sister is a physician and has to wear an n95 mask. She has had problems with facial rashes. Masks have side effects too. Hypoxia can lead to fainting which can lead to falls, concussions, or worse.
Mask requirements in community settings have very weak evidence supporting them. They are more about virtue signaling.
Meso- “I think people are unconsciously or consciously treating mask wear as being highly protective. That was an attitude that was a concern when mask wear was first debated.”
Meso – I know we have argued in the past about the effectiveness of masks, Though you make a very valid point which is similar to the main point I have tried to previously make. While I agree that masks are highly protective in a high risk environment, The difference is that “highly protective standard” doesnt provide much, if any, additional protection in a low risk environment.. In other words, if you are in an environment where there is little, if any, transmission of covid, then the incremental reduction in risk of transmission provides by wearing a mask is inconsequential.
“Meso – I know we have argued in the past about the effectiveness of masks, Though you make a very valid point which is similar to the main point I have tried to previously make. While I agree that masks are highly protective in a high risk environment, The difference is that “highly protective standard” doesnt provide much, if any, additional protection in a low risk environment.. In other words, if you are in an environment where there is little, if any, transmission of covid, then the incremental reduction in risk of transmission provides by wearing a mask is inconsequential.”
Which is why I don’t wear a mask in my home, or when out walking in the neighborhood. Although, in the latter case, given the nice weather in Arizona, I sometimes feel like a quarterback, dodging people who blithely will get too close to me. Even outdoors, an encounter within 6 feet does have a risk – not a big one.
As for how protective – typical masks are not highly protective, but they do help, and add to the other measures. From a public health standpoint, having people take more measures translates into a lower Rt, which means slower growth or more rapid decline of cases. I wear procedure masks – intermediate between N95 and a lot of cloth masks. And unfortunately, it’s really hard to tell which cloth masks on the market are effective and which are not – they don’t disclose much about them, unfortunately.
I like the graphic at the link below, for the concept. I know that you are already aware of these concepts, but it’s a nice graphic.
https://virologydownunder.com/wp-content/uploads/2020/12/SwissCheese-ver3.0_MUG-version.png#main
> In other words, if you are in an environment where there is little, if any, transmission of covid, then the incremental reduction in risk of transmission provides by wearing a mask is inconsequential
If an infectious person walks into a market where there are 150 people, and wearing a mask prevents an infection of just one other person in that supermarket, and that one non-infected person then doesn’t go home and infect any of her pod members, and each of those pod members doesn’t infect anyone else they’re in contact with, and none of those infected people infect anyone else, and so on….the reduction of risk isn’t inconsequential.
And suppose wearing a mask DIDN’T make any difference for that one trip to the market – which may be quite likely. That would be just one infectious person at the market for one limited time horizon, on just one day. In a locality where there is a lot of community spread, the exponential growth of just a very small incremental reduction can bake a very big difference in illnesses and lives, in pressure on jealthcare workers, and on the economy.
“I think that’s because those things are a pain to wear”
I don’t think that’s what’s happened. When I ask for a FDA N95 I always also ask if other people are asking to buy them too. Even the store employees have been saying the same thing for months “We can’t even order them, they don’t even have a stock # in my order system.”.
The FDA N95 masks I got at a dental supply has 20% more coverage than the simi-ridged shell/cup design and they have a small bendable foam sealing strip along the top edge. It works pretty good since I rarely have my glasses fog up.
I sympathize with the beard lovers but I guess we will just have to avoid being around them till the pandemic subsides. That put the burden back on me and you – wear your N95 mask and stay safe.
Josh – did you flunk math or the basic concept to marginal cost/marginal benefit or are just misapplying math out of an irrational fear.
You should be focused on closing the big leaks first. Instead you are focusing on every possible leak, no matter how trivial or remote, simply because any possibility of transmission, no matter how remote, justifies wearing a mask.
Your comment reflects an irrational and poor evaluation of the relative risks of transmission in various environments. I was in colorado hiking and your comments reflect someone that believes the mask mandate should be enforced on the mountain hiking trails or walking across the grocery store parking lots simply because the remote possibility of transmission does exist.
A good hint to your misunderstanding is the behavior of the so-called experts when they think they are not on camera.
> Instead you are focusing on every possible leak, no matter how trivial or remote,
The potential exponential growth from a small incremental risk is not trivial.
Further, for an individual decisiom the incremental change in risk might be small. If I don’t wear a mask on one particular trip to the market the impact on risk is small. But the small risks add up.
> You should be focused on closing the big leaks first.
I said nothing about prioritizing risks.
There no reason why you can’t focus on big risks and potential exponential growth of small incremental risks.
> I was in colorado hiking and your comments reflect someone that believes the mask mandate should be enforced on the mountain hiking trails or walking across the grocery store parking lots simply because the remote possibility of transmission does exist.
The question of enforcement becomes a matte if risk/reward ratio. Do you cause a backlash. Are there associated risks from wearing a mask. I personally wouldn’t wear a mask when honking on a trail unless it was heavily trafficked. I wouldn’t support a uniform mandate for wearing masks when outside irrespective of the conditions. But one problem with such policy scenarios is how to realistically design and implement them to provide for variation.
But sure, as for the supermarket scenario and the honing scenario.rhr balancing is quite different. There’s no inherent reason to treat them as the same. So for you to jump from the one to the other is a non-sequitur.
> A good hint to your misunderstanding is the behavior of the so-called experts when they think they are not on camera.
Because you can find some examples of experts who don’t follow official public health advice all the time is no reason to generalize unless you have some generalozable evidence. Outrage mining and extrapolating from unrepresentative sampling is inherently unscientific.
But even still, we know thst people often don’t act in ways that are consistent with cogent risk analysis for any variety of reasons. That doesn’t necessarily mean that the risk analysis is in error. Even if all “experts” weren’t following the best practices that wouldn’t mean they aren’t the best practices. Perhaps you think that experts are advising one set of practices and not acting accordingly because of some dishonesty thst is politically motivated. Well, that wouldn’t exactly be surprising to find someone here at Climate Etc. that has such a conspiratorial view of the world, without actually having evidence to support the conspiracy theory, now would it?
> I was in colorado hiking and your comments reflect someone that believes the mask mandate should be enforced on the mountain hiking trails or walking across the grocery store parking lots simply because the remote possibility of transmission does exist.
The question of enforcement becomes a matte if risk/reward ratio. Do you cause a backlash. Are there associated risks from wearing a mask. I personally wouldn’t wear a mask when honking on a trail unless it was heavily trafficked. I wouldn’t support a uniform mandate for wearing masks when outside irrespective of the conditions. But one problem with such policy scenarios is how to realistically design and implement them to provide for variation.
But sure, as for the supermarket scenario and the honing scenario.rhr balancing is quite different. There’s no inherent reason to treat them as the same. So for you to jump from the one to the other is a non-sequitur.
> A good hint to your misunderstanding is the behavior of the so-called experts when they think they are not on camera.
Because you can find some examples of experts who don’t follow official public health advice all the time is no reason to generalize unless you have some generalozable evidence. Outrage mining and extrapolating from unrepresentative sampling is inherently unscientific.
But even still, we know that people often don’t act in ways that are consistent with cogent risk analysis for any variety of reasons. That doesn’t necessarily mean that the risk analysis is in error. Even if all “experts” weren’t following the best practices that wouldn’t mean they aren’t the best practices. Perhaps you think that experts are advising one set of practices and not acting accordingly because of some dishonesty thst is politically motivated. Well, that wouldn’t exactly be surprising to find someone here that has such a c*nspiratorial view of the world, without actually having evidence to support the c*nspiracy theory, now would it?
But even still, we know that people often don’t act in ways that are consistent with c*ogent risk analysis for any variety of reasons. That doesn’t necessarily mean that the risk analysis is in error.
Even if all “experts” weren’t following the best practices that wouldn’t mean they aren’t the best practices. Perhaps you think that experts are advising one set of practices and not acting accordingly because of some dish*nesty that is politically motivated. Well, that wouldn’t exactly be surprising to find someone here that has such a c*nspiratorial view of the world, without actually having evidence to support the c*nspiracy theory, now would it?
my comment > You should be focused on closing the big leaks first.
Josh’s response – “I said nothing about prioritizing risks.”
“There no reason why you can’t focus on big risks and potential exponential growth of small incremental risks.”
Josh virtually every one of your responses you havent comprehended the concept of prioritizing risks. Your second sentence highlights you lack of comprehension of the subject matter.
> Josh virtually every one of your responses you havent comprehended the concept of prioritizing risks. Your second sentence highlights you lack of comprehension of the subject matter
I said nothing about prioritizing risks. There is no inherent reason why addressing risks of varying magnitude would be mutually exclusive. To the extent that they are in some situations (say because of limited resources), of course you prioritize the most significant risks.
There’s no reason that wearing masks to the supermarket prevents addressing more significant risks
The science is clear that masks don’t make any difference in community settings. Maybe Josh should have been wearing a mask 24/7 since birth because of the constant small risk of airborne pathogens. This last series of comments takes the cake for drivel and lack of common sense.
-snip-
Not all countries have this problem. Taiwan massively scaled up its manufacturing of masks at the start of 2020, such that by April every citizen received a fresh supply of high-quality masks each week, and the distribution system was regulated by the government. Taiwan’s COVID-19 death rate per capita is more than 1,000 times lower than that in the U.S. Hong Kong has been distributing patented six-layer masks (the efficacy of which has been laboratory tested) to every citizen. Singapore is on at least its fourth round of distributing free, reusable, multilayer masks with filters to everyone—even kids, who get kid-size ones. In Germany, Bavaria has just announced that it will be requiring higher-grade masks. If all of these places can do this, why can’t we?”
-snip-
https://www.theatlantic.com/health/archive/2021/01/why-arent-we-wearing-better-masks/617656/?utm_source=pocket-newtab
ACE inhibitors and receptor blockers in COVID-19:
https://jamanetwork.com/journals/jama/fullarticle/2775280?utm_source=silverchair&utm_campaign=jama_network&utm_content=covid_weekly_highlights&utm_medium=email
Discontinuing ACE inhibitors had no beneficial effect.
researchers project that, due to the pandemic deaths last year, life expectancy at birth for Americans will shorten by 1.13 years to 77.48 years, according to their study published Thursday in the Proceedings of the National Academy of Sciences.
That is the largest single-year decline in life expectancy in at least 40 years and is the lowest life expectancy estimated since 2003.
The declines in life expectancy are likely even starker among minority populations. For Blacks, the researchers project their life expectancy would shorten by 2.10 years to 72.78 years, and for Latinos, by 3.05 years to 78.77 years.
Whites are also impacted, but their projected decline is much smaller — 0.68 years — to a life expectancy of 77.84 years.
Overall, the gap in life expectancy between Blacks and whites is projected to widen by 40%, from 3.6 to more than 5 years — further evidence of the disease’s disparate impact on disadvantaged populations.
https://www.pnas.org/content/118/5/e2014746118
Latest stats from Sweden
https://lockdownsceptics.org/2021/01/24/latest-news-264/#swedens-per-capita-deaths-in-line-with-the-european-average-in-2020
tonyb
Thanks for the info:
“Will Jones has taken another look at the situation in Sweden. He finds that the country does indeed show that lockdowns aren’t needed…A number of states in America fit this description this winter, such as Florida, Texas, North Dakota and South Dakota.”
Nonsense. There is inadequate data to prove that – especially since nobody bothers to define “lockdowns.”
Look, people have cherry picked data from Sweden from the start to show that herd immunity was reached, or that no mitigations were needed, etc.
Guess what: Sweden has the most stringent mitigations it has had since the pandemic started. Lockdowns? Nope. Lockdowns only work if they lock pretty much everyone down, hard – like they do in China. No western country has tried that – our “lockdowns” are not that strict. Also, when people seek to tease out the impact of mandates, they usually ignore the fact that as the virus prevalence goes up, people *voluntarily* take stronger measures, in ways that are hard to measure.
So I don’t buy it. It’s the same old argument, hashed over yet again. Since the site is “lockdownskeptics” – i wouldn’t expect a balanced argument.
Hi Tony
A bit OT. I’ve been AWOL from these COVID19 threads for several months because it might take years to get definitive scientific results on some of these issues. But one question still intrigues me. How many died from COVID19 versus with COVID19.
It seems you were one of the first to raise this issue when the Lombardy Region in Italy was being hard hit last spring. As usual, the question became politicized and I don’t remember last year there being any resolution.
My interest was piqued again last week when I read an April, 2020 interview with the State of Illinois Health Department Director, in which she said if someone who died in Illinois and tested positive for COVID19, they were counted as a COVID19 death, which left unresolved what was the actual cause of death.
Based on your observations of these discussions have we come closer to answering that question?
Again, OT. For the first time since last summer, I read my state’s COVID19 data. It showed that 2/3 of the total COVID19 deaths were Long Term Facility residents. Of the 15,000 deaths 10,000 were listed as COVID19 deaths. It seems last Spring and Summer it was 30-40%. The higher number surprised me.
Ceresco
I have not been a frequent visitor either.
Our area has had a number of deaths –mostly very elderly people-we have a lot of Care homes. This level of deaths and infection fortunately remains low compared to the national average.
If you go here to our local news from Thursday evening 5th November, click on ‘Evening News’
https://www.bbc.co.uk/programmes/b006pfr1/episodes/player
It is confirmed the hospital is not overwhelmed (its busier now but still functioning) and Intriguingly there is an interview with Ian Currie, Medical Director of our local Hospital-speaking of the deaths. The interview with him follows immediately on from the start of the programme ;
The interviewer asks him if patients are dying OF covid or WITH covid;
“Many who have died are unfortunately elderly and have other medical conditions and are dying because of these other medical conditions WITH Covid and some patients are dying of Covid”
I think this is the first time i have heard anyone so senior say a number are dying WITH covid and some are dying OF covid.
Which brings us right to the heart of the matter of being able to officially differentiate between the two, as dying OF covid itself seems to be listed in the stats as the same as dying WITH Covid thereby inflating the numbers
Mr Currie also remarks that ‘people have contracted covid inside hospital, which is common to many other hospitals in the region.”
Good to see an honest admission that WITH and OF are not the same thing, and that hospitals themselves are not the places of safety we had expected, back in March.
The UK produces figures in a different manner to most other countries whereby anyone who has died within 28 days of testing positive is considered a death due to the virus. The average life expectancy is 80 whilst the average age of a virus death is 82 with several existing serious illnesses.
They have often gone into hospital either for something else or have caught CV in hospital. So even though they overwhelmingly have died of their pre existing conditions they are, for reasons beyond me, counted as a cv death.
So those who actually died solely of cv is much smaller than the overall figures seem to show. Some 380 people under 60 have died OF covid with no other underlying conditions.
This is not to minimise the virus-where the outcomes range from people not knowing they have it through to death, with the vast majority at the bottom of the scale, but flu –normally a huge factor-has been absent this year.
When adjusted for population and age our ‘excess deaths’ were worse than 2020 for every year from 1890 to 2008. They really need to get their stats in order, report in the same manner as most other countries, protect care homes and hospitals and stop scaring everyone.
Incidentally as with Italy and Sweden I understand that 40% of deaths in the first ‘wave’ occurred in Care homes as people were decanted into them from hospital, and up to 20% in hospital who didn’t have it when they went in.
All in all, if proper protection could be provided to the genuinely vulnerable- whether in care homes, hospitals, those in their own homes with serious medical conditions etc- then the rest of us could sensibly and safely go about our business and try to rescue the economy, education, our mental health, our jobs and finances, our personal health by proper exercising, and enjoy some normality by having a social and family life.
The idea that people should stay cooped up indoors rather than venture safely into the wide open spaces is beyond mad.
tonyb
Ceresco
I wanted to home in on your comment
“Again, OT. For the first time since last summer, I read my state’s COVID19 data. It showed that 2/3 of the total COVID19 deaths were Long Term Facility residents. Of the 15,000 deaths 10,000 were listed as COVID19 deaths. It seems last Spring and Summer it was 30-40%. The higher number surprised me.”
I remember posting a link last spring to an Italian report showing the huge number of care home deaths which were mirrored in the UK and Sweden. Perhaps in other countries also but I have not seen the stats.
They needed protection but instead in order to clear the decks for other incoming covid patients we decanted elderly people from the hospitals into care homes even though it turn out many were infected.
Without this action the deaths in the 3 countries named would have been far smaller and presumably a different policy would have been taken by the govt to protect those who most needed protecting and not lock down the fit and heathy
tonyb
Tony –
It seems that for Sweden you have the causal mechanism pretty much exactly backwards.
https://judithcurry.com/2021/01/10/covid-19-why-did-a-second-wave-occur-even-in-regions-hit-hard-by-the-first-wave/#comment-941350
From all the way back in May:
-snip-
Now, increasing numbers of workers are also coming forward to criticise regional healthcare authorities for protocols which they say discourage care home workers from sending residents into hospital, and prevent care home and nursing staff from administering oxygen without a doctor’s approval, either as part of acute or palliative (end-of-life) services.
https://www.bbc.com/news/amp/world-europe-52704836
Cesco – “My interest was piqued again last week when I read an April, 2020 interview with the State of Illinois Health Department Director, in which she said if someone who died in Illinois and tested positive for COVID19, they were counted as a COVID19 death, which left unresolved what was the actual cause of death.”
Let me partly respond to you comment.
Some individuals would argue that the final cause of death would be covid since it was covid that gave the final tiny push to cause the death. Others would argue that covid did not accelerate the date by any meaningful time frame.
Some individuals go into LTC because they need “long term” care. Others go into LTC for end of life care. The “average” life expectancy when entering a LTC is a little over 2 years. While the “median” life expectancy for a male is less than 4 months and less than 8 months for a female. Emphasis is on Median life expectancy, not average life expectancy.
For approximately half of LTC deaths ,Its hard to argue that death with covid was the actual cause of death.
Joe –
You might find this interesting:
https://www.nature.com/articles/s41598-021-81419-w
-snip-
Autumn COVID-19 surge dates in Europe correlated to latitudes, not to temperature-humidity, pointing to vitamin D as contributing factor
-snipg
The Vitamin D link is amplified in this paper.
https://virologyj.biomedcentral.com/articles/10.1186/1743-422X-5-29
Just another illustration of why R0 is a useless concept. Rt varies all over the place due to many many factors. There is no “normal” condition.
“No western country has tried that – our “lockdowns” are not that strict.” – mesocyclone
The UK is in it’s third lockdown with the end date legally extended to July 17th. Small and medium businesses are going bust by the day. It’s extremely distressing. The chief medical scientist keeps repeating that the UK strain is more deadly and more transmissible so everyone should stay indoors.
Trust me, it’s intense. Furlough will end and millions of people will likely become unemployed.
The Swedish model was the better option. The UK is still a LONG way from opening up again. It’s not looking good for the next 4 months..
“The UK is in it’s third lockdown with the end date legally extended to July 17th. Small and medium businesses are going bust by the day. It’s extremely distressing. The chief medical scientist keeps repeating that the UK strain is more deadly and more transmissible so everyone should stay indoors.”
Yes, pandemics really does suck. The chief medical scientist appears to be correct – the strain is clearly more transmissible, and apparently more deadly. Your emotional arguments are like railing against the sun for being too bright when you go outside. COVID19 doesn’t care about politics.
The argument about lockdowns has been hysterical and imprecise. “Lockdown” need to be defined, before asserted as good or bad, because there are so many variants. And, the people arguing mostly ignore the elephant in the room: voluntary measures which themselves are significant and hard to measure, and which also have economic implications.
It is obvious that lockdowns work – both from basic epidemiology (the “physics” of respiratory virus transmission) and from the experiences of China (which had a very severe first outbreak), and Australia (which has had lesser outbreaks, which it has defeated).
All this “lockdowns down’t work” ranting is really torturing the science, and using statistical arguments, based on very uncertain measurements, in favor of one approach or the other, in a Manichaean way.
Lockdowns work, if harsh enough. They are economically and psychological harmful. They are one of several measures, all of which involve either social distancing, masking or sanitation. And, they, like the other measures have different levels and styles of implementation.
Why not start from that, and avoid the weak statistical arguments which are just not convincing in either direction.
“The Swedish model was the better option. The UK is still a LONG way from opening up again. It’s not looking good for the next 4 months..”
We do not know that. There are way too many variables to assert that. Maybe it was the better option for Sweden, although it’s hard to tell given the lack of measured alternatives. Does that mean it would work in the UK or the US? No, it does not mean that.
Well meso, Feasible lockdowns have very limited evidence of effectiveness. Elsewhere I linked to a few papers on this subject. Locking everyone in their homes would work but cause millions of deaths from starvation, preventable illnesses, etc. OK, so say people can go to “essential businesses.” Then there will be transmission, especially between workers at these businesses. Is health care OK? If not there will be millions of future deaths. How about gyms? If not we will have a new obesity epidemic. We already have a suicide epidemic that directly correlates with lockdown. I posted a reference on this too. There is huge collatoral damage from mass unemployment.
The debate is mostly driven by fear and dishonest politicians. Basically, covid is not going to go away. There are many new strains now some of which are claimed to be worse than the old strains. Existing vaccines may or may not work on these new strains similarly to the flu.
In the US there seems little correlation between epidemic severity and severity of mitigation.
I personally worry that the response to this pandemic has been extreme and unhinged. Never before have the healthy been quarenteened in mass.
Few have even tried to quantify the massive damage to our future. Hyperinflation has historically been extremely harmful.
Tony
> I understand that 40% of deaths in the first ‘wave’ occurred in Care homes as people were decanted into them from hospital.
It seems with Sweden you have that almost exactly backwards?
https://www.google.com/amp/s/amp.france24.com/en/20200916-they-sacrificed-the-elderly-how-covid-19-spread-in-sweden-s-care-homes
Further to my comments above on 16th where I was looking at data for UK cases, adjusting raw cases, plotting graphs to see if there were any possible conclusions in there for herd immunity vs lockdowns bringing things under control, if immunity lasts, if it applies to new variants. Mostly people seem to be following the raw cases data which I suspect is not always showing the right trends or enough detail. I’ve also just started including the deaths data but not sure if what I’m seeing is anything new/different and everyone knows all this anyway.
I’m using the daily cases, number of tests and percent of tests positive. The UK data is good in having a consistent layout as well as enabling me to at least back calculate the value they use for the population in each area. Also I wanted to have a look at some places in US and Sweden in the same way. Unfortunately the problem seems to be getting all the data I need and in consistent formats and tables – recorded cases, number of tests, % of tests positive, and working out the population number they used.
I looked at Los Angeles county as I thought I’d heard there may be a new variant circulating there the last few weeks and I managed to find a good source of the data for there except I can’t decide whether the population value I need to use is 6.99 million based on looking at a graph that looked like around 14% cumulative cases but I can’t find the exact % of population number vs total cases and if I google the population I find it is 10.04 million.
Where the raw LA county data shows quite a big dip between about December 20th and January 4th and because testing dropped but % positive rose more sharply my adjustment equation converted that in to a nice continuation to and fall from a single peak on December 30th. The date seems interesting too as most of the UK locations I’ve been looking at seemed to also peak at exactly that same time (30th and 31st). And, depending which population value I use it peaked at 16.5% (and is currently at about 21%) or 13.6% (currently about 17%) cumulative, my adjusted numbers which potentially puts it in line with some of the UK locations I was thinking may have peaked around 16-18% due to a kind of herd immunity in the non adequately shielded group, or a bit lower if they hadn’t mixed so much at Christmas. On other hand so many places peaking on last 2 days of December probably suggests something else.
My adjustment equation is not adjusting the LA County data up nearly as much as the UK data, but LA seems to be doing a lot more testing than UK, so perhaps US figures are quite a bit closer to reality than UK numbers.
What I found most interesting was; due to a larger population LA county was the first place I tried plotting deaths on same graph and to my amazement, offsetting it 12 days it was an incredibly good match for the winter wave with my adjusted cases curve, and almost perfect match after a tweak to my adjustment equation: Poor correlation with the raw data, number of tests and % positive changing a lot, but run all three through a simple equation and I get a much better looking curve and it has incredibly good match to the deaths curve. Almost seems too good to be true making me concerned I’m doing something wrong.
An offset of 12 days between recording cases and deaths also seems too short, I thought it was more like 17 to 21 days? Maybe there’s quite a few days lost in the data?
I’ve just done same for England, not quite as good as the LA result, it is still showing quite a lot of agreement between deaths and adjusted cases, with same offset of 12 days, and I’m wondering if the discrepancies are due to what I’m calling a ski jumping effect where deaths follows the (adjusted) cases curve incredibly closely as they rise, but then stays ‘airborne’ after passing the peak roughly about 30 days delay. Is that perhaps dependent on how we classify deaths due to or with covid and the cut off time. For LA the ski jump had almost landed from the summer wave before latest wave, but in England we had a wave in October/November too close to the December wave for the ski jump to fall much.
However for both LA and England there appears to be no ski jumping after the end of the December peak, i.e. it is keeping firmly on the ground, if real I wonder what that means. But I suspect actually it’s too early and more data has yet to be retrospectively added. I’m also trying to figure out if I can use this to plot mortality against time, but the ski jump effect is complicating it.
I’m finding this very interesting and potentially seems a very useful technique to get a lot of useful information out of the data urgently before the vaccine starts confusing the issue. Or is everyone already doing this kind of approach and much more advanced; I would have thought they must be?
Paul
Do you read the UK site, lockdown sceptics? Variable articles but some very good and informative. As you know, here in the UK cases had reduced before the first 2 lockdowns and this was admitted by whitty and johnson in press conferences.
The same thing has happened in the third lockdown with as near as we can judge the peak being very much the last day of Dec or the first ones of January as you mention.
If lockdowns worked we wouldn’t keep needing them. Lets hope the vaccine works else this could go on indefinitely as the govt cant admit they chose the wrong policy
tonyb
Tony –
> If lockdowns worked we wouldn’t keep needing them.
Unnecessarily binary.
NPIs might have a limited value. The definition of “lockdown” and “worked” need specificity to be meaningfully interpreted.
> here in the UK cases had reduced before the first 2 lockdowns and this was admitted by whitty and johnson in press conferences.
Asusuming that claim stands up to scrutiny, the problem is that remains an unsupported counterfactual assumption.
NPIs could create a differential longitudinal effect even if a downward trend in infections preceded the implementation of interventions.
Tony –
> The same thing has happened in the third lockdown with as near as we can judge the peak being very much the last day of Dec or the first ones of January as you mention.
What is your reasoning on that? Looking at Worldometers, the peak in infections is about Jan 10th. And the shape of the trend shows dramatic rise up until the 10th and dramatic decline thereafter.
Tony –
Further – looks to me like hospitalizations and hospital admissions follow a pattern consistent with a very strong dampening effect from the initiation of the NPIs:
https://ourworldindata.org/local-covid-uk
joshua
Those words are straight our of the mouths of our chief medical officer and Our Prime Minister given at press conferences.
You will be familiar with Einsteins words on the definition of failure
tonyb
Thanks Tony I wasn’t aware of that website.
Joshua, the point I’ve been trying to make is the raw data may not show the true picture.
I’ve just done a quick google and looked at what I presume is a graph for whole world – my quick reading is it shows a peak on 24th December, a drop then another peak on about 10th January as you say. Now that may be a good representation of reality, or is it the case that the amount of testing dropped a lot during those two peaks as it did in many of the locations I’ve looked at, and % of tests showing positive went up at same time? I don’t know I haven’t got the data. I don’t know if the same applies to the “whole world”, but if it does then accounting for a drop in testing and increase in % positive perhaps it wasn’t really a dip during the holiday period, but an increase to a peak around 30th/31st?
>here in the UK cases had reduced before the first 2 lockdowns and this was admitted by whitty and johnson in press conferences.
>Asusuming that claim stands up to scrutiny
It doesnt.
Neither, of course does “lock down sceptics”.
But really, life is too short.
Paul –
> Joshua, the point I’ve been trying to make is the raw data may not show the true picture.
Sure. A of positive tests in proportion to tests performed is the key number.
But I think that the pattern all over the world of inctions rising an falling in close correlation with interventions, or social patterns that parallel the outcomes of interventions amp official mandates (such as reduction in mobility in NYC prior to SIPs being issued) is pretty compelling.
As to whether in various countries you’d be able to manifest those social patterns without mandates is another matter. Also. As to whether there are associated costs with mandates is another matter (I see a lot of claims made in that regard but from what I’ve seem they all look like specious conflation of correlation with causation because they assume counterfactuals that tie negative impact to mandates without being able to distinguish the impact of the pandemic itself).
BTW. I wrote this comment recently to Frank in continuation of our discussion about whether there an evident signal of “herd immunity.”
https://judithcurry.com/2021/01/10/covid-19-why-did-a-second-wave-occur-even-in-regions-hit-hard-by-the-first-wave/#comment-941289
“The definition of “lockdown” and “worked” need specificity to be meaningfully interpreted.”
Classic progressive- if what I said turns out to be untrue, we need to change the meaning of the words in the sentence I said.
The only thing funnier than that line is JackSmith’s claim that he can’t get an N95 mask. Well, except all the ones available at THAT store, but certainly not the ones at THIS store. And, of course ignoring the fact that Fauci says any “face covering” works, at least when he’s not saying that only N95’s work, or when he’s not saying that all masks are pointless for the general public to wear. But other than that, N95s are “available” if the political whim o’ tha day requires it and “not available” if it doesn’t. Just like “lockdowns” will “work” or even exist at all depending on momentary needs of partisan activists on MSNBC panels.
There is an even funnier example of this right now. Joe Biden, who has been credibly accused of assault by more than one woman, has issued an edict vowing “zero tolerance for assault” in his administration. And not a soul even hesitates in shame. Because, to paraphrase Joshua, “the definitions” of assault and administration “need specificity to be meaningfully interpreted.”
Jeff –
> Classic progressive-
Yeah. That’s why meso made the same (obvious) point about the definition of “lockdown.” Because he’s such a “progressive.”
Good point. Lol.
An easy way to plot deaths and cases on the same graph is to use 91-divoc.com. Very customizable graphs. For a lag, just eyeball it.
And another one for our resident squirrels about the very real human costs of locking people in their houses.
https://adc.bmj.com/content/early/2020/06/30/archdischild-2020-319872
What is amazing is how easy this kind of thing is to find. Perhaps its a case of people allowing their biases and partisanship determine their conclusions without bothering to do any research into the science.
people allowing their biases and partisanship determine their conclusions without bothering to do any research into the science.
Very apposite.
David –
lol. You cherry-pick some studies in your rhetorical “research,” where you obviously skip over ANY that don’t confirm your bias (there are many out there that run the other direction so your bias confirmation is obvious) to link some preprints, one by an ocranographer from 9 months ago who says that police enforced home confinement might not be terribly effective, another that says the most strict interventions might not always be a good idea, etc.
Yes, some well-credentialed researchers have published studies that indicated the vague notion of “lockdowns” are not a slam dunk – but anyone who claims that the body of research as a whole is even remotely conclusive either isn’t really paying attention or is just overwhelmed by their confirmation bias.
I certainly think that there’s a lot of uncertainty, especially since people start out with different notions of what a “lockdown” means and all the myriad confounds across different countries with different definitions and interventions implemented and different levels of compliance and different precipitating conditions when interventions are implemented, etc. That’s why I stress that this is best approached as a matter of decision-making under conditions of extreme uncertainty until, maybe, possibly we get a lot more high-quality evidence.
Any reasonably comprehensive literature review, of the type one would expect in an academic approach, would addeess the ENTIRE field of reseaech.
It’s not clear why you bother with that kind of charade and call it “reseaech” just to repeat the same insults and appeal to authority that you’ve stated so many times in this thread. It’s not necessary – you could just repeat the insults.
I suppose it’s because you’re very motivated by the many times that I’ve pointed out the obvious errors thsr you’ve made, and documented them.
And that’s your right – but as someone who thing is of you fondly I suggest to you that a sloppy approach to reviewing the literature certainly doesn’t mitigate the obvious mistakes that yicd made and would likely just make the situation worse to anyone open- minded who’s paying attention.
For me the biggest issues are (1) the vagueness of the starting definitions of how people are approaching the analysis. As I think meso has discussed, people are disagreeing violently without even agreeing what it is that they’re disagreeing about. For me, that’s a HUGE red flag for motivated reasoning – in this case ideologically motivated reasoning. I think it’s highly unlikely that the strongly predictive signal of the ideological predisposition of the analysts views on the efficacy of NPIs is just purely coincidental. Nor do I think it is determined by one side or the other employing categorically superior science.
(2) The lack of consideration of the relationship between the precipitating conditions and the stringency of the interventions – of course one would expect, generally, more severe interventions where the pandemic was most out of control, and thus you can’t just compare outcomes as if the precipitating conditions weren’t predictive of severity of outcomes independent of the interventions implemented, and
(3) the lack of attention to how people are working from largely unsupported counterfactual presumptions about what would have happened absent interventions,
(4) the willingness to compare across vastly different conditions to asses the efficacy of interventions without controlling for confounding variables
(5) the lack of good cluster analyses to help look at interactions and mediation/moderation effects between and among different interventions.
I think that basically this whole situation is one big mess.
I started out with thinking I had two biggest issues but couldn’t get out without 5, and if I thought about it more I’m sure I’d come up with more even bigger ones.
Of course, if you’d like me to David, I’d be happy to think about it more and share my thoughts with you. I always so look forward to your carefully thought out feedback.
https://ourworldindata.org/covid-vaccinations
Josh, This comment shows your deep bias and motivated reasoning. There is little difference in the ramp up of vaccinations between the US and the UK. There’s no evidence Biden has done anything to accelerate it.
David –
I didn’t say that Biden has done anything to accelerate the rates of vaccination.
I said that he only recently has indicated an expectation of increasing the rate beyond where it was during the Trump administration. I would consider anything less than a significant increase to be disappointment, if not a failure. If only because you’d expect them to get better at it as the go along
I consider it to be a failure of government that the Trump administration didn’t bring more federal resources to bear in the distribution of vaccines. If Biden fails to do so, I would consider that to be a failure as well.
You may choose to use distribution in other countries that have done relatively poorly as a benchmark for judging the US. On the other hand, I think we should do far better than most other countries and should actually be the best if not close to the very best (in terms of absolute metrics, not relative metrics) – given the level f resources we have to bring to the table.
> There is little difference in the ramp up of vaccinations between the US and the UK.
Interesting that @35% worse than the UK (by that chart) – a country with considerably fewer resources to bring to bear – is “little difference.”
J
There is 2 step process in getting people vaccinated. Production and distribution. What makes you think the problem is distribution?
Kid –
> What makes you think the problem is distribution?
Fair question and actually I don’t know. I would imagine Thera some of both. I was using “distribution” to mean end point of care distribution. I don’t know what explains why it wasn’t better this far. That said, there is this…
https://twitter.com/ezraklein/status/1354105105688064001?s=20
Where someone with a vested interest is interviewed… so I’ll take it with a grain of salt…
Particularly hopeful is there may be signs that they’ll ramp up rapid antigen testing.
Note that he says that supply isn’t the primary chokepoint. Take that for what it’s worth.
Josh, lets review.
1. I link 5 papers finding pretty much the same thing as the latest Ioannidis paper on NPI’s and their effectiveness.
2. Josh in typical rambling evidence free fashion asserts that other papers say other things. I am well aware of this and dealt with it with one of the linked papers.
3. Josh does not present any evidence whatsoever.
4. One of my links is a devastating critique of Flaxman’s paper purporting a huge effect of lockdowns. It’s quite conclusive in demonstrating that Flaxman’s conclusion is just obviously wrong.
5. Earlier, Josh quotes Gellman who relies heavily on Flaxman’s expert opinion on lockdowns. Other than that, Gillman has no substantive criticisms of Ioannidis’ et al latest paper.
6. Josh spends many paragraphs dishing out vagaries about bias even though he has no idea what that means in real scientific terms.
It’s an embarrassing waste of time and you should have a higher regard for yourself. The most obvious explanation of this is that Josh has a predetermined narrative (both about science and about a few scientists) and doesn’t have the intellectual discipline to look critically at that narrative. You are not “fond” of me. You disrespect me and other denizens by cluttering up Judith’s comment threads.
David –
As I said, an academic literature review surveys and cites the entire body of research. You referenced only studies that support one side of the debate. There are many that run counter to that. And some of the papers you cited are rather limited in relevance anyway, such as the preprint from 9 months ago, written by an oceanographer – that made a fairly obscure conclusion that police enforced home confinement might have limited benefit.
I think that there is a lot of uncertainty regarding the efficacy if NPIs. But to have a useful discussion, I think one should start with a clear definition of what it is you’re discussing when you talk of a “lockdown.” If you’re talking about police-enforced home confinement – as that one paper was examining to conclude limited value, then the application is quite limited regardless of the merits of the analysis. Of what use is such an examination w/r/t the context of the vast number of cases where NPIs have been implemented, and where no such intervention was even considered?
I listed a few other considerations that I think are important. If you think that simply citing studies that come to conclusions on one side of the debate, and ignoring any studies that support different conclusions, and appealing to the authority of some of the authors while ignoring the lack of authority of the others, is more interesting than discussing the context in which the different findings might apply, that’s certainly your right. Since you’re so much smarter than I, and a skilled scientist, obviously your antipathy-filled appeals to authority are of much greater value than any thoughts I could muster. Nonetheless, because your typical comment is so chock-full of sage insight, I offer my thoughts on the hope that maybe you could steer me on the right path should since obviously my thinking must be in error.
At any rate. David, maybe we should focus on the positive: It looks like the infection rate in the US is starting to drop, and we can expect the average number of deaths (and hospitalizations) to follow. Irrespective of what we think about the likely efficacy of interventions, I’m sure we can agree that’s a relatively positive development.
I will note this – Biden only recently started taking responsibility for significantly increasing the rate of vaccination beyond what was happening under Trump’s administration.
If he does reach his more recently stated goal, that would be a marginal improvement. Not sufficient, IMO, however, and if he doesn’t do better still than that I would consider it to be a disappointment bordering on the failures of the Trump administration.
dpy,
Do try to avoid further childish insults. You demean only yourself.
You have claimed that the near total suppression of influenza this year is unrelated to covid suppression.
You have claimed that anywhere with declining covid infections has reached herd immunity, regardless of the level of immunity in the population.
You claim that herd immunity can be reached and then turned around repeatedly during the course of a year.
These are very obviously unscientific and frankly bizarre assertions. Repeated attempts to point this out with references have been met with insults and obfuscation.
Rational discourse with you is entirely impossible.
Just to recap since this “interaction” is replicated many many times on the internet including many times here with the same suspects. There are two of these types who have a massive comment footprint on this post.
Anonymous nonscientist sees something by a real scientist that is an interesting contribution but that might harm his narratives. Activist joins conversation.
1. He quibbles about definitions, provides no evidence, and denies the citations of the scientist from other scientists. He prefers more simplistic definitions.
2. Since the activist doesn’t do real science or math he is reduced to proof texting various pseudo-scientific sources such as that leading scientific journal, Twitter.
3. He then goes on the attack a lion of the medical field with vague falsehoods and other rhetorical devices to cast doubt on an excellent body of work.
4. Then he shows his utter ignorance of statistics by stating that doing a statistical analysis involving data from many different counties means the authors are saying they are equivalent.
5. Activist ignores real scientific analysis of HIT, R0, R and whether these are useful concepts or not.
6. Activist distorts what scientist says by claiming he said things he never said.
It’s pathetic, but its a pattern that repeats itself thousands on the internet.
Hi dpy.
I see your previous reply was moderated. I doubt this one will stay either, given the content. I suggest you reflect on why that is.
Influenza remains an infectious disease suppressed by covid measures.
The UK remains under the protection of herd immunity by your definition, despite needing a hard lock down to prevent growth of the pandemic.
Countries and regions continue on a near daily basis to enter and leave herd immunity by your definitions.
Your leonine hero’s output on COVID, as you are well aware, has been the subject of highly critical responses from the scientific community, and much of it has been overturned by subsequent events.
But really, you just need to take your own advice. I commend it to you.
Another content free rhetorical comment. Why waste time on this?
I have cited support for everything I have said here. You haven’t cited much aside from Twitter. Yet you claimed you did.
You also claimed I said things I did not say. You claimed that “You have claimed that the near total suppression of influenza this year is unrelated to covid suppression.” I never said that. I said no one knows why influenza “disappeared.”
It’s not really possible to have a conversation in which facts are manufactured and rhetoric dominates your responses.
Some of Ioannidis’ covid body of work has been criticized especially the Santa Clara study and they published a very careful further analysis using more data on the test used to refine their confidence intervals. Have you read that? I thought not.
But its a controversial area so disagreement is normal. It’s called science as perhaps you are unaware how that works. Certainly Ioannidis’ work has fared better than Flaxman’s really obviously wrong paper on the effect of lockdown in England and Ferguson’s early work. Both were just almost scientific malpractice.
https://www.frontiersin.org/articles/10.3389/fmed.2020.580361/full?&utm_source=Email_to_authors_&utm_medium=Email&utm_content=T1_11.5e1_author&utm_campaign=Email_publication&field=&journalName=Frontiers_in_Medicine&id=580361
If you have a substantive response I’d entertain it. I’m not going to respond to content free vague and purely rhetorical responses.
dpy,
Claiming a citation supports you does not make it so.
Your attempt to obfuscate on the near total suppression of seasonal flu by covid suppression measures is noted. It lacks any citation to support it, of course,
Your continued belief that anywhere with reducing prevalence is in “herd immunity” is noted, and of course is entirely unsupported by any citation.
Your extremely odd position that herd immunity can be entered and exited multiple times is, of course, also entirely unsupported by any citation.
To end on a positive, you have at least managed to control your unprofessional tendency to lace your posts with insults.
Another content and evidence free response from vtg.
You did it again. You claimed I didn’t cite anything on the flu. I cited a very careful and lengthy paper pointing out how little we really know. Certainly your assertion on it is worthless without some scientific support. Just to repeat a third time the citation:
https://virologyj.biomedcentral.com/articles/10.1186/1743-422X-5-29
It’s not possible to have a real conversation with someone who can’t remember even the most basic facts of the previous statements (or perhaps deliberately distorts them).
I hope you realize that anyone reading this thread will immediately see the difference between activist rhetoric and science.
For those who have not read or can’t read and understand the paper it offers strong support for the idea that we really know little about the flu beyond crude mechanistic narratives with little real scientific support.
But I forgot, activist anonymous nonscientists are to be believed over real scientists.
You claimed I didn’t cite anything on the flu.
No dpy, I claimed you cited nothing which supported your position, that being that the near elimination of seasonal flu is not related to covid suppression. That paper does not support your position, which is entirely incredible.
Likewise, you have provided nothing to support your other unscientific assertions:
(1) that anywhere where infections are reducing is in herd immunity
(2) that poopulations can come repeatedly in and out of herd immunity.
As to “activist rhetoric”, I suggest you read your own postings.
On Ioannidis, I suggest you read, more carefully this time, the previously cited critique of his work on covid.
dpy,
I have read all of your references. They do not support your positions, and in most cases actively contradict them.
If your position on seasonal flu elimination is not what I stated, you have a very simple remedy: state what your position is, without obfuscation. I shall await it with interest. The paper you cited does not support your position that it is not related to covid suppression, however often you repeat it.
If you dislike rhetoric, I suggest your use of “lion of the medical sciences” might be a teensy bit unwise, no?
As to ethics, actually read the paper on Ioannidis contributions. There is a relevant passage.
And you’re back to unprofessional insults again. Disappointing, but unsurprising alas. Try again?
VTG –
> As to ethics, actually read the paper on Ioannidis contributions. There is a relevant passage.
The recruitment of participants for the Santa Clara study by Bhattacharya’s wife on Facebook, promising the potential of “immunity passorts” was bad enough. I can’t believe that it could have passed IRB review.
But much worse, IMO, was where they tested people without fully informing them about the problem with false positives, and allowing them to go home to infect grandma. And then rejected the advice of their colleagues to re-test.
No wonder those scientists who were originally participating in the study quit in protest, and why one of them lodged a whisteblower complaint.
I don’t think you have understood any of my references. That’s the most likely explanation for why you have quoted nothing from them. I’ve been much more detailed.
I regret that you have such a poor memory. I just restated in the previous comment for the 3rd or 4th time my position on the flu. But again for the benefit of anyone else foolish to waste time on this. No one really knows why the flu “disappeared.” You provided no evidence that covid mitigation had anything to do with it.
Children often interpret truth telling as an insult. Adults will be more logical and try to refute the statement. Repeating many many times an obvious falsehood is bad faith.
You vaguely cite a “telling passage” from some paper or another. That’s a purely rhetorical device and typical of an activist who is acting as a merchant of doubt.
Since you make a highly unlikely claim, what do you think of the Vitamin D hypothesis for the flu? No vague rhetorical responses please.
dpy,
Your continued obfuscation on flu is telling.
There has never been a year for flu like the last one, and it is transparently obvious that covid suppression measures have also suppressed flu – actually more so due to its lower Ro.
The extent to which vitamin D normally influences flu I have no opinion on, I have not researched the subject.
Now, please: your opinion on the correlation of flu near elimination and covid. No obfuscation. A clear answer.
Well at least you have stopped peddling obvious untruths about what I said over and over again.
This is so typical of you VTG. You ask simple minded questions which are based on your exceptionally superficial understanding of science and media partisan narratives. It’s another rhetorical device in the merchants of doubt playbook.
There are lots of alternative explanations for flu “disappearing.” One that I give some credence to is that due to the health care system being overloaded and the number of covid cases that flu testing has dropped off dramatically. There has been a massive focus on covid testing.
Testing for the flu is always very spotty and the flu numbers you are naively believing are based on models derived from very limited samples. CDC flu numbers are quite unreliable.
Summarizing: Your question is based on ignorance and a gullibility about what you read in the media. I’m so sure you read media with a variety of points of view, being an activist merchant of doubt and all.
dpy,
You can’t bring yourself to admit that suppressing covid has also suppressed flu, worldwide (Clue: the CDC is not the only organisation in the world to test for flu).
You can’t write a post without insulting other people, whilst simultaneously trying to claim the to be a scientist(!).
It’s remarkable to observe.
Whatever next?
VTG
I haven’t really been following this part of the conversation but surely extensive mask wearing and keeping your distance from possibly infected people and not being inside crowded public places would have a dramatic effect on curtailing flu?
https://www.health.com/condition/cold-flu-sinus/surgical-mask-flu-prevention
Somewhat wistfully the link seems to be from a year ago today and only makes a passing reference to cv.
Tonyb
Tony –
> I haven’t really been following this part of the conversation but surely extensive mask wearing and keeping your distance from possibly infected people and not being inside crowded public places would have a dramatic effect on curtailing flu?
Wow. Who could possibly thought of that? And I thought it was only because medical practioners never test anyone for the flu anymore – even people who have flu-like symptoms but test negative for COVID.
I mean David made an argument like that and he cherry-picks scientific literate and…well…science.
And authority.
And lions.
…surely extensive mask wearing and keeping your distance from possibly infected people and not being inside crowded public places would have a dramatic effect on curtailing flu?
One would certainly think so, and the evidence overwhelmingly supports it.
Apparently, though such views are, alas
…based on ignorance and a gullibility about what you read in the media
Hey ho.
Just for the record since our local anonymous activist nonscientists seem to have a compulsion to denigrate Ioannidis’ work I’ll review some well known facts. These two have fixated on old and now irrelevant issues in a generally magnificent opus of scientific work. They don’t care about vastly more egregious errors from Flaxman and Ferguson that had vastly bigger implications.
1. The Santa Clara study was not perfect. In response to criticism, they addressed some of the issues in later publications that I don’t think Josh or VTG are even aware of.
2. There are by now numerous other studies that show pretty much the same IFR for the early epidemic. Two such were Los Angeles county and Miami-Dade county.
3. Ioannidis himself has a fairly recent paper surveying all the serological studies worldwide.
4. Ioannidis has at least a dozen papers since the start of the pandemic this spring. That’s an impressive volume of work.
The real truth is that the IFR is highly variable depending on many variables. Simple minded fixations on a single universal number are as misguided as thinking that R0 is meaningful in any practical sense.
Perhaps the most important point here is the obvious truth that there is no science beyond crude mechanistic narratives that fully explains most of the characteristics of flu epidemics.
And since its been misrepresented by VTG several times here, what I actually said is that no one really knows why flu seems to have “disappeared” in 2020. There has been nothing here that calls that into question.
Re: “2. There are by now numerous other studies that show pretty much the same IFR for the early epidemic. Two such were Los Angeles county and Miami-Dade county.”
Nope. For example, the Miami-Dade study had lower seroprevalence than initially reported, and thus entails a higher IFR than you claimed. You’ve been corrected on this multiple times, dpy, but keep repeating your falsehood anyway. How telling.
Old and debunked, April 24:
“Our data from this week and last tell a very similar story. In both weeks, 6% of participants tested positive for COVID-19 antibodies, which equates to 165,000 Miami-Dade County residents.”
https://www.miamidade.gov/releases/2020-04-24-sample-testing-results.asp
Newer update:
2.2% – 2.8% seroprevalence
http://www.sparkc.info/
https://twitter.com/AtomsksSanakan/status/1287606941863354368
A little over a year into this pandemic and there are over 700,000+ daily chances (infections) for SAR-COV-2 evolve into a more deadly strain or become vaccine resistant. Let’s hope these new designer vaccines can keep up with the mutations.
https://nextstrain.org/ncov/global
Must be fun to model?
Where’s Ragnaar? Witness protection?
https://ourworldindata.org/grapher/daily-covid-deaths-7-day
Joshua
Interested to get your reaction to this article which claims diet has been a key factor for covid deaths as I guess the comments on sugar and obesity apply to both our countries
https://www.conservativewoman.co.uk/instead-of-lockdown-cut-sugar-and-hand-out-vitamin-d/
The author is a former pharmacist and Olympic medallist so knows about diet
Tonyb
https://judithcurry.com/2021/01/10/covid-19-why-did-a-second-wave-occur-even-in-regions-hit-hard-by-the-first-wave/#comment-941599
re Ivermectin – I hope you’ve read Frank’s useful comments at Climate Etc.
Also, someone just posted this on another blog:
There was good reason for initial skepticism on ivermectin. In contrast to some of the other repurposed drug candidates, ivermectin has a well-understood and very specific action (in its usual use), blocking glutamate-gated chloride channels that arthropods and worms have but vertebrates don’t. This selective action gives it an incredible therapeutic index, but would also make one a bit dubious that it would work anywhere near as well on coronaviruses as on worms. On top of that, the first published in vitro experiments showing effects against SARS-2-CoV were at much higher concentrations than the usual treatments would achieve in humans (orders of magnitude higher) but still were being quoted as supportive, which is a bit of a red flag. Derek Lowe wrote about it here:
https://blogs.sciencemag.org/pipeline/archives/2020/05/11/whats-up-with-ivermectin
Now, just because there were reasons to be dubious doesn’t mean it *couldn’t* work in real life, and maybe it does, but the doubts were at least reasonable at the time.
———-
https://statmodeling.stat.columbia.edu/2021/01/26/three-coronavirus-quickies/#comment-1680845
Tony –
> “If the health authorities had used the viral epidemic to advise everyone to avoid ultra-processed food, sugar and refined carbohydrates they could have avoided thousands of Covid deaths ”
There certainly seems to be much evidence that metabolic diseases are associated with risk for mortality from COVID. Anything that public health officials can do to reduce obesity would obviously have huge social and economic benefit.
On the other hand, the notion that “advice” from public health officials would significantly alter the prevalence of metabolic disease seems to me to be very simplistic and naive. The idea that it would do so on a time frame to be reflected in COVID morbidity and mortality outcomes seems to me to be EXTREMELY implausible.
> “and simultaneously reduced the burden of metabolic diseases, from which millions suffer.”
Once again, advice from public health officials might have a marginal effect in that regard, but tackling the public health problem of metabolic diseases seems to me to be much, much more involved than a matter of public health officials offering advice.
> “There is no good evidence that lockdowns work in the long run. ”
Hmm. Not sure what “long run” means there. I think it’s reasonable to speculate that interventions can reduce the rate of spread of the disease.
That allows for time for the development of better therapeutics, and thus outcomes per infection can improve. But with the development of vaccines, then the whole ball game changes – and the definition of “long run” becomes extremely important. Take a situation like Sweden vs. Finland. If infections and deaths ended tomorrow in Sweden, it would take years, maybe decades, for Finland to reach the same level of COVID population fatality as Sweden, but with vaccines it would likely NEVER happen. So if the difference in rate of infection and death can largely be attributed to interventions, then the statement I highlighted above looks quite dubious w/r/t “long run” benefit.
Of course, there’s a lot of uncertainty there. How do we know if it is actually the interventions that explain the different levels of morbidity and mortality in Finland compared to Sweden (close to an order of magnitude difference)?
I don’t think that we can, but neither do I think that we can say it isn’t a reasonable possibility.
> “However, there is clear evidence that prolonged lockdowns make the entire population more susceptible to severe symptoms when they catch a respiratory virus. American Professor Sheldon Cohen has worked for decades on the relationship between prolonged stress and more severe viral infections. He regards chronic social isolation and fear as the most potent stressors for damaging the immune system. Inevitably, lockdowns increase social isolation and fear. Perhaps the new coronavirus variants are not more transmissible; they just seem to be so because now we are all more susceptible.”
The problem here is that it takes “lockdowns” out of context. The context is a raging pandemic. Is there significantly more of the problems from chronic social isolation in Finland than there is in Sweden, since the onset of the pandemic? I doubt it. Perhaps we wouldn’t be able to evaluate for quite a few years yet, but certainly that limitation would apply to your conservative woman just as it would to everyone else.
The problem here is that people are trying to make these interpretations in a logically problematic manner. They’re taking gross-level correlations and pinning causation onto them – with sufficient evidence to do so. They’re relying on counterfactual assumptions – that things wouldn’t have been worse absent interventions than they are now. Or that many of the same problems from depression and unemployment and social isolation and etc., wouldn’t have manifest with a more out of control pandemic. We don’t know that. Is it possible? I can’t say. But I don’t think that anyone else can either. So when I see rhetoric like that, I think that I see someone who isn’t sufficiently respecting the uncertainties in play.
> “No harm-to-benefit analysis has been performed for lockdowns, which demonstrates staggering ineptitude and negligence. ”
That seems to me like an odd statement. People are attempting to perform such analyses. But it’s a remarkably difficult task.
> “Other researchers, however, have done this work and state that any benefit does not warrant the harm and lockdowns should be eased. ”
I dunno Tony, that seems like a very unscientific pronouncement to me.
There seems to be a lot of evidence in support of using vitamin D as a treatment for COVID. As to whether using it as a prophylactic or even as a therapeutic would significantly alter the societal trajectory of morbidity and mortality from COVID seems to me to be highly uncertain.
I don’t know what to think about Ivermectin, frankly. I tend to be reflexively skeptical about claims of miracle cures. I’m also inherently skeptical of grand conspiracy theories such as that the medical establishment is irresponsibly ignoring or hiding a cheap and effective miracle treatment. But I’m certainly not in a position to say whether it is an effective treatment. There seem to me to be some problems with just using it on a large scale without carefully controlled studies. But sure, there’s a pressing issue w/r/t how to proceed during an ongoing large-scale medical emergency with treatments that have anecdotal evidence of success, provided by reasonably credible sources.
Joshua
Thanks. I think you have the same body positive movement over there that we do, whereby over the last 5 years or more being obese has been seen as a valid lifestyle choice and celebrated by some and it cant be criticised or else you are guilty of all sorts of “isms”.
I think there was sufficient time between being overweight, unfit and lacking Vitamin D as known health factors in April 2020, and the second wave in October of that year for many people to have done something about it.
Whether that would have saved people I don’t know as the effects of such conditions cant be immediately overcome. However there is considerable resistance to govt edicts affecting lifestyle choices and the belief people should be allowed to eat what they want.
Smoking became socially unacceptable and was heavily taxed. Whether-like smoking- eating too much of the wrong food and not exercising will come to be seen as socially unacceptable we shall have to wait and see as this fat is beautiful mantra will need to wither away first
tonyb
Tony –
> I think there was sufficient time between being overweight, unfit and lacking Vitamin D as known health factors in April 2020, and the second wave in October of that year for many people to have done something about it.
I think you’re both vastly overestimating the reach of the “body positive” movement and underestimating complicated nature of addressing obesity. The lifestyle changes needed to address obesity, by altering both levels of physical activity as well as diet, are very complicated and significant and the impact of “advice” from public health officials is far from something that would see significant effect on a population-wide scale in a matter of one year.
As for Vitamin D, I would be less willing to weigh-in on the impact over one year of advice from public health officials, but I will say that from what little I’ve read, the jury is still out as to whether taking vitamin D in pill form, as opposed to increasing levels from exposure to sunlight or even from diet, is likely to have a large impact on reducing the prevalence of COVID morbidity and mortality. It’s an incredibly complex mechanistic process, and simple associations of Vitamin D deficiency with COVID morbidity does not easily lead to a conclusion of “Just take come Vitamin D and we’ll have a significant impact population-wide.”
>However there is considerable resistance to govt edicts affecting lifestyle choices and the belief people should be allowed to eat what they want.
No doubt. And that’s party of the reason why I think it’s extremely unrealistic to think that some simple advice from government officials would create a strong signal in obesity rates, let alone COVID morbidity and mortality as the result of a significant change in obesity rates.
> Smoking became socially unacceptable and was heavily taxed. Whether-like smoking- eating too much of the wrong food and not exercising will come to be seen as socially unacceptable we shall have to wait and see as this fat is beautiful mantra will need to wither away first
This notion that what is “socially acceptable” leads to a simple cause-and-effect with obesity rates is incredibly simplistic. First, the physiology of obesity is incredibly complex. A simple “energy balance” model is useful to a degree, but in reality the mechanisms of individual physiology, especially once you figure in factors like “food deserts” and other macro-scale societal changes that have correlated with the obesity crisis requires a much more sophisticated approach.
Obesity is still widely considered extremely undesirable in our culture. The idea that the obesity epidemic is somehow attributable to a relatively small movement of “body positive” just seems to me to be extremely implausible, and more some kind of politically-based argument than a scientific one.
As for Vitamin D, I would be less willing to weigh-in on the impact over one year of advice from public health officials, but I will say that from what little I’ve read, the jury is still out as to whether taking vitamin D in pill form, as opposed to increasing levels from exposure to sunlight or even from diet, is likely to have a large impact on reducing the prevalence of COVID morbidity and mortality. It’s an incredibly complex mechanistic process, and simple associations of Vitamin D deficiency with COVID morbidity does not easily lead to a conclusion of “Just take come Vitamin D and we’ll have a significant impact population-wide.”
>However there is considerable resistance to govt edicts affecting lifestyle choices and the belief people should be allowed to eat what they want.
No doubt. And that’s party of the reason why I think it’s extremely unrealistic to think that some simple advice from government officials would create a strong signal in obesity rates, let alone COVID morbidity and mortality as the result of a significant change in obesity rates.
I hate to interrupt this conversation but there is actually some real evidence on Vitamin D and flu transmission. In fact this paper hypothesizes a controling influence of winter Vitamin D deficiency on flu.
https://virologyj.biomedcentral.com/articles/10.1186/1743-422X-5-29
Tony –
As we can see from the article that David links, there is some interesting information out there about the influence of Vitamin D.
But to my knowledge, there isn’t much that’s conclusive and in particular very little that’s conclusive as to the impact of taking Vitamin D orally as a way to prevent or mitigate a COVID infection.
Given the information that is out there, I’m not sure why there hasn’t been more public health messaging advising people to take Vitamin D as a kind of precautionary measure.
But still, a strong conclusion that public health officials advising people to take Vitamin D orally would show a significant signal in the trajectory of the pandemic seems very premature to me.
Meanwhile, Tony – you might find this interesting:
https://twitter.com/EricTopol/status/1354290584454799362
There is a difference between October 2020 and October 2019. The pandemic did not reach the world or the US until January 2020 or thereabouts. Typically the US virus (flu) season starts in Mid-October. It can be seen that in Mid-October 2020 – the lowest numbers were achieved and at that same point started to ramp up to the largest, January 13th, 2021 at 3,355 deaths reported in the 7-day average. It should be noted that the new case infections have dropped each day from Jan. 11th until today (Jan.26 report – a slight 700 case bump the day before). This Mid-October ramp up to this day suggests a cause as the New Season hitting much like with the New Flu season begins. January the US “missed” the October strike. By April 2019 when the virus was in full swing and without the Mid-October New season strike, the US contracted a weaker virus through to April with more than 1,000 fewer deaths per 7-day occurring compared to today’s.
.
correction … By April 2020 not 2019…
Joshua
A seemingly extraordinary finding from new York University
https://www.dailymail.co.uk/sciencetech/article-9192869/Schizophrenics-THREE-TIMES-likely-die-Covid-19.html
tonyb
Fascinating findings on a couple of levels, including having mental disorders after leaving hospital.
Without being flippant, if you have this illness plus some of the other known conditions such as obesity that would be very bad news.
tonyb
Tony –
Having cared for a brother with schizophrenia for decades, I am somewhat skeptical of their findings – to the extent that they’re suggesting a physiological causal mechanism – but it’s certainly an interesting study.
Here’s a better overview.
https://nyulangone.org/news/schizophrenia-second-only-age-greatest-risk-factor-covid-19-death
And here’s a link to the study.
https://jamanetwork.com/journals/jamapsychiatry/fullarticle/2775179
BTW – that headline at the article you linked (and some aspects of the article itself) is disturbingly misleading. I think it merits the label of misinformation, perhaps disinformation or fake news.
Joshua
I don’t pretend to know anything about it but having read up a little about it the illness seems to be treatable but not curable and that those with it seem to be more susceptible to such things as heart disease and diabetes. Which of course are covid warning factors in themselves.
Any idea why they should be more susceptible?
Tonyb
Tony –
My brother was almost completely unaware of what was going on with his body. Or surpressed that awareness for a variety of reasons. I seem to recall reading something about that as a more general characteristic of schizophrenia.
In general, they are going to be remarkably unable or unlikely to adopt healthy behaviors, eat well, sleep well, practice good hygiene, go to the doctor if they’re ill, etc. Smoking in that population is incredibly high. So the list of confounds you’d have to control for, to isolate that particular mental disorder as compares to others would be enormous. They are also quite likely to rarely or never get medical care. If they do get medical care regularly, they’re quite likely to be taking a regimen of powerful drugs that have many side effects over very long time periods.
Consider the people sleeping under a bridge in your community. My guess is they’re relatively quite likely schizophrenic – even as compared to the likelihood of having other mental disorders like anxiety. Would you predict they’d have poor health outcomes?
I have no doubt my brother would have died decades prior to when he died if he didn’t have a have a family who took good care of him. Many, many schizophrenics aren’t that lucky.
They’re also prone to anxiety and depressions and many other predictors for poor health outcomes, along with the many SES predictors.
Now in the study they tried to control for known medical history… but still. I’m not saying they didn’t control for confounds well – just that it would be an incredibly daunting task.
So would the susceptibility to heart disease or diabetes or COVID be physiological as opposed to being, effectively, environmental? I’m certainly not in a position to give a scientific opinion. But I think that there’s a very high bar to be set before making conclusions. The study is interesting and I wouldn’t dismiss it – but I think in the least the one study rule should apply.
Joshua
I have got to know some of the ( thankfully very few) homeless round here who sell a magazine called Big issue. Reading through the magazine it is remarkable how many die in their forties and fifties.
They are often reluctant to get the free medical treatment available or to go into a hostel except in exceptionally cold conditions.
I had never thought before that they might have schizophrenia.
Tonyb
Yes Tonyb, In the US its scandalous how many of the mentally ill are homeless and not institutionalized. There are a significant proportion however who are simply drug addicts who prefer living under underpasses. It is also quite true that homelessness in the US strongly correlates with government tolerance and indeed subsidization of it. It also correlates with decriminalization of petty crime.
Like many other issues, its one surrounded with woke ideology that is divorced from reality.
More on Manaus – with troubling implication to the “let it rip” philosophy of rightwing armchair epidemiologists (and lions also) :🦁
https://www.washingtonpost.com/world/2021/01/27/coronavirus-brazil-variant-manaus/
We took a hard look at the winter surge in California and came to the conclusion that Winter connectivity levels exceeded those at the onset of the outbreak … our results are largely consistent with the high level conclusions here. Our model includes both connectivity and susceptibility probability distributions as well as dynamic models for connectivity over time by age group. https://covidplanningtools.com/understanding-the-california-winter-surge/
Thanks for the interesting report, and the use of an agent based model.
But… what measurements did you use, as opposed to model parameters?
One concern I have had with heterogeneity is the lack of actual data on both connectivity and individual susceptibility. And if you have enough free parameters, a model can fit pretty much anything.
Do you have access to some measurements not commonly available?
Tks..
This is indeed an interesting article. I would caution however that attributing changes in infections with connectivity alone is a simplification. Or maybe they used a more sophisticated model. Can you enlighten me on this?
There might be other effects like seasonal differences in underlying health or vitamin D levels. Particularly in Northern California there is a big seasonal difference in hours of sunlight.
But generally, I also tend to think that California and some other states will hit the HIT for “summer” conditions before vaccination will have much effect.
I also hypothesize that 2021 will see significantly lower mortality than 2020 and than 2019 due the effects of a large number of people who were seriously ill dying somewhat earlier due to covid19. Could be wrong of course, but I will be following up to see if its correct or not.
“Key Lesson: Ro at the Outbreak May NOT represent the Bounding Case
The key lesson is that human connectivity driving infectious spread varies seasonally as a function of weather and human behavior. We had assumed that the initial R0 conditions for a given geography / demographic were a static upper bound. However, our assessment of the winter outbreak required an increase of connectivity to a level ABOVE the levels used to match the initial R0 conditions. In simple terms, the level of potentially infectious human interaction during the cold winter months, especially during the holidays, exceeds the levels in early spring. Likewise, very hot weather also drives an increase in connectivity as people spend more time inside in air=conditioned homes, but likely not to the extreme of the colder winter months during the Thanksgiving, Christmas and New Year’s holiday season.”
I do agree with this bit. R0 is not a very meaningful number.
“I do agree with this bit. R0 is not a very meaningful number.”
Replace “meaningful” with precise and I’d agree. It is a very meaningful number. The issue is how you measure it. And, they measured it at a time when there was probably already some change in connectivity due to awareness of the virus. After all, early March was when sanitation supplies started disappearing, so people were altering behavior due to the anticipation of the disease. I was already trying to find sanitation supplies (and masks) wearing gloves when out, sanitizing more than I do normally during flu season, and substantially limiting store visits. The shortages suggest I was hardly alone.
Also, the SoCal surge has a new, more infectious variant, and that would, by definition, have a higher R0.
We have been over this many times meso. Please try to respond to what I cited from the literature. For covid19 according to Kwok 1-1/R0 varied between countries from 5% to 78% this spring. Likewise it varies by at least a factor of 2 because of seasonal factors. “Normal conditions” is not meaningful and you have made no attempt to define it. It’s like trying to define “normal weather” with a single number. It is just very very simplistic and not useful.
As a check on your thinking can you please define what “normal conditions” means?
I’ll bit one more time.
“We have been over this many times meso. Please try to respond to what I cited from the literature. For covid19 according to Kwok 1-1/R0 varied between countries from 5% to 78% this spring. Likewise it varies by at least a factor of 2 because of seasonal factors. “Normal conditions” is not meaningful and you have made no attempt to define it. It’s like trying to define “normal weather” with a single number. It is just very very simplistic and not useful.”
There is no way in hell that R0 varied that much. Asserting as such can mean either ignorance, or a use of the term R0 to mean Rt.
“As a check on your thinking can you please define what “normal conditions” means?”
Normal conditions are when nobody is immune (immunologically naive) and no precautions are being taken. That defines R0.
One could quibble about summer or winter or whatever, and get somewhat different values. But the best evidence is that seasonality is more related to the changes in human behavior with the season than changes in factors which affect transmission, such as humidity or temperature.
The R0 is somewhat different for different populations, due primarily to levels of social interaction.
So R0 is not a universal constant, but it is a pretty good approximation. And, as different populations are measures, an average R0 converges.
And, R0 is useful, and your assertions otherwise are wrong. Generally, when one makes such a sweeping statement that goes against the vast majority of scientists and practitioners in a field where experiments can be done (nature does them for us) and measured (unlike some of the climate change world), it is appropriate to consider that person a crank.
I’ll try this one more time as I’ve cited it a couple of times before.
“Reaching herd immunity depends in part on what’s happening in the population. Calculations of the threshold are very sensitive to the values of R, Kwok says. In June, he and his colleagues published a letter to the editor in the Journal of Infection that demonstrates this4. Kwok and his team estimated the Rt in more than 30 countries, using data on the daily number of new COVID-19 cases from March. They then used these values to calculate a threshold for herd immunity in each country’s population. The numbers ranged from as high as 85% in Bahrain, with its then-Rt of 6.64, to as low as 5.66% in Kuwait, where the Rt was 1.06. Kuwait’s low numbers reflected the fact that it was putting in place lots of measures to control the virus, such as establishing local curfews and banning commercial flights from many countries. If the country stopped those measures, Kwok says, the herd-immunity threshold would go up.”
This was in a Nature and was one of the first to come up in a search on I think herd immunity threshold. They clearly calculated HIT from Rt, not from R0 because R0 does’t mean anything in terms of whether an epidemic is growing or fading away.
I don’t think you said much in your last response to dispute that R0 can be highly variable. For the flu, it varies strongly seasonally and that’s likely the case for covid10 too.
You have yet to offer a definition of R0 that is well defined. It obviously is highly variable and not some universal constant for any virus.
From your own citation: ““Key Lesson: Ro at the Outbreak May NOT represent the Bounding Case”
Certainly you would agree that Rt is a vastly more meaningful number for predicting things.
Re: “You have yet to offer a definition of R0 that is well defined”
It’s been defined for you multiple times now, for months.
Once again:
R0 is the basic reproduction number under baseline conditions, meaning no additional public health interventions nor additional behavior changes beyond what would typically be present. In laymen’s terms for the COVID-19 pandemic, that means what conditions would have been at the same time of year in 2019. So if you have a substantial level of additional behavior changes and/or public health interventions beyond what you had at the same time of year in 2019, then you aren’t at baseline conditions and R0 doesn’t apply. For a novel pathogen, baseline conditions are typically present right near the beginning of the outbreak when the pathogen is first introduced, with virtually no one infected and before there’s enough time for behavior + interventions to change much:
Ferguson et al., March 2020:
“In the (unlikely) absence of any control measures or spontaneous changes in individual behaviour, we would expect a peak in mortality (daily deaths) to occur after approximately 3 months (Figure 1A). In such scenarios, given an estimated R0 of 2.4, […]”
https://www.imperial.ac.uk/mrc-global-infectious-disease-analysis/covid-19/report-9-impact-of-npis-on-covid-19/
“This quasi-equilibrium is maintained not because of herd immunity but because of changes in behavior. […]
The peaks occur at levels of infection far from that associated with herd immunity. Post-peak, shoulders and plateaus emerge because of the balance between relaxation of awareness-based distancing (which leads to increases in cases and deaths) and an increase in awareness in response to increases in cases and deaths.”
https://www.pnas.org/content/117/51/32764
“This allows us to “bend the curve” and predict temporary equilibrium states, far away from the equilibrium state of herd immunity, but stable under current conditions […]. Yet, these states can quickly become unstable again once the current regulations change.”
https://link.springer.com/article/10.1007/s00466-020-01880-8
It’s telling that this needs to be explained for the umpteenth time, since many of ‘the herd immunity threshold (HIT) is really low!’ folks, including Nic Lewis, approvingly cited Tom Britton’s June – August 2020 work, when Britton made much the same points about R0, HIT, and baseline conditions. To his credit, Britton later admitted he was wrong on Sweden reaching a low HIT:
Britton, June – August 2020:
“Herd immunity is defined as a level of population immunity at which disease spreading will decline and stop even after all preventive measures have been relaxed. If all preventive measures are relaxed when the immunity level from infection is below the herd immunity level, then a second wave of infection may start once restrictions are lifted.”
https://science.sciencemag.org/content/369/6505/846.full
Britton, June 2020:
“In June the journal Science published a study that incorporated a modest degree of heterogeneity and estimated the herd immunity threshold for COVID-19 at 43% across broad populations. But one of the study’s co-authors, Tom Britton of Stockholm University, thinks there are additional sources of heterogeneity their model doesn’t account for.
“If anything, I’d think the difference is bigger, so that in fact the herd immunity level is probably a bit smaller than 43%,” Britton said.”
https://www.quantamagazine.org/the-tricky-math-of-covid-19-herd-immunity-20200630/
citing Britton, December 2020:
“Whether or not Sweden publicly admitted its strategy was to pursue herd immunity, other countries began to cite its approach as such. In July, according to a report in Politico, White House advisors promoting herd immunity referenced a June study by Sweden’s pandemic modeler, Tom Britton, which said that herd immunity could occur after just 43 percent of a population became infected—an estimate far lower than what most other epidemiologists have put forward. Britton told Foreign Policy that his calculations that Sweden would reach herd immunity turned out to be incorrect. Britton now says that U.S. government officials misinterpreted his study and that using his June research to promote herd immunity was wrong, adding that “too many people will die in order to reach herd immunity.””
https://archive.is/2NJIk#selection-1607.0-1623.302
‘Behavior changes and/or public health interventions’ include:
– cancelling large gatherings
– closing universities and high schools / secondary schools with respect to in-person schooling, when they typically would have been open
– people interacting much less in public, as reflected in mobility data, surveys of the public regarding their behavior, etc.
– restrictions on air travel
etc.
Stockholm, and Sweden in general, had all of that at levels beyond the same time of year in 2019, and thus they were not under the baseline R0 conditions of herd immunity. Nic Lewis was thus wrong when he said Stockholm achieved herd immunity just because their cases/day, COVID-19 deaths/day, and hospitalizations/day decreased sometime in the spring ( https://archive.is/h96zK#selection-23077.0-23105.283 ). It wasn’t herd immunity that caused that, but the additional behavior changes and/or interventions that occurred.
“Assuming no population immunity and that all individuals are equally susceptible and equally infectious, the herd immunity threshold for SARS-CoV-2 would be expected to range between 50% and 67% in the absence of any interventions.
[…]
Initially, some local authorities and journalists described this as the herd immunity strategy: Sweden would do its best to protect the most vulnerable, but otherwise aim to see sufficient numbers of citizens become infected with the goal of achieving true infection-based herd immunity. By late March 2020, Sweden abandoned this strategy in favor of active interventions; most universities and high schools were closed to students, travel restrictions were put in place, work from home was encouraged, and bans on groups of more than 50 individuals were enacted. Far from achieving herd immunity, the seroprevalence in Stockholm, Sweden, was reported to be less than 8% in April 2020,7 which is comparable to several other cities (ie, Geneva, Switzerland,8 and Barcelona, Spain9).”
https://jamanetwork.com/journals/jama/fullarticle/2772167
And when Sweden’s conditions got back too close to baseline later in the year, Sweden’s government responded with even stronger interventions. So it’s simply a myth to claim non-vaccine-mediated herd immunity saved Sweden in the spring and summer of 2020. Nor is it going to save them in the future, since if they ever reach HIT, it’s going to be via a vaccine.
Same point holds for Lewis’s claims on Stockholm, and Sweden overall, and Geneva, and New York City, and London, and… ( https://archive.is/1j3vS#selection-6021.0-6312.0 ). Geneva and Sweden are particularly telling examples since their subsequent waves were stronger than their 1st, which goes beyond even Lewis’ false prediction of them not having strong 2nd waves:
https://www.covid19.admin.ch/en/epidemiologic/death?detRel=abs&detTime=total&detGeo=GE
https://ourworldindata.org/coronavirus-data-explorer?zoomToSelection=true&time=earliest..latest&country=SWE~DNK~FIN~NOR®ion=World&deathsMetric=true&interval=smoothed&perCapita=true&smoothing=7&pickerMetric=location&pickerSort=asc
That’s what happens when your interventions and behavior changes stop the first wave, but with a high infection rates and many people still infectious. When those interventions and behavior changes lapse to close to baseline (ex: due to colder weather mitigating distancing as people head indoors together), the virus can spread rapidly again, but now with the head start of having more infectious people to start with than when it was initially introduced to the population at the start of the 1st wave. A stronger 2nd wave is not what happens with herd immunity, as previous pathogen outbreaks have shown:
“If significant herd immunity developed following initial major water supply contamination, a multipeaked and/or prolonged epidemic would not be expected to occur.”
https://www.sciencedirect.com/science/article/abs/pii/S0168827894802297
More on Sweden’s additional interventions above those that would have been present at baseline:
“Sweden on Monday announced a ban on public events of more than eight people at a press conference where ministers urged the population to “do the right thing”.
[…]
The new limit is part of the Public Order Act and therefore is a law, not a recommendation like many of Sweden’s coronavirus measures. People who violate the ban by organising larger events could face fines or even imprisonment of up to six months.
[…]
“There should not be social situations with more than eight people even if they are not formally affected by the law. This is the new norm for the whole society, for all of Sweden. Don’t go to the gym. Don’t go to the library. Don’t have dinners. Don’t have parties. Cancel,” he said.”
https://swedishchamber.nl/news/sweden-bans-public-events-of-more-than-eight-people/
“Sweden’s upper secondary schools are again closing their doors to students from Monday until after Christmas.
[…]
They said it was “necessary” in order to curb a resurgence of the coronavirus across Sweden.”
https://www.thelocal.se/20201203/swedens-schools-for-over-16s-shift-back-to-remote-learning
https://academic.oup.com/cid/article/71/12/3174/5866094
https://academic.oup.com/ectj/article/23/3/323/5899049
https://www.oecd-ilibrary.org/sites/85e4b6a1-en/index.html?itemId=%2Fcontent%2Fcomponent%2F85e4b6a1-en
https://ourworldindata.org/policy-responses-covid
https://www.folkhalsomyndigheten.se/the-public-health-agency-of-sweden/communicable-disease-control/covid-19/
https://twitter.com/DrKatrin_Rabiei/status/1339988276229206017
Your definition doesn’t mean much as I’ve pointed out at least twice.
The moment one gives a moment’s thought to the concept of R0 the less useful it appears.
In fact, the number of individuals who’re on average infected by someone with the disease is virtually impossible to determine.
1. In the early stages of this epidemic, testing is scarce and so case numbers were dramatic underestimates.
2. Different countries or even different counties within a State will have dramatically different transmission rates.
3. Most viruses are strongly seasonal. Thus do we want to say that R0 can vary by a factor of 4 depending on which month the epidemic started in? Of course, its impossible to determine exactly when an epidemic started because initial growth is very small.
So, you can average over all seasons, all countries, etc. and get a number that will be incredibly difficult or impossible to compute accurately. This number is useless for predicting anything at a local or even country level where it’s actually important.
To be fair most epidemiologists know this and don’t really care much about R0. It seems to be mostly used as a communication tool to convince people epidemiology is a mature science that can predict things. The papers Nic has discussed here show that much more sophisticated models offer better results.
Re: “1. In the early stages of this epidemic, testing is scarce and so case numbers were dramatic underestimates.
2. Different countries or even different counties within a State will have dramatically different transmission rates.
3. Most viruses are strongly seasonal. Thus do we want to say that R0 can vary by a factor of 4 depending on which month the epidemic started in? Of course, its impossible to determine exactly when an epidemic started because initial growth is very small.”
1) Irrelevant to the fact that R0 is about baseline non-mitigated conditions. You simply don’t need widespread testing of much of the population to determine R0 since, for instance, you can find it by tracking contacts of an initial identified case to see how many of those were infected overtime.
2) And so will have different R0 values. There will be similarities in transmission that will make R0 similar in many places. But none of this makes R0 useless, anymore than temperature being different in some places, and more similar in others, makes temperature useless. It gives one information about that site, and sites like it. It will also be useful in comparing to the R0 of other pathogens in that area, such as seasonal influenza viruses.
3) This was already explained to you, dpy. Once again, R0 is about baseline conditions for that time of year:
“R0 is the basic reproduction number under baseline conditions, meaning no additional public health interventions nor additional behavior changes beyond what would typically be present. In laymen’s terms for the COVID-19 pandemic, that means what conditions would have been at the same time of year in 2019.”
Re: “To be fair most epidemiologists know this and don’t really care much about R0.”
Unlike you, I actually read papers in epidemiology, so I know you just invented a false claim. One of the first things epidemiologists did was try to figure out R0 when discussing herd immunity. So your claim is baseless. For instance:
“R0: the average number of secondary infections caused by a single infectious individual introduced into a completely susceptible population”
https://www.sciencedirect.com/science/article/pii/S1074761320301709
“(R0; the average number of persons infected by an infected person in a fully susceptible population)”
https://jamanetwork.com/journals/jama/fullarticle/2772167
“The basic reproduction number R0 denotes the average number of infectious contacts that an infected individual has before recovering and becoming immune (or dying).”
https://science.sciencemag.org/content/369/6505/846.full
You’re literally commenting on a post, dpy, in which Nic Lewis uses R0. So you’re not paying attention again:
“The herd immunity threshold (HIT) depends positively on the basic reproduction number R0 and negatively on heterogeneity in susceptibility.”
Re: “It seems to be mostly used as a communication tool to convince people epidemiology is a mature science that can predict things. The papers Nic has discussed here show that much more sophisticated models offer better results.”
You mean the papers in which people are still using R0, and cited by a person who’s still using R0?
Anyway, Lewis hasn’t offered “better results.” For example, he projected there would only be ~1000 more COVID-19 deaths in Sweden, at a time when they already had ~5400 deaths. Instead, Sweden has had close to 8000 more deaths, and increasing. So Lewis was wrong by about a factor of 8, and it’s likely to still get worse. How’s that for “better results”?
That occurred for the reason I already explained to you: he incorrectly thought Sweden was at the conditions of herd immunity, when actually mitigations and behavior changes pushed them from the baseline R0 conditions of herd immunity. When those mitigations and behavior changes were not enough to keep R below 1, the large 2nd peak Lewis failed to predict began. He tries to say this was just a matter of seasonality. It wasn’t. It was him messing up on what baseline conditions of R0 are.
“In the absence of a change in trends, it seems likely that the epidemic will peter out after a thousand or so more deaths, implying an overall infection fatality rate of 0.06% of the population (0.04% excluding COVID-19 deaths of people in care homes).”
https://judithcurry.com/2020/06/28/the-progress-of-the-covid-19-epidemic-in-sweden-an-analysis/
“I also projected, based on their declining trend, that total COVID-19 deaths would likely only be about 6,400. Subsequent developments support those conclusions.”
https://judithcurry.com/2020/07/27/why-herd-immunity-to-covid-19-is-reached-much-earlier-than-thought-update/
Sweden’s COVID-19 deaths:
https://ourworldindata.org/coronavirus-data-explorer?zoomToSelection=true&time=earliest..latest&country=~SWE®ion=World&deathsMetric=true&interval=total&smoothing=0&pickerMetric=location&pickerSort=asc
https://coronavirus.jhu.edu/data/mortality
https://covid19.who.int/region/euro/country/se
I was perusing my open windows in Chrome and found this really excellent article about how epidemic modeling has failed. It’s a supremely important point.
https://forecasters.org/blog/2020/06/14/forecasting-for-covid-19-has-failed/
I found the article to be tendentious and somewhat incorrect.
Examples: “For models that use exponentiated variables, small errors may result in major deviations from reality.
Epidemic growth is inherently exponential – it is akin to a chain reaction. The only way for it to not be exponential is for mitigations to balance it at some even level or linear rate – impossible to do in practice. Epidemic growth will tend to be essentially exponential in time (except over longer time scales when herd immunity kicks in, or mitigations), and declining exponential when shrinking.
“The core evidence to support “flatten-the-curve” efforts was based on observational data from the 1918 Spanish flu pandemic on 43 US cites. These data are >100-years old, of questionable quality, unadjusted for confounders, based on ecological reasoning, and pertaining to an entirely different (influenza) pathogen that had ~100-fold higher infection fatality rate than SARS-CoV-2.”
Nobody knows the infection fatality rate of the Spanish flu. But the case fatality rate was only a bit more than 2.5X that of COVID19 in the US. From a mitigation and impact point of view, this disease is similar to the 1918 influenza, with the exception of age-dependent severity.
Regarding flatten the curve – We already had witnessed the Wuhan and Italy disasters, where hospitals were overrun.
“Many models used by policy makers were not disclosed as to their methods; most models were never formally peer-reviewed and the vast majority have not appeared in the peer-reviewed literature even many months after they shaped major policy actions”
Not disclosed is true, but doesn’t make the models incorrect. Not peer reviewed was because there wasn’t time.
“Complex code can be error-prone and errors can happen even by experienced modelers; using old-fashioned software or languages (e.g. Fortran) can make things worse;”
Actually, Fortran is a decent language for modeling – it was written as a scientific applications language, after all. Python is often preferred today, because it is easier to code. I have done perfectly find modeling in Fortran, long ago.
Overall, that article (not a scientific paper, BTW) is a laundry list of things that might go wrong, thrown up against all models of the epidemic. It is essentially making the argument that modeling early in an epidemic is useless, without offering any alternative. It appears to be pretty much an argument from the “let ‘er rip’ world, with a strong bias towards the slow moving traditional academic research world. That bias is interesting, since some of the models came originally from that world (e.g. Ferguson).
There was only one reason that the models “failed” in terms of forecasts of deaths: mitigations, including importantly, voluntary mitigations taken independent of government mandate.
Otherwise, the curve would not have been flattened, and the medical system would have been overwhelmed, as it subsequently was in Tijuana and other Mexican Cities, Manaus (twice, the second after it was thought to have reached herd immunity) and Ecuador.
Meso, The article I cited is obviously correct about exponential variables. Any first year calculus student knows this. As an exercise for you calculate on your calculator 1.05^10 and 0.95^10. You will find that the final results differ by over 100%. The epidemeological models are ill posed so that small changes in R cause large changes in the result.
The rest of your comment seems to not offer anything substantial or even scientific.
“Meso, The article I cited is obviously correct about exponential variables. Any first year calculus student knows this. As an exercise for you calculate on your calculator 1.05^10 and 0.95^10. You will find that the final results differ by over 100%. The epidemeological models are ill posed so that small changes in R cause large changes in the result.
The rest of your comment seems to not offer anything substantial or even scientific.”
I’m not sure why I bother to react to your insults. I don’t need a calculator to understand exponentials, I I did first year calculus while I was 15, and plenty more after that.
Pointing out that exponentials are sensitive is tendentious. First of all, it is obvious. Second, it doesn’t offer an alternative. And third, that exponential nature does not effect the herd immunity threshold at all, only the HIT overshoot in a raging epidemic. The main effect of the exponential is that the *rate of growth* of the epidemic is sensitive to the base of the exponential, and the serial interval (which is in the exponent).
You failed to respond to my other points, claiming they were not substantial.
Meanwhile, you continue to assert that R0 is meaningless. I suggest that you do not know the meaning of the term “meaningless.”
I accept you’re surrender.
Well my point is that epidemiological models are ill-posed and so forecasts are usually totally wrong. In that the Ioannidis article is correct. And the track record for covid is extremely bad. I won’t do the search but you can find very scathing criticisms all over of Ferguson’s model. It was really disgraceful.
Their other points are also widely accepted for example about model documentation and terrible code maintenance. That was a major criticism of the Imperial College model.
There are plenty of alternatives to modeling. Gathering better data is vastly superior. Doing a better job of testing is also superior.
I see the same thing in my field. There is dramatic confidence in models that are all over the place. Laymen tend to fall for the colorful and deception of many marketers of models.
“Well my point is that epidemiological models are ill-posed and so forecasts are usually totally wrong. In that the Ioannidis article is correct. And the track record for covid is extremely bad. I won’t do the search but you can find very scathing criticisms all over of Ferguson’s model. It was really disgraceful.”
I disagree. Yes, Ferguson’s model had a lot of crappy code in it – it was a dog’s breakfast – a “kluge” in the parlance. But, had there been no mitigations, its predicted death rate would have been fairly accurate. You don’t need to be precise in this area, especially early on, but it’s important to get a ballpark idea.
For a quick look, you only need well known differential equations. I wrote a python model that used that approach, which used 54 lines of python, and again got results similar to Ferguson’s overly complex, buggy model.
But, I wanted one where I could add in mitigations or other things hard to do with diff-E, so I wrote a more complex SEIR model. In order to make it flexible beyond the basic diff-E approach, I used an event step approach which I have used in my professional field. It was so simple that it was easy to verify the code – 750 lines of python, about half of that being just graphical display. I got very similar results to Ferguson. And that’s not surprising, because in the absence of mitigation, you really only need a few differential equations to generate the results. And, naturally, it got pretty much the same results. Not surprisingly, the HIT was that predicted by the basic (non-heterogeneity) formula, but it got there my simulation.
Now, these models didn’t have a value for heterogeneity, so they came in somewhat high in total cases. But using a reasonable infection hospitalization rate, both predicted a disastrous, early overload of the medical systems.
That overload was demonstrated in reality in Wuhan, and later in Italy, in Tijuana, in Manaus, and many other places.
I would love to have had real data on heterogeneity in both contact rate and susceptibility. If I had that, I could have treated an agent based model and had lots of fun. But then it takes a huge amount of computing power, especially because those models may require stochastic techniques, requiring a potentially huge number of runs to achieve statistical significance. I have also done such models professionally. But in this epidemic, I don’t want to pay for the computing power to engage my intellectual interest, especially since I don’t have that data.
You must be in a field where models have to be exact to be useful at all. For example, engineers have requirement like that. But for designing policy early on, it’s better to have approximate models than nothing. And, for those purposes, yes, R0 is meaningful and important. The R0 could have been well below the best estimates, and with no mitigations, the hospital systems would have been crushed.
“There are plenty of alternatives to modeling. Gathering better data is vastly superior. Doing a better job of testing is also superior.”
Those are not alternatives. There are additional, important measures that we have failed badly at in the US – especially testing. But you need models for planning. You need a way test estimate required resources – health care, morgues space, vaccines, etc – and a way to estimate mitigation needs and effects.
Well in CFD requirements differ. But like epidemic modeling and weather forcasting the problem is ill-posed in very real senses. What that means is that naive simple models are badly wrong as time goes on. This is a truism that even laymen realize. I’m a little surprised that you seem to dispute it even though I have noticed a nonspecificity to your verbal rhetoric.
My main point is that simple epidemeological models are bound to be badly wrong as time goes on. In weather modeling the limitations are understood by everyone and the models are based on very sophisticated and detailed data sets for initial data. In an epidemic, the quality of the data is very poor by comparison.
Weather modeling is very sophisticated and the investment massive. Despite this, we all know that weather forcasts are not reliable after about 7 days at best. Epidemic modeling is primitive by comparison and relies on simplistic models, the data quality is very poor, and the results are of limited value.
Ioannidis’ paper documents these stark failures. But its really obvious based on mathematical first principles. I’m a little surprised you seem to disagree. Can you point to any model that has been skillful in predicting the course of this epidemic?
Annan’s model for example can’t predict much because it uses fits of past data to extrapolate into the future. No model that I know of predicted the winter second wave. Annan’s didn’t even show model results for more than a few weeks into the future. That’s because he knew they were worthless.
And no model can successfully predict the severity of a given flu season. No model I know of even predicts accurately the seasonality of viral epidemics. There are vague and crude mechanistic narratives such as “in the winter people stay indoors and have closer contact with other people.”
In reality epidemeology is a crude field with most predictions being little better than educated guesses. Did Fauci or anyone else say there would or would not be a winter second wave or how severe it would be? I don’t think they did because even experts know that there was no scientific basis for such quantifiication.
Re: “Meso, The article I cited is obviously correct about exponential variables.”
Outbreaks are often sigmoidal, not exponential. There’s an accelerating phase with R > 1, then R drops to 1, and then there’s a decelerating phase with R < 1. That's the case regardless of whether R drops below 1 due to non-vaccine-mediated herd immunity, lockdowns, increased mask-wearing, vaccination, voluntary social distancing, etc. Other shapes can occur, such as when behavior changes and/or public health interventions wax and wane, or as the pathogen mutates to forms with a higher R0 or which can re-infect people (leading to multiple accelerations and deaccelerations).
But it's ridiculous to criticize the use of exponential variables in pandemic modelling as if that's some deep insight. It only works on those who engage in motivated reasoning while not knowing what a "sigmoid function" is. But I guess I shouldn't expect much better when the lead author is John Ioannidis, a.k.a. Dr. 40K, who falsely forecasted less than 40,000 USA COVID-19 deaths and then started blaming others when his forecast failed:
“Data from Iceland and Denmark, which have done the best random sampling, also point in the same direction, Ioannidis said. “If I were to make an informed estimate based on the limited testing data we have, I would say that covid-19 will result in fewer than 40,000 deaths this season in the USA,” he told me.”
https://www.washingtonpost.com/opinions/without-mass-testing-were-flying-blind-through-this-crisis/2020/04/09/bf61e178-7a9b-11ea-a130-df573469f094_story.html
[ https://archive.is/dT97F#selection-2211.202-2219.279 ]
Another baseless comment in a long record of irrelevancies.
I did not say that exponential variables were not the best one could do in epidemic modeling. What I said was that the model is ill-posed in that small changes to R will result in dramatically different outcomes. Thus modeling epidemics is not likely to be very skillful.
You need to stop using straw man arguments. They are a sophist’s tool.
The sensitivity of models to exponentials, chaos and so on is hardly new. Lecturing about it is a waste off time.
Some of understand how to deal with the range of uncertainty.
You argue that R0 is meaningless or useless. That in itself utterly disqualifies you from this discussion.
“Outbreaks are often sigmoidal, not exponential. There’s an accelerating phase with R > 1, then R drops to 1, and then there’s a decelerating phase with R < 1."
Yes.
The ideal shape depends on whether one is talking about the total number of infections, or the number for a given time interval (epi curve). I've seen that confusion around.
If we assume no transmission changes, then:
–The total cases vs. time curve is sigmoidal.
–The per-time-interval curve will look exponential over short time periods, with either a positive or negative base (R). But for constant R (i.e.no change in the susceptible/population ratio, no mitigations), it is a pure unbounded exponential.
The latter ends up looking like Farr's Law, although that law was developed before infection spread was understood at all – germ theory wasn't around – so it was a descriptive law rather than one based on the "physics" of epidemics.
The impact of either mitigations or increasing immunity have the same effect – reducing R – which changes the shape of the curve, of course. The exception would be a change in heterogeneity.
And, of course, this a still a trivially simplistic model, but useful.
The real world is hardly ideal, hence neither is a curve.
This is pretty incredible. Jay Bhattacharya (and other GBD lions 🦁) is on the advisory board of a group that published the following:
-snip-
“Currently, there is no one for whom the befit would outweigh the risk of these vaccines – even the most vulnerable elderly, nursing home patients.”
-snip-
er…that is before they scrubbed it.
@Joshua
Re: “Jay Bhattacharya (and other GBD lions 🦁)”
In addition to being a co-author on the garbage Santa Clara seroprevalence study and a host of other distortions, Bhattacharya + one of his Santa Clara co-authors (Bendavid) defended a ~0.01% IFR for SARS-CoV-2 in the USA. Since the USA suffered >440,000 COVID-19 deaths, that means Bhattacharya + Bendavid need at least 4 billion people to have been infected in the USA. That’s more than 10 times larger than the USA’s population. Unsurprisingly, they said this nonsense in the Wall Street Journal, a haven for right-wing disinformation on COVID-19, AGW, second-hand smoking, etc.
These people are clowns just looking to downplay COVID-19 to avoid policies they dislike, like lockdowns. They’re on par with people who pretended smoking didn’t cause cancer or that second-hand smoking was not a serious health risk, so they could avoid cigarette taxes, bans on smoking in public places, etc.
“As of March 23, according to the Centers for Disease Control and Prevention, there were 499 Covid-19 deaths in the U.S. If our surmise of six million cases is accurate, that’s a mortality rate of 0.01%, assuming a two week lag between infection and death. This is one-tenth of the flu mortality rate of 0.1%. Such a low death rate would be cause for optimism.”
https://www.wsj.com/articles/is-the-coronavirus-as-deadly-as-they-say-11585088464
https://archive.is/QLmJt#selection-2607.523-2607.882
USA reported COVID-19 deaths (under-estimates actual COVID-19 deaths, as per excess deaths):
https://covid19.who.int/region/amro/country/us
https://coronavirus.jhu.edu/us-map
You should avoid terms like “garbage” because its applicable to your own comments. Many people were really wrong in March of last year. Citing guesses people gave in the press is cherry picking. No one had any idea at that time what the IFR was except based on the Diamond Princess ship which Ioannidis did a good job of analyzing. I think the estimate was somewhere near 0.2-0.3% but there were corrections needed to account for how many might ultimately die.
The Wall Street Journal is center right publication and its information is relied on by millions around the world including many on Wall Street. It does not contain “right wing disinformation.” This mud slinging shows you are a supplier of left wing disinformation.
I’m going to ignore many the content-free portions of your reply, dpy, since I’m not responding to you for your benefit.
Re: “Many people were really wrong in March of last year. Citing guesses people gave in the press is cherry picking.”
I’m well within my rights to point out false claims people make, especially when they continue to make false claims to suit their agenda. And the WHO was doing pretty well on IFR back in mid-February 2020, as I showed you elsewhere ( https://archive.is/WJmbY#selection-40753.0-40817.100 ). Which makes it even worse that Bhattacharya + Bendavid were still screwing up a month later on March 24:
WHO, in February 19 (Situation Report # 30):
“[…] current IFR estimates […] range from 0.3% to 1%.”
https://reliefweb.int/report/china/coronavirus-disease-2019-covid-19-situation-report-30-19-february-2020
[ https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports ]
Re: “No one had any idea at that time what the IFR was except based on the Diamond Princess ship which Ioannidis did a good job of analyzing. I think the estimate was somewhere near 0.2-0.3% but there were corrections needed to account for how many might ultimately die.”
No, the IFR from the Diamond Princess was ~2.0%, since 14 people died, contrary to Nic Lewis’ prediction of 8 – 9 deaths and about on par with Verity et al.’s prediction of 12 – 13 deaths ( https://archive.is/xKEBL#selection-299.91-307.242 ). Age-stratifying that to a general population in places like France, China, or Great Britain, gives a median IFR in the range of ~0.5% – ~0.9%. Getting 0.2% – 0.3% from that for the USA was ridiculous. That was shown in papers such as:
https://www.eurosurveillance.org/content/10.2807/1560-7917.ES.2020.25.12.2000256
https://www.thelancet.com/journals/laninf/article/PIIS1473-3099(20)30243-7/fulltext
https://science.sciencemag.org/content/369/6500/208
“The IFR estimates from Verity et al.12 have been adjusted to account for a non-uniform attack rate giving an overall IFR of 0.9% (95% credible interval 0.4%-1.4%).”
https://www.imperial.ac.uk/mrc-global-infectious-disease-analysis/covid-19/report-9-impact-of-npis-on-covid-19/
There were also other papers showing similar IFR estimates using modelling + other data sources, before Bhattacharya + Bendavid’s March article. For instance:
“[…] the infection fatality risk (IFR), i.e., the risk of death among all infected individuals, would be on the order of 0.5% to 0.8%.”
https://www.mdpi.com/2077-0383/9/2/523/htm
“The infection fatality risk (IFR) […] is therefore 0.3% to 0.6% […].”
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7074297/
Note how that’s all consilient with the WHO’s IFR range of 0.3% – 1.0% from February 19, 2020.
So no, it was clear even in March that IFR was not as low as Bhattacharya + Bendavid’s absurd March 24th value of ~0.01% for the USA. IFR also wasn’t as low as Ioannidis was claiming, as I again showed elsewhere ( https://archive.is/WJmbY#selection-40511.0-41017.1 ). Not surprising, since Ioannidis likely then used his under-estimated IFR to under-estimate the number of COVID-19 deaths the USA would suffer:
“Data from Iceland and Denmark, which have done the best random sampling, also point in the same direction, Ioannidis said. “If I were to make an informed estimate based on the limited testing data we have, I would say that covid-19 will result in fewer than 40,000 deaths this season in the USA,” he told me.”
https://www.washingtonpost.com/opinions/without-mass-testing-were-flying-blind-through-this-crisis/2020/04/09/bf61e178-7a9b-11ea-a130-df573469f094_story.html
[ https://archive.is/dT97F#selection-2211.202-2219.279 ]
You don’t get to re-write history to make Ioannidis, Bhattacharya, and Bendavid look less wrong than they actually were, dpy.
Atomsk, Just to take the most obvious of your misrepresentations from this latest long winded comment concerning the Diamond Princess.
1. Ioannidis calculated what the DP statistics implied about the US population by doing an age adjustment. That’s the number I was referirng to. When you logic chop you miss little subtleties require careful thought.
2. Ioannidis’ assumed there might be additional deaths and multiplied his IFR by a factor of 2 to account for a number of effects. It turns out his adjustment was almost exactly correct.
3. The DP passengers were mostly from older age cohorts and so citing 2% as an IFR is very misleading. Yet you did it.
The problem with such long winded comments is twofold. First, no one is going to read it. Second you don’t have time to actually think about the random nonsense you are writing. Just some friendly advice.
The other very obvious fallacy is that IFR will be highly variable depending on the population. It will vary from country to country.
> 1. Ioannidis calculated what the DP statistics implied about the US population by doing an age adjustment. That’s the number I was referirng to. When you logic chop you miss little subtleties require careful thought.
Extrapolating from the DP was very amateurish – which is why I might expect it from Nic, but coming from Ioannidis it was just weird.
The DP would obviously not be a representative sampling for a variety of reasons. Age, which was adjusted for is only one factor. But even more than that, (1) it COVID on board could have been entirely comprised if an unrepresentative COVID variant and, (2) the conditions (pre- and during-infectiousness) were about as unrepresentative as you could get. Extrapolating from live in a cruise ship to people living in high rise buildings in NYC or farms in Iowa???
Weird. Particularly from someone criticizing other scientists who were drawing conclusions from poor quality data early on in the pandemic.
It was not weird at all because it was perhaps the first large dataset that could inform about covid19. It would have been malpractice to not try to use it to estimate IFR.
But I know confounders confounders and uncertainties. Josh, grow up, early science can be more uncertain but very valuabel.
Re: “Ioannidis calculated what the DP statistics implied about the US population by doing an age adjustment. That’s the number I was referirng to.”
I already explained to you that age-adjustment would give an IFR in the range of ~0.5% – 0.9%, not the 0.2% – 0.3% you cited for Ioannidis. That includes for the USA, as shown when Ferguson et al. extended Verity et al.’s Diamond-Princess-based IFR to the USA, with age-stratification.
Re: “Ioannidis’ assumed there might be additional deaths and multiplied his IFR by a factor of 2 to account for a number of effects. It turns out his adjustment was almost exactly correct.”
His IFR was wrong for the USA, as noted above. You’ve already been shown that the USA’s IFR up to December 2020 was, at minimum, ~0.5%, even if one goes with the CDC’s over-estimated number of infections:
https://archive.is/TH5uq#selection-36441.0-36521.110
Earlier in the pandemic, the CDC’s IFR was higher, at 0.65%, and they cited Gideon Meyerowitz-Katz’s IFR research to support that:
http://archive.is/w2xC7#selection-2331.0-2331.6
Re: “The DP passengers were mostly from older age cohorts and so citing 2% as an IFR is very misleading. Yet you did it.”
I already showed that age-stratifying that ~2.0% IFR gives a median IFR in the range of ~0.5% – ~0.9%.
Re: “The problem with such long winded comments is twofold. First, no one is going to read it”
My comments are not made for your benefit, and others read them. The fact that you don’t bother to pay attention before responding is your issue; no need for you to extend that to those who aren’t you.
As mesocyclone said in reply to a comment of mine:
“You are correct – thank you very much for the links.”
https://judithcurry.com/2021/01/10/covid-19-why-did-a-second-wave-occur-even-in-regions-hit-hard-by-the-first-wave/#comment-942297
And the basic research isn’t the problem. The problem is trying to extrapolate from that population to make a global IFR projection.
It was extrapolated to the US by doing various adjustments, not the world. You really need Josh to try to avoid obvious errors and misstatements.
> It was extrapolated to the US by doing various adjustments, not the world. You really need Josh to try to avoid obvious errors and misstatements.
bu global I meant generalized.
Did they adjust for SES, David? Race/ethnicity? Medical history?
The population on board a cruise isn’t remotely generalizable.
Because Sanakan is doing his usual cherry picking it’s important to go back and see what was actually said by Ioannidis who has the most cited medical sciences paper of the last 20 years. Sanakan has zip, zero, nada.
‘Projecting the Diamond Princess mortality rate onto the age structure of the U.S. population, the death rate among people infected with Covid-19 would be 0.125%. But since this estimate is based on extremely thin data — there were just seven deaths among the 700 infected passengers and crew — the real death rate could stretch from five times lower (0.025%) to five times higher (0.625%).”
He then goes on to say that a mid range estimate is 0.3%.
David –
One of the fundament tenets of epidemiology is to not extrapolate from non-representative samples. You can’t get more non-generalizable than a bunch of cruise passengers and the treatment conditions of a cruise are very idiosyncratic.
Shows fundamental confusions about scientific method Josh. Science uses the best data available no matter how flawed and then tries to improve with time.
David –
> Science uses the best data available no matter how flawed and then tries to improve with time.
Science uses the best data available. So study that data. Don’t extrapolate from it if it isn’t representative. Basic science.
You of all people don’t get to tell scientists what to do and not do. You have no idea what any tenet of epidemiology is.
David –
> You of all people don’t get to tell scientists what to do and not do.
I’m telling you, because you don’t seem to understand.
Ensuring representativeness of sampling is a fundamental tenet of epidemiology. If representativeness isn’t an intrinsic quality of the sample, then you do post-stratification adjustments. They did that with age, but failed to do so on a large # of other variables that are predictive of health outcomes – such as SES or medical history.
So sure, do an analysis of the Diamond Princess data. It has some value. But extrapolating from that data is fundamentally bad epidemiological science. That’s why it was so surprising to see Ioannidis do that (just as it was with the Santa Clara data), if not someone who has no apparent background in epidemiology like Nic.
https://twitter.com/malar0ne/status/1353934042236784640
Josh is spreading dangerous misinformation again. Twitter is a garbage medium where most things are false or taken of context. I went to the page and what it actually says is this:
“PANDA supports the use of safe and efficacious vaccines as part of a focused protection strategy. When vaccines are in short supply, older high-risk people should be prioritized together with nursing home and hospital staff.”
The 🦁s scrubbed it.
Here’s how another 🦁 responded when called on it:
https://twitter.com/MLevitt_NP2013/status/1354106007849463809?s=20
It was archived before the edit, dpy.
“Once this is established, individuals, with the help of their physician, should do a cost-benefit analysis based on their individual profile, their age and their health status to determine whether taking the vaccine would be more beneficial than getting a natural infection.
Currently, there is no one for whom the benefit would outweigh the risk of these vaccines—even the most vulnerable, elderly nursing home patients.”
https://web.archive.org/web/20210122171642/https://www.pandata.org/you-asked-we-answered/
Learn to admit when you’re wrong, for once in your life.
Atomsk –
> Learn to admit when you’re wrong, for once in your life.
It’s quite fun to watch. I predict next he’ll to to the “What I probably meant to say” mode to explain it.
As he did here:
https://judithcurry.com/2021/01/10/covid-19-why-did-a-second-wave-occur-even-in-regions-hit-hard-by-the-first-wave/#comment-941025
Joshua, I expect the same. It’s telling that all three ‘Great Barrington Declaration’ contrarians (Gupta, Bhattacharya, + Kulldorff) served as scientific advisors to PANDA while this was going on, along with other known crackpots on COVID-19 like Altas, Levitt, Sikora, and Yeadon:
https://web.archive.org/web/20210126054236/https://www.pandata.org/scientific-advisory-board/
I’ve dealt with anti-vaxxers for over a decade, so I know their tropes. PANDA is clearly an anti-vaxxer organization, right down to their JAQing off:
“PANDA – Pandemics Data and Analytics
[…]·
The Covid vaccine comes at a cost to our immune systems. Is it worth it?”
https://www.facebook.com/112123760487193/posts/the-covid-vaccine-comes-at-a-cost-to-our-immune-systems-is-it-worth-itwhen-we-ar/215760306790204/
You could literally put that in a textbook or peer-reviewed paper on JAQing off. That’s how blatant it is. Robert F. Kennedy, Jr would be proud.
“Not just asking questions: Effects of implicit and explicit conspiracy information about vaccines and genetic modification”
https://www.tandfonline.com/doi/full/10.1080/10410236.2018.1530526?af=R&
“Just asking questions (also known as JAQing off) is a way of attempting to make wild accusations acceptable (and hopefully not legally actionable) by framing them as questions rather than statements.”
https://rationalwiki.org/wiki/Just_asking_questions
Further context on JAQing off:
“Beyond belief: The social psychology of conspiracy theories and the study of ideology”
“Conspiracy and bias: argumentative features and persuasiveness of conspiracy theories”
“Propagating and debunking conspiracy theories on Twitter during the 2015–2016 Zika virus outbreak”
Dr. Gold on experimental covid vaccines being given now …
Basically, the vaccine isn’t of any use to anyone under 60. Over 70 or so, it would make sense. There are therapeutics that work if taken early.
> Basically, the vaccine isn’t of any use to anyone under 60
Do you still not understand the idea of herd immunity?
Why do you not want our economy to normalize?
Lol.
She won’t call them vaccines – calls them “experimental biological agents.”
Amazing what COVID turned up among “skeptics.” Almost as revealing as the election.
So – she was arrested for participating in the Capitol Hill Mob riot. She says that the vaccines will result in millions of people testing positive.
She’s a real winner you’ve got there, Jim.
She did enter the Capitol and gave an impromptu speech. You can see Capitol police standing by the door and don’t seem concerned. At one point they walk over to her, but then move off. It’s a peaceful scene and I think she may have thought it was legal to enter through the door with Capitol police standing by the door unconcerned.
https://twitter.com/Autre_Vierge/status/1348152363765305344
As usual, you slime the person, not what she says. What she says sounds like the truth. The vaccine wasn’t tested long enough to determine if it would have adverse effect like the mRNA SARS vaccine did. She’s not saying it will happen, she’s saying no one knows – a perfectly true statement.
She has first hand experience with hydroxychloroquine and zinc to treat early COVID. I’m guessing you have zero experience with that.
So, all in all, she’s got chops.
Jim –
Why is it “smearing” her to point out that she was part of the Capitol Hill Mob riot?
-snip-
Medical News and Free CME Online
Home
Specialties
Back
Allergy & Immunology
Anesthesiology
Cardiology
Critical Care
Dermatology
Emergency Medicine
Endocrinology
Gastroenterology
Genetics
Geriatrics
HIV / AIDS
Hospital-Based Medicine
Infectious Disease
Nephrology
Neurology
Nursing
OB / GYN
Oncology / Hematology
Ophthalmology
Orthopedics
Pain Management
Pathology
Pediatrics
Primary Care
Psychiatry
Pulmonology
Radiology
Rheumatology
Surgery
Transplantation
Urology
CME/CE
Health Policy
Meetings
COVID-19
Videos
Opinion
Condition Centers
Careers
Society Partners
Back
AAD Reading Room
ACR Reading Room
AGA Reading Room
ASCO Reading Room
Endocrine Society Reading Room
IDSA Reading Room
enter search terms
Enter Search Terms
Washington-Watch > Washington Watch
Simone Gold Arrested for Role in Capitol Insurrection
— Physician faces charges for entering restricted grounds, disorderly conduct
by Amanda D’Ambrosio, Staff Writer, MedPage Today
January 20, 2021
share to facebook
share to twitter
share to linkedin
email article
A photo of Simone Gold, MD, JD, looking through handcuffs
Simone Gold, MD, JD, founder of the notorious pro-hydroxychloroquine, anti-vaccine group America’s Frontline Doctors, was arrested Sunday for participating in storming the U.S. Capitol earlier this month, according to the Department of Justice.
Following Gold’s confirmation that she entered the Capitol building during the riot on Jan. 6, DOJ officials arrested her in California on Sunday. Gold faces charges of entering a restricted building and for violent entry and disorderly conduct.
Department representatives did not return a query as of press time as to whether Gold has been released on bail.
Along with Gold, federal officials also arrested John Strand, communications director for America’s Frontline Doctors, who was pictured with the physician at the event.
Gold joined the mob that stormed Capitol Hill in an attempt to disrupt the 2020 election certification process. She spoke at protests in Washington, D.C., leading up to the riot, casting doubt on the vaccines and claiming that COVID-19 is non-fatal. Gold stated that citizens must not comply with taking “an experimental, biological agent deceptively named a vaccine.”
The California-based physician told the Washington Post that she was indeed inside the Capitol, as she followed a crowd and assumed it was legal to do so. Several photos of Gold at the insurrection have circulated online, as well as a video of her making a speech to rioters inside the federal building.
Starting on the day after the riot, federal investigators received photographs of Gold and Strand during the riot. The pair was also captured on video at the doors of the federal building, in the middle of a crowd attempting to push past law enforcement officials to get inside. In this footage, one law enforcement official appeared to be pulled down by someone in the crowd and landed right where Gold and Strand were standing, according to the FBI’s affidavit supporting the arrest warrant.
https://www.medpagetoday.com/washington-watch/washington-watch/90778
From the arrest warrant:
-snip-
21. Video posted to the Getty Images database online (https://www.gettyimages.com/detail/video/mob-of-pro-trump-rioters-and-protesters-break-intothe-u-news-footage/1294938866) shows what appears to be STRAND and GOLD in a large crowd attempting to push past multiple officers blocking the entrance to the Capitol, which had visibly broken windows at the time. One of the officers, who had been pinned near the doors to the Capitol, appears to be pulled down by someone in the crowd and lands near where STRAND and GOLD were standing.
Nic, I may be a simpleton, but what about consideration of Vitamin D deficiency, which is more pronounced in the Winter?
I linked earlier several times to a paper on influenza showing some rather strong evidence on the Vitamin D deficiency flu link.
https://virologyj.biomedcentral.com/articles/10.1186/1743-422X-5-29
dpy,
The three most notable things about Ioannidis’ Covid work are that
(1) it has been rapidly disseminated and publicised in the right wing media, pushed by people including yourself.
(2) It has consistently been an outlier in the rest of the science
(3) It has attracted very strong criticism from the rest of the field
“The sampling method, the statistical tests, the sensitivity/specificity of the tests, are all reported to be seriously flawed”
https://arxiv.org/pdf/2012.12400.pdf
You should familiarise yourself with the whole of the field rather than fixate on an outlier. I have suggested review articles for you previously.
Your continued focus on insulting people who disagree with you is again noted.
Here’s an excellent video where experts at the UCSF go into quite a bit of detail to explain the flawed convenience sampling methodology used in the Santa Clara study.
https://www.youtube.com/watch?v=NTXgbN6uB1I
It’s almost like that study was conducted by rank amateurs. But of course, they were amateur lions.
dpy,
You really do need to
(1) stop insulting people
(2) take your own advice on content of interactions
If you can manage a civil post, I’ll consider responding.
“dpy,
You really do need to
(1) stop insulting people
(2) take your own advice on content of interactions
If you can manage a civil post, I’ll consider responding.”
I second the motion.
David –
> suggesting that a true leader in the medical sciences is a “rank amateur.
It seems you misunderstood my point.
Our 🦁 is most certainly *not* a rank amateur. But he signed on to a study where the sampling methodology was at the level of a rank amateur. They explain why in the video if you care.
Of course, he was only one of quite a few authors and I’d guess he was only minimally involved in the sampling – but it’s just sad that such a 🦁 would sign on to a study with such poor methodology.
The fact that he’s a 🦁 just makes it that much more incriminating.
It’s important to respond to the “paper” VTG posted above because its very weak and offers nothing positive in terms of better estimates of IFR. They are mostly critiquing a meta-analysis by Meyerowitz-Katz and Merone. There is a short section on the Santa Clara study which is a rehash of old criticisms and the authors are apparently unaware of the improvements to their methods that Ioannidis et al have published. Finally I will note that the authors don’t appear to have any medical qualifications and so are reduced to irrelevancies such as “the serologic test was not approved by the FDA.”
In short, this paper is a purely negative critique of one meta-analysis. It offers nothing to actually improve IFR estimates. It is obvious that the authors disagree with the policy implications of a lower IFR estimate. It’s very similar to Gellman’s critique in which he admits that he is not an expert but that the author’s are the experts on medical issues.
I did prepare some time back a comment on IFR. For the scientific illiterates here, such as Josh and VTG, I will repost it in sections because its long. Basically IFR estimates will be all over the map because of the strong age dependence of susceptibility to Covid19.
David –
It’s good to see that you’ve written a couple of comments that weren’t filled with insults. Congratulations.
While you’re considering what’s “important.”
We have 25 million identified cases in the US. Estimates of unidentified cases vary, but lets say its 4 to 1 as that’s within the range of what the CDC says is plausible. So 100 million cases. We have about 440k dead from COVID already, and we could easily say another 30k will die among those already infected. So let’s put it at 470k dead out of 100 million infected.
So conservatively speaking the IFR would be 0.47%. And that’s now, not when Ioannidis was making his estimate. Obviously, because of improved treatment, the fatality rate has been reduced at least somewhat – meaning that when he made his estimate he was even further off.
Now it’s hard to say whether the IFR for the US would be generalizable. We have in some way a fairly unhealthy population. We are older than the populations of some other countries. We don’t have great life expectancy compared to some other countries. On the other hand, we have relatively good healthcare facilities. And, more importantly, Ioannidis himself at times, extrapolated from sampling (non-representative, at that) from the US to generalize an IFR. His meta-survey did use worldwide seroprevalence analyses to base is IFR, but there are legitimate criticisms of how he did that. Of course, you won’t look at those criticisms – but it doesn’t really matter – because as Nic made projections that showed his reliance on erroneous analysis, so did Ioannidis.
He made statements indicating that there might be as few as 40k deaths in the US.
He likened COVID to the seasonal flu in terms of virulence.
He might be a 🦁. But he is a 🦁 who made errors. It’s easy to see that he did so. Lot’s of people make errors. It’s not the end of the world. The more interesting question is why you’re so invested in acknowledging the obvious errors that he made.
And btw –
For Ioannidis’ IFR to be correct, given the number of COVID deaths we’ve seen (and will see from those already infected) the population infection rate would have to be over 50% currently. And obviously that number is going to go up.
What’s funny is that you think that Nic was right about the low “herd immunity threshold” AND that Ioannidis was right about the low IFR. But they aren’t compatible.
Something’s gotta give. At least one of your 🦁’s has to be wrong.
Joshua, Ioannidis under-estimates IFR and many of the informed experts paying attention know this. His estimate is an outlier that mostly uses studies on non-representative samples such as blood donors, non-targeted volunteers, dialysis patients + other people in hospitals, etc. Some other analyses suffer from this problem to a lesser degree, but Ioannidis takes it to an extreme: ~70% of the material he uses shouldn’t be used to calculate IFR.
Or at the very least, Ioannidis should have graded the studies for risk of bias, as per the PRISMA guidelines he himself co-authored. Then he should have done a sensitivity analysis / outcome-level assessment in which the most non-representative, lowest-graded samples were removed, and IFR was calculated with the more representative, higher-graded samples:
PRISMA guidelines, co-authored by Ioannidis:
“Item 19: Risk of bias within studies
Present data on risk of bias of each study and, if available, any outcome-level assessment (see item 12).”
https://www.bmj.com/content/339/bmj.b2700
GidMK (Gideon Meyerowitz-Katz) does this grading and sensitivity analysis in his own peer-reviewed IFR work, as per Ioannidis’ PRISMA guidelines. Yet Ioannidis’ doesn’t bother to for his study. So GidMK did it for him, finding that removing the most egregiously non-representative studies roughly doubles Ioannidis’ median IFR to ~0.5%. I did a similar analysis and found the same thing.
Basically, Ioannidis’ median IFR estimate of 0.23% is nonsense, as is the process Ioannidis used to get that, and he tried to get away with it by violating the very guidelines he helped establish. Sources below explaining others issues with Ioannidis’ study, along with covering higher IFR estimates:
“Serious flaws and errors in the methods and data render the study conclusions misinformative. The results and conclusions of the ideal study are at least as likely to conclude the opposite of its results and conclusions than agree. Decision-makers should not consider this evidence in any decision.
[…]
In addition to the major concerns described above, I must comment that this paper does not follow the norms of systematic review. For example, it does not include a PRISMA checklist, it does not include a flow diagram, and the description of the search strategy appears to be incomplete and insufficient for replication.(6)”
https://rapidreviewscovid19.mitpress.mit.edu/pub/p6tto8hl/release/1
“The author combined all seroprevalence estimates into one analysis, and this is a central limitation of this study. […] There is great variability in the quality of studies performed. A quality assessment of each study would have been beneficial to ensure that the included studies were of adequate quality for inclusion, and so that any bias was characterized.
[…]
However, the infection fatality rate estimates by Ioannidis et al. are lower than those reported in multiple other studies.”
https://www.publichealthontario.ca/-/media/documents/ncov/research/2020/12/synopsis-ioannidis-studies-covid-19-infection-fatality-rates.pdf?la=en
Sweden’s Public Health Agency:
“Globally, it is estimated that 0.5–1 percent of those who are infected with COVID-19 die.”
http://web.archive.org/web/20200901075237/https://www.folkhalsomyndigheten.se/the-public-health-agency-of-sweden/communicable-disease-control/covid-19/
October 2020, WHO official:
“Several of these analyses have used published or pre-print seroepidemiologic results and they all converge around a point estimate of around 0.6%.
That may not sound like a lot but that is a lot higher than influenza and the infection fatality ratio increases dramatically with age so we can provide you with these papers, these papers are published but there is a big increase in the infection fatality ratio by age but overall these converge around 0.6%.”
https://www.who.int/publications/m/item/covid-19-virtual-press-conference-transcript—12-october-2020
In the peer-reviewed literature, there are many IFR estimates better than Ioannidis’ low-ball outlier. For example:
GidMK in IJID:
0.8% (0.4% – 1.2%) [various regions, for studies with low risk of bias] : “A systematic review and meta-analysis of published research data on COVID-19 infection-fatality rates” [all studies: 0.7% (0.5% – 0.8%) ; updated to 1.1% (0.7% – 1.5%) for increased number of low-risk-of-bias studies in GidMK’s tweet below]
https://www.sciencedirect.com/science/article/pii/S1201971220321809
https://twitter.com/GidMK/status/1285020775892709377
GidMK on the IJID paper above and his EJE age-specific metagression to come later:
“I think our initial estimate from the IJID paper above was good at using the available data at the time but if you use the more updated figures from our metaregression of COVID IFR by age you can get quite specific death rates in a region”
https://archive.is/PebfT#selection-5511.0-5511.238
GidMK’s age-specific metaregression in EJE:
0.5% – 2.7% [developed countries] : “Assessing the age specificity of infection fatality rates for COVID-19: Systematic review, meta-analysis, and public policy implications” (figure 6; updates the IJID meta-analysis above)
https://link.springer.com/article/10.1007/s10654-020-00698-1
Okell et al.:
0.5% – 1.0% [western Europe] : “Have deaths from COVID-19 in Europe plateaued due to herd immunity?”
https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)31357-X/fulltext
[peer-reviewed, as per: https://www.imperial.ac.uk/news/198238/declines-covid-19-cases-herd-immunity-says/ ]
O’Driscoll et al.:
0.8% (0.7% – 0.9%) [various regions in Europe, USA, South America, + Africa] : “Age-specific mortality and immunity patterns of SARS-CoV-2”
https://www.nature.com/articles/s41586-020-2918-0
https://twitter.com/GidMK/status/1316511734115385344
https://twitter.com/AtomsksSanakan/status/1341183815176364038
Your quite wrong point about low HIT and low IFR is also unscientific and simple minded. HIT was probably low this summer when the epidemic almost died out in many areas. This winter its a lot higher because Rt went up because there were more infectious strains, etc., etc.
David –
OK. Have it your way. You say some 170 million Americans have already been infected. With another 20 milliion vaccinated…so close to 2/3 of the population are already immune.
I guess you think that the epidemiologists were pretty close after all in their estimates that a high % would need to be infected or vaccinated for us to reach herd immunity.
Another cheap rhetorical flourish. You have no idea how many have been infected and neither do I. What I actually said was that HIT could have been achieved in the summer and then gone away in the winter.
Joshua, even with COVID-19 deaths being under-estimated (especially early in the pandemic, consistent with excess mortality estimates), the CDC’s stated number of infections implies an IFR of ~0.5%, with 81 millions infections and >400,000 COVID-19 deaths. One should remember to include the lag between infection and reported deaths, due to infected people taking awhile to progress to death, the time it takes to report + log deaths, etc. That makes it even more ridiculous that Ioannidis abused his low fatality rate estimate to predict 400,000 reported COVID-19 deaths:
https://covid19.who.int/region/amro/country/us
https://coronavirus.jhu.edu/us-map
“In all these places, the numbers of infections (many with no symptoms) — when adjusted for the U.S. population as a whole — suggest a fatality rate that is actually similar to that of the seasonal flu. Data from Iceland and Denmark, which have done the best random sampling, also point in the same direction, Ioannidis said. *“If I were to make an informed estimate based on the limited testing data we have, I would say that covid-19 will result in fewer than 40,000 deaths this season in the USA,” he told me.”
https://www.washingtonpost.com/opinions/without-mass-testing-were-flying-blind-through-this-crisis/2020/04/09/bf61e178-7a9b-11ea-a130-df573469f094_story.html
[ http://archive.is/dT97F#selection-2211.0-2219.279 ]
Ioannidis:
“If we assume that case fatality rate among individuals infected by SARS-CoV-2 is 0.3% in the general population — a mid-range guess from my Diamond Princess analysis — and that 1% of the U.S. population gets infected (about 3.3 million people), this would translate to about 10,000 deaths.”
https://www.statnews.com/2020/03/17/a-fiasco-in-the-making-as-the-coronavirus-pandemic-takes-hold-we-are-making-decisions-without-reliable-data/
Dpy loved citing the CDC in support of Ioannidis when he though the CDC showed an IFR of 0.26%. I wonder what the excuse is going to be now?
dpy:
“Omitted were current CDC estimates […]”
https://judithcurry.com/2020/06/28/the-progress-of-the-covid-19-epidemic-in-sweden-an-analysis/#comment-920111
David –
> You have no idea how many have been infected and neither do I.
We have 450k dead from COVID. You think the IFR is 0.27%. Do the math.
David –
Also. The CDC thinks that maybe 80 million have been infected, or it could be 1/3 the population. So the IFR then would between @ 0.59% and @ 0.43% (allowing that some 30k or so of those already infected will die).
But you know better than the CDC.
Because you’re a scientist and 🦁s.
Given what’s going on with variations – which may well necessarily be a function of the number of infections – the “let it rip” advocacy may turn out to be the among most ill-advised, irresponsible, and destructive ideas in history.
But 🦁s.
We are back to anonymous political activists using Twitter tweets as proof texts.
In addition they use the bad faith tactic of misrepresenting the results of the scientist they smear. Ioannidis’ meta-analysis found IFR ranged from 0.02% th 0.67% depending on a number of factors.
Atomsk –
> Dpy loved citing the CDC in support of Ioannidis when he though the CDC showed an IFR of 0.26%. I wonder what the excuse is going to be now?
-snip-
Across 32 different locations, the median infection fatality rate was 0.27% (corrected 0.24%).
-snip-
It’s really funny that David was over and over telling me how ioanniidis’ best estimate of 0.27% IFR was certainly the most credible (he’s a 🦁 doncha know) and now when I point out the implausibility of that number he references the range in Ioannidis’ estimates.
I do agree that a single number has limited utility. It’s almost like it’s different diseases in different age cohorts. However, it is the “let it rip” crowd that has been promulgating that low IFR to justify their policy advocacy. Hence, they should show accountability for the death and suffering that is resulting from their erroneous analyses.
Joshua, it is pretty ironic, especially since dpy was once telling us how IFR was not as high as 0.5% in the “best” studies. Now he’s trying to expand the range to include 0.5% so he can avoid admitting he was wrong. How telling. It’s like he expects us not to remember what he said. For example:
dpy:
“There are by now at least 10 serologic data sets around the world. They pretty much uniformly show an IFR less than 0.5% with the best ones showing perhaps 0.12% to 0.31%.”
https://statmodeling.stat.columbia.edu/2020/05/08/so-the-real-scandal-is-why-did-anyone-ever-listen-to-this-guy/#comment-1333886
dpy:
“There are at least now 10 more meaningful serological studies from around the world.
1. There is a Danish one of blood donors Joshua pointed out. IFR is 0.08.
2. There is the Santa Clara study which was strengthened by a revision.
3. There is a Los Angeles County study which shows a low IFR too.
4. Miami Dade county which shows an IFR of 0.17-0.31% even when I took fatalities from 21 days after the mean testing date.
5. State of Arizona comes in around 0.28%.
[…]
I personally think the US as a whole is more similar to the 4 US datasets I mentioned above.”
https://judithcurry.com/2020/05/06/covid-discussion-thread-vi/#comment-916993
But like I said, Ioannidis’ median IFR of 0.23% is nonsense + an outlier in the literature. His low estimate likely contributed to his ridiculous prediction that the USA would have <40,000 COVID-19 deaths that season:
“The median infection fatality rate across all 51 locations was 0.27% (corrected 0.23%).”
https://www.who.int/bulletin/volumes/99/1/20-265892/en/
Oh well. Ioannidis can head back to rightly warning us about the dangers of anthropogenic climate change, as he should and as he has done before, despite his debunked views of COVID-19. I wonder how dpy will try to get around that?
Ioannidis:
“The anti-vaccine movement and climate emergency deniers are already drawing ammunition from the reversals of opinion and policy during the pandemic. Clearly the strategy of deniers is inappropriate, since the knowledge we have about covid-19 and how to handle it is still evolving, while we have solid evidence about the efficacy and safety of the measles, mumps, and rubella vaccine or the dangers of climate emergency.”
https://www.bmj.com/content/371/bmj.m4048.full
Ioannidis on how anthropogenic climate change is on the same level of certainty as smoking killing people:
17:17 – 18:22 of: http://rationallyspeakingpodcast.org/show/rs-174-john-ioannidis-on-what-happened-to-evidence-based-med.html
Ioannidis:
“Many fields lack the high reproducibility standards that are already used in fields such as air pollution and climate change. […] It is a scandal that the response of governments to climate change and pollution has not been more decisive.”
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5933781/
Everything that Sanakan quotes from me in May was true and correct at the time. I mentioned specific studies and correctly summarized their results.
It’s mind reading to suggest that I am now trying to “change” the upper bound from May. More data from serologic studies has come in since the quotes you quote mined from last May. It is indeed true that the range of IFR estimates has broadened. That’s called science.
People must beware that Sanakan cherry picks a few statements in an area where there is tremendous controversy and in which the science is changing rapidly. The median IFR is not very meaningful. The range is 0.02% to 0.67%. The median is not “ridiculous.” There are many countries with very young populations where IFR will be very low such as in Africa.
Just as one example he quotes from a study by Meyerowitz-Katz but in this very thread VTG has a paper that is very critical of this very study. Sanakan then asserts he did a similar study. Was it equally flawed? Or perhaps in a controversial and politicized area of science, there is a lot of harsh criticism, much of it unjustified.
In any case, IFR is strongly dependent on the age structure of the population as any expert knows. Ioannidis’ was one of the first to point how important that was. It is no mystery here. It is expected that IFR estimates will have a broad range. Using such uncivil characterizations as “ridiculous” is itself an act of rhetorical theatre.
Re: “Everything that Sanakan quotes from me in May was true and correct at the time. I mentioned specific studies and correctly summarized their results.”
No, your claims were wrong from the moment you made them. I knew that since I’m an immunologist who knows this field well. For example, we’ve known for decades that blood donors, non-targeted volunteers, etc. are not representative of the population, and thus studies based on them will not give accurate population-wide fatality rates. After all, SARS-CoV-2 is not the first pathogen for which we’ve needed to do seroprevalence studies or needed to calculate fatality rates. There was HIV, the 2009 H1N1 pandemic, etc.
Yet you, dpy, went on citing studies with non-representative sampling of blood donors (ex: Denmark), samples that left out older people (ex: Denmark), non-targeted volunteers (ex: Santa Clara, Arizona), faulty tests and/or faulty reporting (ex: Miami-Dade [ https://umiamihealth.org/en/coronavirus/spark-c ]), and so on. That would have been as wrong in 2010 as it was in 2020.
You really need to look up what “representative/randomized sampling” is. The following should get you started, for whenever you finally decide you actually want to learn about this topic:
“Sampling frame
To assess prevalence in the general population, a study should be specifically designed to utilize a random sample using standard survey procedures such as stratification and weighting by demographic characteristics. Other sampling frames may be useful for specific purposes such as sentinel surveillance but not well-suited for assessing prevalence due to substantial risk of systemic bias. Consequently, our meta-analysis excludes the following types of studies:
Blood donors. Only a small fraction of blood donors are ages 60 and above—a fundamental limitation in assessing COVID-19 prevalence and IFRs for older age groups—and the social behavior of blood donors may be systematically different from their peers [13, 18]. These concerns can be directly investigated by comparing alternative seroprevalence surveys of the same geographical location. As of early June, Public Health England (PHE) reported seroprevalence of 8.5% based on specimens from blood donors, whereas the U.K. Office of National Statistics (ONS) reported markedly lower seroprevalence of 5.4% (CI: 4.3–6.5%) based on its monitoring of a representative sample of the English population [19, 20].
Dialysis centers. Assessing seroprevalence of dialysis patients using residual sera collected at dialysis centers is crucial for gauging the infection risks faced by these individuals, of which a disproportionately high fraction tend to be underrepresented minorities. Nonetheless, the seroprevalence within this group may be markedly different from that of the general population. For example, a study of U.K. dialysis patients found seroprevalence of about 36%, several times higher than that obtained using a very large random sample of the English population [21, 22]. Similarly, a recent U.S. study found a seropositive rate of 34% for dialysis patients in New York state that was more than twice as high as the seroprevalence in a random sample of New York residents [10, 23].
Hospitals and urgent care clinics. Estimates of seroprevalence among current medical patients are subject to substantial bias, as evident from a pair of studies conducted in Tokyo, Japan: One study found 41 positive cases among 1071 urgent care clinic patients, whereas the other study found only two confirmed positive results in a random sample of nearly 2000 Tokyo residents (seroprevalence estimates of 3.8% vs. 0.1%) [24, 25].
Active recruitment. Soliciting participants is particularly problematic in contexts of low prevalence, because seroprevalence can be markedly affected by a few individuals who volunteer due to concerns about prior exposure. For example, a Luxembourg study obtained positive antibody results for 35 out of 1807 participants, but nearly half of those individuals (15 of 35) had previously had a positive live virus test, were residing in a household with someone who had a confirmed positive test, or had direct contact with someone else who had been infected [26].”
https://link.springer.com/article/10.1007/s10654-020-00698-1
https://wwwnc.cdc.gov/eid/article/26/9/20-1840_article
https://link.springer.com/article/10.1007/s11357-020-00253-6
https://cdc.gov/nceh/casper/overview.htm
http://web.archive.org/web/20201016231417/https://apps.who.int/iris/bitstream/handle/10665/331656/WHO-2019-nCoV-Seroepidemiology-2020.1-eng.pdf?sequence=1&isAllowed=y
https://sci-hub.se/https://www.jclinepi.com/article/S0895-4356(12)00079-0/fulltext
https://ncbi.nlm.nih.gov/pmc/articles/PMC7454696/
https://twitter.com/AtomsksSanakan/status/1341286778884550656
Sanakan jumps the shark with this whopper. He says what I said in May is incorrect but didn’t quote a single thing that was wrong. That’s just proof by assertion and means nothing.
Re: “Sanakan jumps the shark with this whopper. He says what I said in May is incorrect but didn’t quote a single thing that was wrong. That’s just proof by assertion and means nothing.”
False. For example, folks can see where I clearly showed you studies with very non-representative sampling (ex: on blood donors in Denmark), and then where I cited evidence on why you were wrong to use such non-representative studies for IFR estimates as if they were credible.
Another example was when you said this:
“By Ferguson’s own estimation, 2/3 of those who die with covid19 would have died within the year because they were already seriously ill. That’s borne out by the high percentage of fatalities among nursing home residents. Thus for policy making, 1 million becomes 333,000, about 13% excess mortality.”
https://archive.is/RvDuP#selection-19645.156-19645.457
“So, the 2/3 number is Ferguson’s estimate and I think is perhaps a lower bound.”
https://archive.is/RvDuP#selection-20281.661-20281.740
That was wrong for a number of reasons. First, 2/3 was not Ferguson’s lower-bound estimate. It was an upper-range of a guess he gave in response to a question and after admitting we didn’t know. So you misrepresented what he said:
““We don’t know what the level of excess deaths will be in this epidemic,” Ferguson said. In other words, we don’t know the extent to which COVID-19 will increase annual deaths above the level that otherwise would have been expected. “By the end of the year, what proportion of those people who’ve died from COVID-19 would have died anyhow?” Ferguson asked. “It might be as much as half to two-thirds of the deaths we’re seeing from COVID-19, because it’s affecting people who are either at the end of their lives or in poor health conditions. So I think these considerations are very valid.””
https://archive.is/RvDuP#selection-20611.0-20615.593
[ https://parliamentlive.tv/Event/Index/2b1c71d4-bdf4-44f1-98fe-1563e67060ee ]
Second, years of life lost (YLL) is much greater than 1, even if COVID-19 deaths disproportionately occurred among elderly people. You likely conflated one’s life expectancy at birth, with how much longer one is expected to live if one reaches an elderly age like 80. If life expectancy at birth is 80 years, then that estimate includes many people who die before 80. But that’s not what’s relevant to YLL. What’s instead relevant is how much longer one can expect to live if one reached 80. That number does not include those who die before 80, thereby yielding a higher YLL than what would naively infer from life expectancy at birth. For example:
“Using the standard WHO life tables, YLL per COVID-19 death was 14 for men and 12 for women. After adjustment for number and type of LTCs, the mean YLL was slightly lower, but remained high (13 and 11 years for men and women, respectively). The number and type of LTCs led to wide variability in the estimated YLL at a given age (e.g. at ≥80 years, YLL was >10 years for people with 0 LTCs, and <3 years for people with ≥6).
[…]
Deaths from COVID-19 represent a substantial burden in terms of per-person YLL, more than a decade, even after adjusting for the typical number and type of LTCs found in people dying of COVID-19."
https://wellcomeopenresearch.org/articles/5-75
Third, excess deaths have been larger than reported COVID-19 deaths in many places. Moreover, some countries with strict lockdowns that kept their COVID-19 deaths low, had negative excess deaths, meaning that less people died than usual. This overall pattern is one would expect from reported COVID-19 deaths under-estimating actual COVID-19 deaths, along with policies like lockdowns mitigating deaths from things like traffic accidents, etc. But this overall pattern goes against your claim on excess mortality.
Fourth, it’s well-established that COVID-19 is a risk not only to the elderly, but to the middle-aged as well. For example:
“These results indicate that COVID-19 is hazardous not only for the elderly but also for middle-aged adults, for whom the infection fatality rate is two orders of magnitude greater than the annualized risk of a fatal automobile accident and far more dangerous than seasonal influenza.”
https://link.springer.com/article/10.1007/s10654-020-00698-1
“Do not neglect SARS-CoV-2 hospitalization and fatality risks in the middle-aged adult population
[…]
These findings support the need for comprehensive preventive measures to help reduce the spread of the virus, even in young or middle-aged adults.”
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7836556/
I could go over some of your other false claims on this, but this should suffice for illustrating your long history of making debunked, ideologically-motivated claims meant to illegitimately downplay the danger of COVID-19.
There’s a reason Andrew Gelman said this when you were rebutted on this topic:
“David:
When people point out false or misleading things you’ve said, it’s inappropriate to give them “the bronze medal for nit picking.” This is a statistics blog. Details matter, and if you don’t care about details, you’re kinda wasting our time here.
If people go to the trouble of carefully reading what you’ve written and they find untrue or misleading statements, you should thank them for their effort, not criticize them for “nit picking.” If you can’t appreciate when people correct you, you should avoid open discussion forums like this blog.”
https://archive.is/RvDuP#selection-20965.0-20973.427
Sources below on the 3rd reason mentioned above:
“Managing COVID-19 spread with voluntary public-health measures: Sweden as a case study for pandemic control”
“Quantifying the impact of non-pharmaceutical interventions during the COVID-19 outbreak – The case of Sweden”
“Reduced mortality in New Zealand during the COVID-19 pandemic”
“Physical distancing interventions and incidence of Coronavirus Disease 2019: Natural experiment in 149 countries”
“Inferring the effectiveness of government interventions against COVID-19”
“Estimating the impact of physical distancing measures in containing COVID-19: an empirical analysis”
“Have deaths from COVID-19 in Europe plateaued due to herd immunity?”
“Lockdown timing and efficacy in controlling COVID-19 using mobile phone tracking”
“Estimating the burden of SARS-CoV-2 in France”
“Nudges against pandemics: Sweden’s COVID-19 containment strategy in perspective”
“Estimating the effects of non-pharmaceutical interventions on COVID-19 in Europe”
“The COVID-19 infection in Italy: A statistical study of an abnormally severe disease”
“Impact of lockdown on the epidemic dynamics of COVID-19 in France”
“Shelter-in-place orders reduced COVID-19 mortality and reduced the rate of growth in hospitalizations”
“Effects of non-pharmaceutical interventions on COVID-19 cases, deaths, and demand for hospital services in the UK: a modelling study”
“Modelling SARS-COV2 spread in London: Approaches to lift the lockdown”
“The origin and underlying driving forces of the SARS-CoV-2 outbreak”
“The effect of large-scale anti-contagion policies on the COVID-19 pandemic”
“Spread and dynamics of the COVID-19 epidemic in Italy: Effects of emergency containment measures”
“The positive impact of lockdown in Wuhan on containing the COVID-19 outbreak in China”
https://www.bloomberg.com/opinion/articles/2020-09-17/child-mortality-covid-19-lockdowns-may-have-saved-kids-lives
“Cross-country comparisons of COVID-19: Policy, politics and the price of life”
“Impact of COVID-19 lockdown policy on homicide, suicide, and motor vehicle deaths in Peru”
Not peer-reviewed: “The effectiveness of lockdowns: Learning from the Swedish experience” ( http://archive.is/yo5Ac#selection-985.0-985.268 )
Non-peer-reviewed: “Do lockdowns bring about additional mortality benefits or costs? Evidence based on death records from 300 million Chinese people”
Non-peer-reviewed: “Sheltering in place and the likelihood of non-natural death”
On excess deaths:
https://ourworldindata.org/excess-mortality-covid
https://www.economist.com/graphic-detail/2020/07/15/tracking-covid-19-excess-deaths-across-countries
https://www.nytimes.com/interactive/2020/04/21/world/coronavirus-missing-deaths.html [ http://archive.is/ZCoO3 ]
https://www.ft.com/content/a26fbf7e-48f8-11ea-aeb3-955839e06441
https://www.bbc.com/news/world-53073046
“Excess deaths from COVID-19 and other causes, March-April 2020”
“Estimation of excess deaths associated with the COVID-19 pandemic in the United States, March to May 2020” [with: https://www.cdc.gov/nchs/nvss/vsrr/covid19/excess_deaths.htm ]
“Excess deaths associated with COVID-19, by age and race and ethnicity — United States, January 26–October 3, 2020”
“Magnitude, demographics and dynamics of the effect of the first wave of the COVID-19 pandemic on all-cause mortality in 21 industrialized countries” (extended data table 1)
“COVID-19 and excess all-cause mortality in the US and 18 comparison countries”
“Estimating total excess mortality during a COVID-19 outbreak in Stockholm, Sweden”
“Excess mortality: the gold standard in measuring the impact of COVID-19 worldwide?”
https://twitter.com/jburnmurdoch/status/1354158357754601472
https://twitter.com/GidMK/status/1336493624490487808
No one will read your long winded and very vague nonsense.
I’ll just point out one misrepresentation about blood donor studies. No one ever said they were “representative” samples. People have said they indicate a low IFR in healthy people under 70. You just twisted what people have said out of all context. A typical rhetorical tactic of political activists.
Early serologic studies were not perfect because time was of the essence. Nit picking them 9 months later is another sophist’s trick. Science always is like this.
Re: “No one will read your long winded and very vague nonsense.”
I’m not posting for your benefit. I’m posting for the benefit of people like mesocyclone, Joshua, and verytallguy, who actually read the evidence cited to them. But really, I’m more posting them for the benefit of other people I cite this material to later, regardless of whether Curry lets the comments through moderation or not.
As mesocyclone said:
“You are correct – thank you very much for the links.”
https://judithcurry.com/2021/01/10/covid-19-why-did-a-second-wave-occur-even-in-regions-hit-hard-by-the-first-wave/#comment-942297
I simply don’t care if you read what I say, dpy. It isn’t for you.
Re: “I’ll just point out one misrepresentation about blood donor studies. No one ever said they were “representative” samples.”
Wrong. You said it when you extended their results to the USA, and as a population-wide IFR. I already quoted you doing that. Here’s another example of you doing that for the Danish blood donor study:
“The more recent work with larger datasets mostly show lower IFR’s; Santa Clara, Los Angeles, Miami Dade, Arizona, German and Danish studies.”
https://judithcurry.com/2020/05/06/covid-discussion-thread-vi/#comment-917440
You used that to claim an IFR of 0.08%:
“There are at least now 10 more meaningful serological studies from around the world.
1. There is a Danish one of blood donors Joshua pointed out. IFR is 0.08.”
https://judithcurry.com/2020/05/06/covid-discussion-thread-vi/#comment-916993
Ioannidis did the same thing when he applied those blood donor results to the general population. He combined that with the laughable claim that using blood donors might under-estimate seroprevalence, and thus over-estimate IFR. In reality, blood donors over-estimate seroprevalence for a number of reasons. For example, minors and older people tend to have lower rates of infection, but both groups usually are not active blood donors (children need to wait until their older to donate blood, and elderly people retire out of blood donation). Moreover, blood donors are willing to go out to donate blood during a pandemic, which makes them more likely to be people who socially interact more and thus are at greater risk of infection.
Ioannidis:
“I included studies on blood donors, although they may underestimate seroprevalence and overestimate infection fatality rate because of the healthy volunteer effect.”
[table 4: he infers a ‘corrected’ population-wide IFR of 0.27% inferred from the Denmark blood donor study]
https://www.who.int/bulletin/volumes/99/1/20-265892/en/
reality:
“Blood donors. Only a small fraction of blood donors are ages 60 and above—a fundamental limitation in assessing COVID-19 prevalence and IFRs for older age groups—and the social behavior of blood donors may be systematically different from their peers [13, 18]. These concerns can be directly investigated by comparing alternative seroprevalence surveys of the same geographical location. As of early June, Public Health England (PHE) reported seroprevalence of 8.5% based on specimens from blood donors, whereas the U.K. Office of National Statistics (ONS) reported markedly lower seroprevalence of 5.4% (CI: 4.3–6.5%) based on its monitoring of a representative sample of the English population [19, 20].”
https://link.springer.com/article/10.1007/s10654-020-00698-1
Ironically, the Danish researchers behind the work you and Ioannidis misused, followed that up with another study that actually included older people (retired blood donors). That yielded an IFR an order of magnitude larger than you claimed, dpy, and 3 times larger than Ioannidis claimed. It’s hard to be even more wrong than Ioannidis on COVID-19, but you managed to pull it off.
– old study you and Ioannidis misapplied:
https://academic.oup.com/cid/advance-article/doi/10.1093/cid/ciaa849/5862661
– update that includes older people:
“The IFR for the adult Danish population aged 17 years or older was 0.81% (95% CI: 0.52%-2.2%).”
https://academic.oup.com/cid/advance-article/doi/10.1093/cid/ciaa1627/5939898
Of course, these blood donor studies are irrelevant for sound estimates of population-wide IFR, since their sampling is so non-representative. Instead, there are multiple Denmark studies with representative sampling of the general population and implying population-wide IFRs of ≥ ~0.5%. That’s even more evidence that you and Ioannidis were wrong, dpy.
figure 5: ~0.8% IFR for Denmark
https://www.sciencedirect.com/science/article/pii/S1201971220321809
SSI, phases 1 – 3:
https://web.archive.org/web/20210204214304/https://covid19.ssi.dk/overvagningsdata/undersoegelser/praevalensundersogelsen
“We test Denmark”, phase 1:
https://web.archive.org/save/https://www.vitesterdanmark.dk/
https://web.archive.org/web/20210204214758/https://www.ssi.dk/aktuelt/nyheder/2020/nasten-en-halv-million-danskere-deltog-i-vi-tester-danmark-nu-gar-fase-2-af-projektet-gar-i-gang
https://twitter.com/GidMK/status/1285020775892709377
A further scientific error in your comment Sanakan is suggesting that the 2 Danish studies contradict each other. That’s wrong. They study different populations and the IFR’s might indeed be different. Also the second study you misrepresent. It is a population of adults and so is not an IFR estimate for Denmark at all. It needs to be corrected for the omission of those under 17.
All in all, you throw up a bunch of dust that shows no contradiction or inconsistency. There are by now probably hundreds of attempts to estimate the IFR (which will be vastly different in different countries and regions and age groups). Yet you are proof texting from this vast literature to try to show someone else was wrong. There is no single right or wrong answer here.
Re: “A further scientific error in your comment Sanakan is suggesting that the 2 Danish studies contradict each other.”
Never said they contradict each other. Instead, they together contradict your repeated attempts to use non-representative sampling to under-estimate IFR, including by excluding older people, relying on blood donor studies, etc.
Re: “Also the second study you misrepresent. It is a population of adults and so is not an IFR estimate for Denmark at all. It needs to be corrected for the omission of those under 17”
Same point applies to the blood donor study you and Ioannidis cited to under-estimate IFR. It omits those under 17, yet neither you nor Ioannidis corrected for that. It’s actually even worse than the later blood donor study, since the study you cited also left out those above the age of 69, which neither you nor Ioannidis corrected for either:
“Danish blood donors aged 17–69 years […]”
https://academic.oup.com/cid/article/72/2/249/5862661
In fact, Ioannidis states he doesn’t need to correct for the omission of those under the age of 17 and over the age of 69. Did you actually read the paper of his you keep citing as being the work of a “lion”?:
“I assumed that the seroprevalence would be similar in different age groups, but I also recorded any significant differences in seroprevalence across age strata so as to examine the validity of this assumption.
[…]
Furthermore, many studies have evaluated people within restricted age ranges, and the age groups that are not included may differ in seroprevalence. Statistically significant, modest differences in seroprevalence across some age groups have been observed in several studies.10,13,15,23,27,36,38 Lower values have been seen in young children and higher values in adolescents and young adults, but these patterns are inconsistent and not strong enough to suggest that major differences are incurred by extrapolating across age groups.”
https://www.who.int/bulletin/volumes/99/1/20-265892/en/
Re: “All in all, you throw up a bunch of dust that shows no contradiction or inconsistency.”
Nope. IFR is higher than you and Ioannidis have been saying for months. Both of you used incorrect methods to under-estimate IFR. You’re not going to re-write history on that.
“This Pub is a Review of
The infection fatality rate of COVID-19 inferred from seroprevalence data
by John Ioannidis
[…]
Misleading. Serious flaws and errors in the methods and data render the study conclusions misinformative. The results and conclusions of the ideal study are at least as likely to conclude the opposite of its results and conclusions than agree. Decision-makers should not consider this evidence in any decision.”
https://rapidreviewscovid19.mitpress.mit.edu/pub/p6tto8hl/release/1
“However, the infection fatality rate estimates by Ioannidis et al. are lower than those reported in multiple other studies.”
https://www.publichealthontario.ca/-/media/documents/ncov/research/2020/12/synopsis-ioannidis-studies-covid-19-infection-fatality-rates.pdf?la=en
This last comment shows why Sanakan is completely untrustworthy and Nic and I expect his comments to be biased.
His penultimate reference says far more than he quotes. The summary of the reference says: “This study finds substantial heterogeneity in the infection fatality rate across different locations. Data are useful and add to the emerging picture on IFR, however, substantial conclusions cannot be drawn.”
Further Flaxman has a terrible track record. His paper on lockdown effects was devastated and destroyed in frontiers for essentially choosing a model form that guaranteed the result. It shows the result is totally wrong, not just biased, but wrong.
https://www.frontiersin.org/articles/10.3389/fmed.2020.580361/full?&utm_source=Email_to_authors_&utm_medium=Email&utm_content=T1_11.5e1_author&utm_campaign=Email_publication&field=&journalName=Frontiers_in_Medicine&id=580361
“The model of Flaxman et al. (3) contradicts this elementary insight. They estimate R(t) from daily deaths associated with SARS-CoV-2 using as an a priori restriction that R(t) may only change at those dates where interventions become effective. Such an approach does not prove that NPIs were effective but rather begs the result, i.e., involves circular logic. The true effective reproduction number declines continuously, and when its estimates are allowed to change only at intervention points, it is clear that profound discontinuities, which attribute strong effects to the interventions, will emerge.”
Flaxman is not reliable in my book.
But the proof texting activist doesn’t have time to examine all relevant data. He cherry picks what is required to support his political agenda.
Since this is the first thing I checked and it was so weak, I am forced to the conclusion that I’ve verified quite a few times. Sanakan’s references and cherry picked quotes are almost always misleading or outright wrong.
And no the two Danish blood donor studies you cite do not contradict Ioannidis’ analysis. All three are of different populations. Another totally unbalanced comment from Sanakan.
I’ll just highlight another tactic by Sanakan, namely, cherry picking early public guesses over the real science. Cherry picking by ignoring the author’s later work. Ioannidis’ meta-analysis which is his later work after vastly more evidence was available finds IFR form 0.67% to 0.02% depending on the setting. That is not an outlier result.
No one has claimed that a Danish blood donor study was a “general population” IFR. To claim otherwise shows a lack of candor and a political or personal motive.
Re: “His penultimate reference says far more than he quotes. The summary of the reference says:”
Which doesn’t change what I quoted from it, showing that Ioannidis work was unreliable and should not be trusted in policy-making. Pointing out a source says X doesn’t change the fact that it also said Y, dpy. You’ve had this explained to you multiple times for years.
Re: “Ioannidis’ meta-analysis which is his later work after vastly more evidence was available finds IFR form 0.67% to 0.02% depending on the setting. That is not an outlier result.”
No, it’s an outlier, as per its median IFR estimate of 0.23%, and its lower-bound IFR is much lower than that of other studies. I already cited IFR estimates showing this earlier in the thread ( https://archive.is/iag5s#selection-34201.0-34405.17 ). So the source I quoted on this was correct.
Also, you’ve repeatedly given this inaccurate “0.67% to 0.02%” range, despite Ioannidis’ paper being cited to you and that paper showing his actual range. So I again suspect you didn’t actually read his paper. Alternatively, you could be cherry-picking his IFR for regions with a lower than average COVID-19 mortality rate, and presenting that as his full range. In that case, you’re misrepresenting his work:
“Inferred infection fatality rate estimates varied from 0.00% to 1.63% (Table 4). Corrected values also varied considerably (0.00–1.54%).
[…]
The median infection fatality rate across all 51 locations was 0.27% (corrected 0.23%). […] Uncorrected estimates of the infection fatality rate of COVID-19 ranged from 0.01% to 0.67% (median 0.10%) across the 19 locations with a population mortality rate for COVID-19 lower than the global average, from 0.07% to 0.73% (median 0.20%) across 17 locations with population mortality rate higher than the global average but lower than 500 COVID-19 deaths per million, and from 0.20% to 1.63% (median 0.71%) across 15 locations with more than 500 COVID-19 deaths per million. The corrected estimates of the median infection fatality rate were 0.09%, 0.20% and 0.57%, respectively, for the three location groups.”
https://www.who.int/bulletin/volumes/99/1/20-265892/en/
Re: “But the proof texting activist doesn’t have time to examine all relevant data. He cherry picks what is required to support his political agenda.”
Ironic, since given what I showed above, you either have not “examine[d] all relevant data” in Ioannidis’ paper, or you “cherry pick[ed] what is required” to show a lower IFR range.
Anyway, Ioannidis’ work suffers from even more obvious, crippling flaws. For example, the population fatality rate, or PFR (i.e. COVID-19 deaths per capita), should always be less than IFR, since you can’t have more people infected than actually exist and herd immunity prevents 100% of people from being infected. Yet in at least 3 instances, Ioannidis gives ‘corrected’ IFRs that are less than PFR. That’s not even touching on the instances in which his IFR is so close to PFR, that he requires more people be infected than Nic Lewis’ low HIT would allow. So Ioannidis’ IFR is such a horrible under-estimate that it requires more people be infected than actually exist. His IFR estimate is therefore mathematically impossible; it also contradicts Nic Lewis’ claim of a low HIT.
Using table 4 from his paper, here are 3 regions in which Ioannidis’ IFR is less than PFR:
– Rio de Janeiro (state)
Ioannidis’ IFR: 0.11%
PFR: 0.19%
https://transparencia.registrocivil.org.br/especial-covid
– Scotland
Ioannidis’ IFR: 0.07%
PFR: 0.15%
https://coronavirus.data.gov.uk/details/deaths?areaType=nation&areaName=Scotland
– San Francisco
Ioannidis’ IFR: 0.00%
PFR: 0.04%
https://coronavirus.jhu.edu/us-map
Interestingly, Nic Lewis did the same thing, where his IFR for Sweden is mathematically impossible. Lewis projected Sweden would end up with an IFR of ~0.06%, even though ~0.12% of Sweden’s population has already died of COVID-19. His IFR is therefore less than PFR, requiring more people be infected than actually exist:
“In the absence of a change in trends, it seems likely that the epidemic will peter out after a thousand or so more deaths, implying an overall infection fatality rate of 0.06% of the population (0.04% excluding COVID-19 deaths of people in care homes).”
https://judithcurry.com/2020/06/28/the-progress-of-the-covid-19-epidemic-in-sweden-an-analysis/
“I also projected, based on their declining trend, that total COVID-19 deaths would likely only be about 6,400. Subsequent developments support those conclusions.”
https://judithcurry.com/2020/07/27/why-herd-immunity-to-covid-19-is-reached-much-earlier-than-thought-update/
Sweden’s 0.12% PFR and ~12,000 COVID-19 deaths, for its population of ~10 million people:
https://archive.is/c3V0V
https://coronavirus.jhu.edu/data/mortality
https://covid19.who.int/region/euro/country/se
Re: “This last comment shows why Sanakan is completely untrustworthy and Nic and I expect his comments to be biased.
[…]
Flaxman is not reliable in my book.”
Ironic, since Nic Lewis’ position is mathematically impossible, something that I don’t think can be said of Flaxman’s work. Yet you still rely on Nic Lewis’ work, along with Ioannidis’ research, despite the fact that Ioannidis’ research is also mathematically impossible and contradicts Lewis’ claims on HIT being low.
Maybe you’re not the best judge of who’s “not reliable”, “untrustworthy,” and “biased”, dpy?
https://twitter.com/AtomsksSanakan/status/1358015764658343937
I did an analysis demonstrating Ioannidis’ statements about IFR because its quite obvious that apparent values will vary over a large range depending on the risk groups infected. That’s why the estimates from serological studies that include demographic information are the best ones.
I will start with Ferguson’s age cohort IFR estimates. Virtually every serologic study in the US shows his IFR’s are at least a factor of 2 too high. This gives me the following values. I have combined the 0-9 and 10-19 cohorts and the 20-19 and 30-39 cohorts.
Age cohort. IFR. % of US population
1. 0-19 0.002% 27%
2. 20-49 0.0275% 28%
3. 50-59 0.3% 14%
4. 60-69 1.1% 14%
5. 70-79 2.5% 11%
6. 80-90 4.6% 4.2%
Total IFR: 0.67%
Its now easy to do some calculations on apparent IFR’s depending on the age profile of those infected. For reference, in the US, expected mortality is about 2,840,000 per annum.
VTG says of Ioannidis’s covid work.
“(1) it has been rapidly disseminated and publicised in the right wing media, pushed by people including yourself.”
This is irrelevant to the science and is unimportant.
“(2) It has consistently been an outlier in the rest of the science”
This is an obvious falsehood. Just to take the most controversial of ioannidis’ papers on the Santa Clara Study. There are many serologic studies that arrive at similar estimates for the IFR, including Los Angeles county and Miami-Dade county. There are others that arrive at higher estimates. Ioannidis has a meta-analysis on this with values ranging from 0.02% to 0.66%.
“(3) It has attracted very strong criticism from the rest of the field”
Once again this is irrelevant and typical of normal science in a controversial field.
Duplicating this so it is easy to find.
I did an analysis demonstrating Ioannidis’ statements about IFR because its quite obvious that apparent values will vary over a large range depending on the risk groups infected. That’s why the estimates from serological studies that include demographic information are the best ones.
I will start with Ferguson’s age cohort IFR estimates. Virtually every serologic study in the US shows his IFR’s are at least a factor of 2 too high. This gives me the following values. I have combined the 0-9 and 10-19 cohorts and the 20-19 and 30-39 cohorts.
Age cohort. IFR. % of US population
1. 0-19 0.002% 27%
2. 20-49 0.0275% 28%
3. 50-59 0.3% 14%
4. 60-69 1.1% 14%
5. 70-79 2.5% 11%
6. 80-90 4.6% 4.2%
Total IFR: 0.67%
Its now easy to do some calculations on apparent IFR’s depending on the age profile of those infected. For reference, in the US, expected mortality is about 2,840,000 per annum.
Scenario. #0
Age cohort % infected. Fatalities. Infections
1. 5% 1141 4,455,000
2. 10%. 2,541 9,250,000
3. 15%. 20,790 6,950,000
4. 40%. 203,280 18,500,000
5. 60%. 544,500 21,780,000
6. 80%. 510,048 11,088,000
Totals 1,282,300 72,023,000
Apparent IFR: 1.78%
Scenario. #1
Age cohort % infected. Fatalities. Infections
1. 10% 2282 8,900,000
2. 20%. 5,082 18,500,000
3. 30%. 41,580 13,900,000
4. 40%. 203,280 18,500,000
5. 60%. 544,500 21,780,000
6. 80%. 510,048 11,088,000
Totals 1,306,772 92,668,000
Apparent IFR: 1.41%
Beware Sanakan’s cherry picking from the science. Ioannidis’ work is thought of very highly in the medical field. Many of the criticisms of his work are by people with no medical expertise. That being said, its a rapidly evolving field and results will change as more information becomes available.
What we are seeing here is a rapidly developing epidemic where data and facts are changing every day so real scientists will have varying results. Sanakan exploits this by quote mining old papers that often were correct at the time.
The other aspect of this is the nasty partisanship involved. Scientists are not immune to this. Sanakan states on his Twitter home page that his purpose is political consensus enforcement.
Scenario. #2
Age cohort % infected. Fatalities. Infections
1. 10% 2282 8,900,000
2. 20%. 5,082 18,500,000
3. 30%. 41,580 13,900,000
4. 35%. 177,870 16,200,000
5. 40%. 370,260 14,520,000
6. 45%. 255,042 5,544,000
Totals 852,116. 77,564,000
Apparent IFR: 1.10%
Scenario. #3
Age cohort % infected. Fatalities. Infections
1. 40% 7128 35,640,000
2. 30%. 7,623 27,720,000
3. 20%. 27,720 9,240,000
4. 15%. 69,300 4,620,000
5. 10%. 90,750 3,630,000
6. 5%. 31,878 693,000
Totals 234,399. 76,593,000
Apparent IFR: 0.31%
Scenario. #4
Age cohort % infected. Fatalities. Infections
1. 60% 10,692 53,460,000
2. 60%. 15,246 55,440,000
3. 30%. 41,580 13,860,000
4. 10%. 50,820 6,300,000
5. 5%. 45,375 1,815,000
6. 5%. 31,878 693,000
Totals 150,216 129,753,000
Apparent IFR: 0.116%
This also demonstrates the imperative to protect nursing homes from infection. It appears in the US that 40% of all fatalities have taken place in residents of these homes. Some governors like DeSantos in Florida did a good job. Others in New Jersey, New York, and Pennsylvania did a terrible job and cost tens of thousands of lives. This also explains the apparently higher IFR in these locales and the much lower IFR in Florida.
CFR’s are much more uncertain because of massive differences in rates of testing. Any statistical analysis is thus subject to more confounding factors that serological studies with a defined demographic distribution. And then there are the “Covid” death numbers. It’s a matter of intense controversy. Colorado just reduced theirs by 25%.
Ioannidis is aware of all this and is attempting to correct his numbers based on the available data on age structure of those tested. More work needs to be done.
Comical, non-peer-reviewed analysis in defense of Ioannidis long-debunked, outlier IFR estimate.
“Assessing the age specificity of infection fatality rates for COVID-19: systematic review, meta-analysis, and public policy implications
[…]
Our metaregression results are generally consistent with the study of Verity et al. [148] and Ferguson et al. [149], which were completed at an early stage of the COVID-19 pandemic and characterized an exponential pattern of age-specific IFRs (see Supplementary Appendix Q). Other subsequent studies have obtained broadly similar patterns of age-specific IFRs using statistical models to describe the dynamics of transmission and mortality using surveillance data in specific locations [69, 150].
Our findings are well-aligned with a recent meta-analysis of population IFR and indeed explain a high proportion of the dispersion in population IFRs highlighted by that study [151]. In contrast, our findings are markedly different from those of another review [Ioannidis’ paper] of population IFR that includes samples that did not satisfy our inclusion criteria [152].”
https://link.springer.com/article/10.1007/s10654-020-00698-1
“However, the infection fatality rate estimates by Ioannidis et al. are lower than those reported in multiple other studies.”
https://www.publichealthontario.ca/-/media/documents/ncov/research/2020/12/synopsis-ioannidis-studies-covid-19-infection-fatality-rates.pdf?la=en
Ioannidis’ results have not been debunked least of all by an anonymous twitter activist such as you. There is continuing controversy which is normal science. Perhaps that’s something you don’t know much about.
My calculations are indeed proof that IFR’s can vary all over the place.
@AtomsksSanakan – Thanks again for the many links. Once again, my suspicions of Ionnidis appear to be amply justified. It’s quite amazing how bad his work appears to be – to this non-expert.
Beware Sanakan’s cherry picking from the science. Ioannidis’ work is thought of very highly in the medical field. Many of the criticisms of his work are by people with no medical expertise.
What we are seeing here is a rapidly developing epidemic where data and facts are changing every day so real scientists will have varying results. Sanakan exploits this by quote mining old papers that often were correct at the time.
The other aspect of this is the nasty partisanship involved. Scientists are not immune to this. Sanakan states on his Twitter home page that his purpose is political consensus enforcement. He’s a professional hack in other words.
@mesocyclone
You’re welcome. Ioannidis’ review paper is basically dead, as I went over elsewhere ( https://judithcurry.com/2021/01/10/covid-19-why-did-a-second-wave-occur-even-in-regions-hit-hard-by-the-first-wave/#comment-941855 ).
Ioannidis needs to learn not to use non-representative samples as if they’re representative of population-wide IFR (ex: blood donors, dialysis patients, non-targeted volunteers, etc.). In fact, that’s one of the main reasons his co-authored Santa Clara study was nonsense, if it’s treated as a population-wide IFR estimate for Santa Clara county: the study uses non-targeted volunteers. That over-estimates seroprevalence in comparison to a randomized representative sample, because non-targeted volunteers are more likely to seek out testing to confirm their suspicion that they were infected (ex: due to prior exposure to an infected person, or due to previously having symptoms).
Randomized targeting ameliorates this some, by giving the targeted people a reason to show up for testing and by excluding non-targeted volunteers who come for testing. But an even better scenario is to have randomized targeting with a high response rate (at least 60%) among those targeted and/or randomized targeting with adjustment for non-response bias. The Santa Clara study failed on both; it doesn’t even have randomized targeting of the general population. Sources on this below, for the curious:
https://www.medrxiv.org/content/10.1101/2020.04.14.20062463v2#comment-5152013204
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7454696/
“In addition to the 2% infection rate among randomized participants, individuals who participated without an invitation showed a 3.3% rate of infection. This is higher than the random sampling because individuals self-selecting for testing are more likely to have been exposed to the virus.”
https://www.uoflnews.com/post/uofltoday/co-immunity-project-shows-covid-19-infection-rate-in-jefferson-county-increased-tenfold-since-september/
“In general, our analysis shows that naïve estimates that do not account for the non-response bias tend to drive prevalence estimates upward. In contrast to the findings in the literature examining the non-response bias in HIV serosurveys, on average participants who are more likely to have antibodies are more likely to participate in COVID-19 surveys [16; 23]. Participants with history of illness in the last 3 months or past history of tests for COVID-19 in the last 3 months were more likely to agree to antibody testing in our study probably seeking external confirmation.”
https://www.medrxiv.org/content/10.1101/2020.11.02.20221309v1.full.pdf
“Active recruitment. Soliciting participants is particularly problematic in contexts of low prevalence, because seroprevalence can be markedly affected by a few individuals who volunteer due to concerns about prior exposure. For example, a Luxembourg study obtained positive antibody results for 35 out of 1807 participants, but nearly half of those individuals (15 of 35) had previously had a positive live virus test, were residing in a household with someone who had a confirmed positive test, or had direct contact with someone else who had been infected [26].”
https://link.springer.com/article/10.1007/s10654-020-00698-1
– Group I seroprevalence, with targeted sampling: 0.97% (95% CI: 0.72−1.30)
– Group II, made up of non-targeted volunteers: 1.94% (95% CI: 0.84−4.42)
https://www.medrxiv.org/content/10.1101/2020.08.24.20181206v1
https://twitter.com/AtomsksSanakan/status/1341187557326004225
I agree. When I saw Ionnidis’ first paper on COVID, I was suspicious, and then the information came out: a test with too high a false positive rate for the prevalence level like to be there (which magnifies the results by a lot – Bayesian stuff); recruitment that was highly biased. So I have not been likely to believe Ioannidis’ stuff unless I see folks who properly criticized the early work not criticizing the new stuff. At this point, I don’t bother to read his studies – I’ll let others do that. It saddens me as a former Stanford staff member.
CDC did some sort of randomize serology survey last fall or summer, including in my zip code. They reached out and contact folks, and went to their homes, I believe – a good way to get both random and a high response rate.
I don’t know the results they came up with. And they picked a time when antibodies were likely to have faded here.
Just in case meso sees this which is very unlikely, a less out of context cherry picking critique can be found here. It’s biased but at least more honest than Sanakan’s dishonest one.
https://www.covidfaq.co/John-Ioannidis-Stanford-School-of-Medicine-14d05ff6d08943fc8b548383e865a9e4
As Nic said in response to a Sanakan comment last summer: This comment contains many obviously wrong and confused items. It’s not worth my time to demolish them.
@mesocyclone
Re: “CDC did some sort of randomize serology survey last fall or summer, including in my zip code. They reached out and contact folks, and went to their homes, I believe – a good way to get both random and a high response rate.
I don’t know the results they came up with. And they picked a time when antibodies were likely to have faded here.”
I already have a thread on this, since an disease ecology expert requested it on Twitter. My guess is you’re referring to the Maricopa County study. Preliminary results are out already:
https://www.maricopa.gov/CivicAlerts.aspx?AID=1876
https://www.maricopa.gov/5607/COVID-19-Serosurvey
https://research.asu.edu/serosurvey-maricopa-county
https://research.asu.edu/study-shows-covid-19-often-undetected-maricopa-county
https://twitter.com/AtomsksSanakan/status/1354634337539518465
“My guess is you’re referring to the Maricopa County study. ”
You are correct – thank you very much for the links. And from your thread, I take it that you consider this high quality randomized study.
So, do you agree that the study implies, reasonably, that the actual case rate was at least 4X to 5X the PCR detected rate?
And a side question… does it make any sense to get an antibody test after getting vaccinated (and waiting), to see if it “took?”
Atomsk –
In the Santa Clara study they didn’t even control for SES and race/ethnicity to make their sample representative.
I was not trying to be rude. Just observing that fewer and fewer people pay attention as things age. That’s particularly true of Sanakan’s long, verbose and almost always misleading comments. He is the king of cherry picking items with no balance or contrast from the other point of view.
1. With respect the criticism of Ioannidis is normal science.
2. This is a rapidly evolving issue and there is tremendous pressure to find things out quickly. That can lead to oversights.
3. The criticisms of Flaxman, Ferguson, and many others have been vastly harsher. Flaxman’s paper on mitigation effects is as close to fraudulent as its possible to come. The model appears to have been selected to show the effect Flaxman wanted.
4. You need to avoid the logical error of judging an entire body of work and the scores of co-authors by a single paper.
@Joshua
Re: “In the Santa Clara study they didn’t even control for SES and race/ethnicity to make their sample representative.”
The Santa Clara is trash, from top to bottom. Fortunately, there’s a better designed, randomized study being done for that region. So folks can continue disregarding Ioannidis’ work for being the garbage that it is.
http://med.stanford.edu/epidemiology-dept/research/covid-research-collaborative/CA-Facts.html
https://www.ca-facts.org/
“To illustrate some of these biases, we examined characteristics of the underlying population for a seroprevalence study conducted in Santa Clara County, California in early April [8]. Participants were recruited using targeted Facebook advertisements and community listserves. Although researchers tried to recruit such that the distribution of participants would accurately reflect the population geographically, persons in wealthier areas were overrepresented. The sample also underrepresented men, persons 65 and older, Hispanics, and Asians. Researchers weighted the sample such that the weighted distributions of participants by ZIP Code, sex, and race/ethnicity would reflect known county demographics, and weighted estimates of seroprevalence were produced. This approach assumes that participants in the study are like a random sample of Santa Clara residents stratified by ZIP Code, sex, and race/ethnicity. There is some evidence to question the validity of this assumption. Age, a factor correlated with COVID-19 risk [13], was not included as a weighting variable. Although 12.9% of Santa Clara residents are reported as being aged 65 or older, only 4.5% of the weighted sample represented this age group. Furthermore, characteristics such as occupation [14] and social distancing practices [15] are likely drivers of COVID-19 infection and were not accounted for in the weighting or analysis. The results of the Santa Clara study and other convenience samples have been publicized widely in the media [5, 7, 16], and thus the estimates from these studies have the potential to influence policymakers.”
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7454696/
@mesocyclone
Re: “You are correct – thank you very much for the links. And from your thread, I take it that you consider this high quality randomized study.”
It’s good. However, I’m still waiting on a pre-print, report, or published paper where they give more details, like the demographic breakdown of their participants, how they addressed non-response bias, what antibody test they used, etc. The last point is one of the most important because immunoassays for antibodies against nucleocapsid protein (in contrast to those against spike protein, including its receptor-binding domain) have markedly reduced sensitivity overtime, leading to them under-estimating of seroprevalence. Abbott’s anti-nucleocapsid test is notorious for this. This led to a lot of silly claims early on about how large numbers of infected people did not seroconvert and were instead protected by T cells, when in reality this was just due to the test’s decreasing sensitivity.
Re: “So, do you agree that the study implies, reasonably, that the actual case rate was at least 4X to 5X the PCR detected rate?”
Yes, about that range, though I still need more information on the study’s design, as noted above. I’ve said for months that the ascertainment rate is higher than the CDC previously claimed; ratios haven’t been as high as 10X for quite some time.
Re: “And a side question… does it make any sense to get an antibody test after getting vaccinated (and waiting), to see if it “took?””
It’s fine, if you wait a minimum of 2 weeks after getting the last vaccine dose, or at least 2 weeks after the beginning of any symptoms you suffer from the vaccine, to give enough time for seroconversion to occur. Researchers often measure antibody levels anyway when assessing vaccines during clinical trials. However, finding an antibody increase is not a guarantee that those antibodies are helpful, since, for example, one needs to check if the antibodies are neutralizing. Many of the antibody tests given to the general public are not focused on whether the antibodies are neutralizing. Some studies below on testing for neutralizing vs. non-neutralizing antibodies:
https://jcm.asm.org/content/59/2/e02257-20
https://www.sciencedirect.com/science/article/pii/S2666991920000329
https://www.eurosurveillance.org/content/10.2807/1560-7917.ES.2020.25.47.2001752
https://wwwnc.cdc.gov/eid/article/27/2/20-4088_article
https://www.medrxiv.org/content/10.1101/2020.08.20.20178533v1
Also, other places where I and others discussed the points above:
https://twitter.com/CovidSerology/status/1294640133250019328
https://twitter.com/AtomsksSanakan/status/1287542476216172544
Thanks for all the info.
It would be nice if they disclosed more early on this study, the way many others have done for COVID19.
The test I’m considering is an ELISA test for the spike proteins: https://covid19antibodytesting.arizona.edu/home
It might be the one used in the study, but I don’t know.
Since it’s free and they are encouraging its use – even from those who have been vaccinated – I’ll make the 20 mile drive when I’m a few weeks past the second dose. The first dose resulted in very mild symptoms for me – lump on my arm, and 24 hours later a bit of unusual fatigue and minor aches.
Atomsk –
As soon as I heard about that Santa Clara study I looked to see what they did for ppst-stratification adjustments to ensure representativeness – since they extrapolated from that study to project some kind of generalized IFR I was really surprised to see they hadn’t done so. Obviously, on top of the convenience sampling problems and the Facebook advertising and having people drive to the testing sites and all of that, not controlling for such basics as SES before extrapolating seemed amateurish beyond belief. I was truly confused. How could I, as someone with no domain specific expertise see a problem that renown experts just completely ignored? Especially Simone with the reputation and work product such as Ioannidis? That David Young was so supportive of the study helped me to think I was right, but still I just didn’t get it. I kept looking for some valid explanation. Still haven’t found one.
Beating up the Santa Clara study from 7 months ago is such a waste of time. Early studies usually use imperfect data. I must say however that this focus is not a waste of talent as those doing it have little of that. They also show that their time has little value or else they would be spending it on something importance.
David –
> Beating up the Santa Clara study from 7 months ago is such a waste of time. Early studies usually use imperfect data. I must say however that this focus is not a waste of talent as those doing it have little of that. They also show that their time has little value or else they would be spending it on something importance.
Don’t be a snowflake. Pointing out the obvious methodological and conceptual errors isn’t “beating [it] up.” It’s just pointing out the errors. Don’t take it so personally.
Yes, they used imperfect data – and that’s one problem only in the sense that their statistical methodology was flawed. Using preliminary data is a regular practice in such situations, but there’s really no excuse for the unrelated bad statistical methodology.
But that’s not MY point. My points are (1) they had a basic and fundamental problem with their convenience sampling methodology and, perhaps more importantly, (2) beyond the study itself they extrapolated from an unrepresentative sampling.
Point (2) actually has nothing to do with the study per se – but their subsequent policy advocacy campaign they conducted on rightwing media. Of course it’s their prerogative to conduct a policy advocacy campaign, I see nothing wrong with that; but when they did so they shouldn’t have made such a fundamental error.
But hey, if you have no problem with a study that contains fundamental methodological errors at multiple levels, and scientists going on media campaigns to promote policy advocacy based on un-scientific practices – that’s your right. I would never try to take that right away from you. In fact, it’s kind of useful that you’re so open about saying that you have no problems with such bad science.
Thanks.
Josh’s latest response is based on an obviously flawed idea of how science works.
It’s analogous to saying that HADCRUT3 was really bad science because there have been continuing significant adjustments since then that did raise the warming rate a lot. The truth is that HADCRUT3 was the best that could be done at the time with the data available. Calling it bad science is quite uninformed.
The Santa Clara study likewise is not “bad” science. They this last fall published a very detailed Bayesian analysis of the blood test they used and sharpened their estimates. It’s the best that could be done in a very short time frame.
I guess biased people who don’t understand science can believe whatever they want, but I don’t think most real scientists care. Josh here is playing the same role as some nonscientists climate skeptics. But wait, Josh attacks those skeptics but then does what they did. Wow.
David –
> The Santa Clara study likewise is not “bad” science.
It was bad science. I’ll take the criticism of people like Gelman at face value: their statistical methodology was flawed.
It was also bad science because they recruitment and sampling flew in the face of accepted epidemiological practices.
It was also bad science because of the ethics of their methods, where they tested people for antibodies under the promise that a positive test would mean immunity – w/o informing the participants of the potential of false positives and without providing them opportunities for retesting. Their colleague quit the study and filed a whistelblower complaint accordingly. That’s bad science.
And then they followed the bad science of the Santa Clara study with the bad scientific practice of going on a publicity campaign to promote policies based on the amateurish science of extrapolating from non-representative sampling (without making adjustments to ensure representativeness.
Bad science followed by bad science.
>The truth is that HADCRUT3 was the best that could be done at the time with the data available
Its not bad science to analyze preliminary data. It’s not bad science to produce analysis that is necessarily limited due to limits in the quality of data that’s available. It’s bad science to use preliminary data to support conclusions that would require better data – such as truly representative samples or properly adjusted samples.
I’ve explained this to you many times but you still seem to be confused. I hope it sticks this time.
dpy,
there’s not much point in responding to any of your posts, as they seem to disappear so frequently. I presume that’s moderation because of your constant insults to those with different points of view, though I can’t be sure.
So I can only suggest, again, that you keep your postings civil if you wish to have any dialogue.
I don’t want to have a dialogue with you VTG. You have misstated at least 3 times what I have said about the flu ‘disappearing.’ You ignore my references to the literature where hit is calculated using Rt, and you generally act in bad faith. You smear scientists and cite obviously weak papers to further the smear.
For the 5th time what I said was that no one knows why the flu ‘disappeared.’ You have given zero evidence that mitigation caused it. It’s not an act of good faith.
dpy,
again, there’s not much point in responding to any of your posts, as they seem to disappear so frequently. I presume that’s moderation because of your constant insults to those with different points of view, though I can’t be sure.
So I can only suggest, again, that you keep your postings civil if you wish to have any dialogue,
VTG, You are just wrong again. Many comments on this thread have been moderated including some of yours. Mostly it’s been Joshy’s. I am grateful Judith is doing it as this post has attracted more than the usual amount of obfuscatory and silly commentss from anonymous activists, mostly from Josh, you, and Sanakan (perhaps the most uncivil twitter activist I’ve encountered). But you are fine with that showing that your concern with civility is not genuine.
State of NY undercounted nursing home deaths. This confirms much speculation about how many patients had died in long term facilities.
https://apnews.com/article/new-york-nursing-home-coronavirus-deaths-a6c214f4467976efdfca9ba75f8adaef
> how many patients had died in long term facilities.
Not offering excuses – it look like horrendous malfeascance – but it looks like the undercounting that took place was LTC residents who died in hospitals.
From USATODAY
“A probe by the New York Attorney General’s Office found COVID-19 deaths of nursing home residents in the state may have been undercounted by as much as 50% as poor infection-control practices and understaffing fueled the coronavirus crisis inside the long-term care facilities.
The bombshell investigation reported the state Department of Health’s controversial policy to only publicly report COVID-19 deaths of residents inside nursing homes and withhold deaths of residents transferred to hospitals hindered attempts to improve conditions inside the facilities.
The true COVID-19 death toll of New York nursing home residents is closer to 13,000, as opposed to the 8,677 reported to date by the state Department of Health, according to the investigation’s findings. Nursing homes might have also undercounted their deaths to the state, the report said.”
kid –
This is what I was talking about:
-snip-
“Preliminary data obtained by O.A.G. suggests that many nursing home residents died from Covid-19 in hospitals after being transferred from their nursing homes, which is not reflected in D.O.H.’s published total nursing home death data,” a summary of the report reads.
-snip-
I was responding to the following sentence (bold added):
This confirms much speculation about how many patients had died in long term facilities.
It doesn’t make the under-reporting any less incriminating.
https://www.nbcnews.com/news/us-news/covid-death-toll-ny-state-nursing-homes-50-percent-higher-n1255997
That’s different than the criticism that was being levied earlier against Cuomo – that he controlled state policy that forced nursing homes to accept previously hospitalized (and possibly infected) patients.
For all I know both might have happened.
Maybe Josh will now admit that Cuomo did a terrible job with the second highest population fatality rate in the country and that De Santos did a much better job with a population fatality rate little more than half that of New York.
The EU’s ire about AstraZeneca vaccine deliveries is deeply irrational, as shown by a cool look at the numbers involved.
How many vaccine doses would it be practically and politically possible to ship from the UK across the channel to the EU? Maybe ten million at most. This corresponds to less than 3% of the EU’s population. And that would be at the cost of damaging political acrimony, as is already happening.
Much better to just get production sorted out in Belgium and the Netherlands. And to learn the lesson that speed and flexibility do really matter sometimes more than rigid organisational orthodoxy.
It is instructive to summarize what happened here because the infestation by anonymous and nasty activists has been more widespread than usual.
Nic provides another fantastic mathematical analysis of the dynamics of the Covid19 epidemic.
Activists with no qualifications think this analysis might challenge their political narratives.
They employ a number of tactics that are unethical and disrespectful.
1. They try to discredit Nic by various claims that he was “wrong” in the past in various ways despite the fact that no one has been able to predict the course of this epidemic with any accuracy at all. Nic has made a huge contribution by helping people learn what the underlying issues are. This is often personal and nasty. They make various silly and false statements about herd immunity and provide no evidence.
2. They misstate and cherry pick statements and refuse to correct their bad faith misrepresentations. This happened here many times with regard to my statements and links on herd immunity, why the flu disappeared and the usefulness of the concept of R0.
3. They smear one scientist, John Ioannidis (who is a true pioneer in the medical sciences) by misrepresenting his body of work on covid19. They cherry pick what is normal scientific discussion and debate and highlight single sentences.
4. They highlight weak papers by authors with no medical credentials and blog posts by people who admit that in fact they are not experts and that Ioannidis’ work might well be correct.
5. They post Twitter droppings that have little basis or that show alleged screen shots of single sentences on a web site they dislike. This sentence is not on the web site and they have no evidence it was ever there.
6. And finally, they provide endless comments that are mostly content and evidence free. These comments employ various dishonest rhetorical devices such as erecting straw men and lying about what someone else actually said.
7. They show disrespect by not reading or acknowledging in any way references to the scientific literature that undermine their statements. You see that might distract from their cherry picked Twitter threads. They dishonestly claim the references don’t say what they plainly say and don’t respond when the references are quoted directly.
And then the most dishonest part is they demand “civility” after being most uncivil and disrespectful themselves.
This is shameful but part of a pattern in the climate world that started with Skeptical Science and their smears of good scientists such as Judith Curry and both Pielkes. It’s become a way that anonymous scientific incompetents can feel like they are “doing” something to support their political narratives or to combat “disinformation”. What they really mean by disinformation is views that disagree with theirs. They are encouraged by activist scientists like Mike Mann and Andrew Dessler who also lies about people he doesn’t like and falsely calls them deniers. Nic has also done good work in proving wrong some of Dessler’s more outlandish science about the pattern effect for example.
I can only suggest, again, that you keep your postings civil if you wish to have any dialogue.
David –
> Everything I said is true.
There’s much in that comment that is clearly inaccurate – but probably the most starkly and easily demonstrated as inaccurate is the following:
> This sentence is not on the web site and they have no evidence it was ever there.
You have been given a URL for the archived statement. It’s really just wild that you’re persisting with this obviously inaccurate assertion. Just click on the URL.
That you continue to make this argument, despite behaving been directed to that link multiple times, obviates any need (or temptation) to debunk the rest of your inaccuracies – because it so clearly outlines that you are content to persist with obvosuly inaccurate statements.
Joshua, I have no idea if the link is accurate and neither do you. On the web there are all kinds of lies and distortions and phony narratives. I checked just after reading the comment and their web site did not say what is alleged.
Maybe it did say that for 5 seconds or 5 hours. This is extreme quote mining. You take a web site with millions of statements and find a single one that is wrong for which the evidence that it was ever there is weak.
In addition there are many out there who dislike this web site because they disagree with the auspicious scientists who run it. Some of them will be willing to manufacture fake controversies.
As other commenters have said above, you need an attitude check.
You are missing a big piece of the puzzle. Herd immunity declines over time.
A person could have a job that causes him to interact with a lot of people. He gets sick and passes the disease on as a super spreader. He later gets removed from his job, and someone else takes his place. The new person becomes the super spreader.
Immunity of highly connected people will not significantly lower the spread when they are no longer highly connected.
Destroying the economy to “slow the spread” increases the rate at which the super spreaders are swapped out.
Interesting observation. Changes in connectivity of people over time can play a large role.
Strange and weird that there are now at least a dozen new strains of this virus around the world but I have yet to see one named as the “US-strain”. The US has around 25% of all identified cases so why are there no (NY,CA,TX,NV,FL)-strains? After over 25mil infections in the USA so far this seems highly improbable. I suspect our early failures to identify and record emergent strains may have been political in nature (Trump and his admin.).
Why would call any new strain from the US “the US -Strain when you cant call the original strain the “Wuhan virus”
Another potential reason for why the “let ‘er rip” approach is potentially a particularly virulent form of political advocacy. Add this problem to the probability of increase in likelihood of variants with the increase in infections, and the as yet unquantified problem of “long covid.”
-snip-
Almost a third of recovered Covid patients will end up back in hospital within five months and one in eight will die, alarming new figures have shown.
Research by Leicester University and the Office for National Statistics (ONS) found there is a devastating long-term toll on survivors of severe coronavirus, with many people developing heart problems, diabetes and chronic liver and kidney conditions.
Out of 47,780 people who were discharged from hospital in the first wave, 29.4 per cent were readmitted to hospital within 140 days, and 12.3 per cent of the total died.
The current cut-off point for recording Covid deaths is 28 days after a positive test, so it may mean thousands more people should be included in the coronavirus death statistics.
-snip-
https://www.yahoo.com/news/almost-third-recovered-covid-patients-180255388.html
Obvious strawman fallacy from Josh on “let er rip” whatever that means.
The “study” is not linked in the press article perhaps because its preliminary. In any case, Josh’s “confounders” arguments apply. Covid19 seems to have affected mostly older people and people with serious pre-existing health issues. It’s not surprising that these people would be likely to have serious health issues.
It’s a press report and low quality evidence for anything.
What I’ve seen indicates that 2% of covid survivors report lingering symptoms 3 months after infection. Reading more widely would dramatically increase the very low information content of your comments. It’s just as easy to use google scholar as vanilla google and the information quality is vastly higher.
dpy,
I continue to commend your own advice to you, and specifically here suggest you reference your figures rather than asserting them. Reading more widely would dramatically increase the very low information content of your comments.
In this case, you’re out by a factor of five or so.
dpy unreferenced assertion:
2% of covid survivors report lingering symptoms 3 months after infection
UK Government’s Office for National Statistics, taken from the Lancet:
around one in five people who tested positive for COVID-19 had symptoms that lasted for 5 weeks or longer, and one in ten people had symptoms that lasted for 12 weeks or longer
https://www.thelancet.com/journals/lanres/article/PIIS2213-2600(21)00031-X/fulltext
Well here’s what I was referring to.
https://covid.joinzoe.com/post/long-covid
There are more links at the end in case you want to gain a more nuanced level of knowledge.
Once again, its possible that numbers will be all over the place depending on population characteristics. I also suspect that “symptoms” may be rather ill defined. It’s simple minded to think in terms of a single number.
From VTG’s link:
“As COVID-19 (and post-COVID-19 syndrome) are still such new conditions, the guideline is adaptive and will be updated as new evidence becomes available from scientific and clinical studies.”
The real truth is that this is evolving rapidly as more information becomes available.
> whatever that means.
I was referring to the idea that the best way through this is to have a lot of younger people get infected quickly. In some fantasy world that might make sense to me. But the main problem with that idea is the implausibility of protecting more vulnerable people when a lot of younger people are getting infected. The article I excepted points to a potential problem that isn’t typically considered. – which is that infections that result in outcomes short of death may wind up having a lot of other problematic complications.
Yes, those problems would be concentrated primarily among older people.
Perhaps you don’t think that’s worthy of much consideration. That’s certainly your prerogative. I’m not sure I understand why it triggers you if I note that it’s a concern for me.
Excellent, dpy, we’re making progress.
I suggest you be sure to continue providing references rather than assertions in future. It will very much improve the information content of your comments.
Next you could offer a perspective on whether the (excellent and useful btw) data from self selecting group or the “nationally-representative sample of the UK community population” from the office for national statistics is more reliable?
Finally, your advice to engage in “Reading more widely ” can help you here. What is the overall weight of evidence on the prevalence of “long covid”? Is your 2% figure from a self selecting survey credible given the overall evidence?
VTG, You are quibbling. Your own reference says the state of knowledge is changing.
Whether the best estimate is 2% or 10%, it gives the lie to Josh’s alarmist press report that 12% will be readmitted and die.
https://www.ons.gov.uk/news/statementsandletters/theprevalenceoflongcovidsymptomsandcovid19complications
Here’s another acknowledgement that this is evolving.
“This is our first attempt at producing these estimates, and the analysis is very much a work in progress. We will seek to further refine the estimates, for example by using more sophisticated statistical techniques to account for the possibility of relapse and, should sample sizes allow, investigate symptoms persisting beyond 12 weeks.
Early next year, a new long COVID question will be added to the COVID-19 Infection Survey, allowing respondents to state the impact long COVID has had on their day-to-day activities, and including an expanded list of symptoms. This new data will allow us to enrich our analysis, for example by estimating the proportion of people with long COVID symptoms who are burdened by the condition.”
https://twitter.com/kamleshkhunti/status/1354708523867901958?s=19
Interesting article that discusses the important of investigating the interaction between demographic and SES factors, and prevalence/severity of COVID infections.
https://blogs.bmj.com/bmj/2021/01/29/the-missing-link-in-ethnicity-and-covid-19-research-time-to-separate-the-risk-of-infection-from-the-risk-of-severe-disease/?utm_campaign
I’ll just add one more thing here to counter the misinformation provided by Josh, VTG, and Sanakan.
https://www.covidfaq.co/John-Ioannidis-Stanford-School-of-Medicine-14d05ff6d08943fc8b548383e865a9e4
It is cherry picked items to make Ioannidis’ body of work seem worse than it is but at least the items don’t misrepresent his overall views too much. They do misrepresent Gellman post as key items are omitted. Our anonymous flacks are much worse cherry picking single sentences or phrases and omitting important context. Really misleading.
“Ioannidis told viewers that the virus has an “infection fatality rate that is in the same ballpark as seasonal influenza.””
https://buzzfeednews.com/article/stephaniemlee/stanford-coronavirus-neeleman-ioannidis-whistleblower
Ioannidis says it from 2:27 here:
https://youtu.be/jGUgrEfSgaU?t=147
He was completely wrong, as was clear even back in February and March of 2020:
WHO, in February 2020 (Situation Report # 30):
“[…] current IFR estimates […] range from 0.3% to 1%.”
https://reliefweb.int/report/china/coronavirus-disease-2019-covid-19-situation-report-30-19-february-2020
[ https://www.who.int/emergencies/diseases/novel-coronavirus-2019/situation-reports ]
WHO, in March 2020:
“Mortality for COVID-19 appears higher than for influenza, especially seasonal influenza. While the true mortality of COVID-19 will take some time to fully understand, the data we have so far indicate that the crude mortality ratio (the number of reported deaths divided by the reported cases) is between 3-4%, the infection mortality rate (the number of reported deaths divided by the number of infections) will be lower. For seasonal influenza, mortality is usually well below 0.1%. However, mortality is to a large extent determined by access to and quality of health care.”
https://www.who.int/emergencies/diseases/novel-coronavirus-2019/question-and-answers-hub/q-a-detail/q-a-similarities-and-differences-covid-19-and-influenza
WHO, in October 2020:
“Several of these analyses have used published or pre-print seroepidemiologic results and they all converge around a point estimate of around 0.6%.
That may not sound like a lot but that is a lot higher than influenza […].”
https://www.who.int/publications/m/item/covid-19-virtual-press-conference-transcript—12-october-2020
In ‘electronic supplementary material’:
“Using the midpoint of that interval, we estimate that the total U.S. incidence of seasonal influenza during winter 2018-19 was in the range of 45 million to 93 million infections and hence that the population IFR for seasonal influenza was in the range of 0.04% to 0.08%–an order of magnitude smaller than the population IFR for COVID-19.”
https://link.springer.com/article/10.1007%2Fs10654-020-00698-1
“The current data in Europe are consistent with an IFR of 0.5–1.0%, which is many times higher than seasonal influenza (<0.1%)."
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7289569/
“The calculated COVID-19 infection fatality rate is 1.63%, which is 10 to 40 times more deadly than the seasonal flu (fatality rate 0.04%-0.16%).”
https://news.ochsner.org/news-releases/ochsner-health-announces-covid-19-jefferson-and-orleans-parish-prevalence-study-findings
[for: https://wwwnc.cdc.gov/eid/article/26/11/20-3029_article ]
“Our estimates of the IFR of COVID-19 are consistent with early estimates and remain substantially higher than IFR estimates for seasonal influenza (<0.1% in the USA) […].”
https://www.imperial.ac.uk/mrc-global-infectious-disease-analysis/covid-19/report-34-ifr/
“Scientific consensus on the COVID-19 pandemic: We need to act now
[…]
The infection fatality rate of COVID-19 is several-fold higher than that of seasonal influenza […]”
https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)32153-X/fulltext
“[…] the symptomatic infection fatality rate from COVID-19 was estimated to be 1.3 percent. This estimated rate is substantially higher than the approximate IFR-S of seasonal influenza, which is about 0.1 percent […]”
https://www.healthaffairs.org/doi/full/10.1377/hlthaff.2020.00455
(under-estimated SARS-CoV-2 IFR by excluding nursing home deaths, and focuses on the 2009 pandemic influenza):
“As a comparison with another pandemic respiratory disease, the IFR estimates for SARS-CoV-2 are approximately 100 times as high as those for influenza A(H1N1pdm09) (1 to 10 deaths per 100,000 infections)”
https://www.sciencedirect.com/science/article/pii/S2666991920000329#bib0090
(focuses on the 2009 pandemic influenza):
“Although this value remains lower than estimates for other coronaviruses, including SARS24 and Middle East respiratory syndrome (MERS),25 it is substantially higher than estimates from the 2009 H1N1 influenza pandemic.”
https://www.thelancet.com/journals/laninf/article/PIIS1473-3099(20)30243-7/fulltext
(under-estimated SARS-CoV-2 IFR by excluding nursing home deaths):
“Our overall infection fatality risk estimates (0.83-1.07%) were about 10 times larger than those for seasonal influenza.”
https://www.bmj.com/content/371/bmj.m4509
(conflates flu’s CFR of ~0.1% with its IFR of <0.1%):
“Based on the estimated total number of infections, the estimated infection-fatality rate was 0.58%, or approximately six times the 0.1% mortality rate for influenza […]”
https://www.cdc.gov/mmwr/volumes/69/wr/mm6929e1.htm
(conflates flu’s CFR of ~0.1% with its IFR of <0.1%, and calculates IFR poorly):
“This estimate is a factor of 3 − 5 larger than the IFR of a typical seasonal influenza, if one assumes its often referred value of ∼ 0.1 %.”
https://www.medrxiv.org/content/10.1101/2020.11.19.20235036v1.full
translation:
“According to our data, the mortality from COVID-19 infections is just under 1% of the individuals with antibody production, many times that for seasonal flu infections.“
http://www.klinikum.uni-muenchen.de/Abteilung-fuer-Infektions-und-Tropenmedizin/download/de/KoCo191/Zusammenfassung_KoCo19_Epi_dt_041120.pdf
That’s really very odd indeed dpy.
The strongest things I’ve said about Ioannidis is that his work is an outlier and has met a lot of criticism from the scientific community.
To “counter misinformation” you link a website which… …points out that his work is an outlier and has met a lot of criticism from the scientific community.
Your link supports what I’ve written on the subject.
In addition to misrepresenting what other say, you seem to have trouble remembering what you said.
VTG says of Ioannidis’s covid work.
“(1) it has been rapidly disseminated and publicised in the right wing media, pushed by people including yourself.”
This is irrelevant to the science and is unimportant.
“(2) It has consistently been an outlier in the rest of the science”
This is an obvious falsehood. Just to take the most controversial of ioannidis’ papers on the Santa Clara Study. There are many serologic studies that arrive at similar estimates for the IFR, including Los Angeles county and Miami-Dade county. There are others that arrive at higher estimates. Ioannidis has a meta-analysis on this with values ranging from 0.02% to 0.66%.
“(3) It has attracted very strong criticism from the rest of the field”
Once again this is irrelevant and typical of normal science in a controversial field.
dpy,
Curioser and curioser. Through the looking glass we go.
First you lionise (literally!) Ioannidis work and complain about “misinformation” from myself and others commenting on the problems with it.
Then you provide a link to a FAQ providing a list of the very same problems with Ioannidis’ work on Covid, which line ul with my own comments on his work.
This FAQ was written, quoting from the very first line of the website, because “The Covid-19 pandemic has brought with it an avalanche of misinformation.” and goes on to characterise Ioannidis and others as being those who “misunderstood the evidence, have been slow to update their beliefs in the face of new evidence, or simply haven’t updated their beliefs at all.”
I mean, that’s pretty damning.
This is a truly remarkable interaction!
Please, continue. I am agog as to what can possibly come next!
The link I provided is very biased and provides little context. It’s transparently political and a hatchet job. What I said is that its more honest in not selecting single sentences and it at least relies on Ioannidis’ ownstatements. I just don’t take seriously what you, Joshy, and Sanakan say as its obviously non expert and politically motivated and has no context or subtlety.
Here’s a representative comment from the preprint of the Santa Clara study from Mazyar Javid who is an actual scientist and presents a more adult and accurate perspective.
“I left a comment for the first version expressing my astonishment on how
many seem to be obsessed with tearing this study apart and discrediting its
findings altogether. I agree that the study has limitations (as a scientist and
a peer reviewer, I am yet to see a “perfect” study). Nonetheless, the authors
have made substantial attempts to address the limitations reasonably and adjust their results accordingly.
Since the publication of the original report, we have seen results of multiple
serologic studies that have largely corroborated these findings: Studies in
less affected areas (e.g. Czech Republic) which indicated very low prevalence.”
This comment has 25 likes and 3 dislikes. Sanakan’s long winded comment has only 1 dislike and no likes.
dpy,
Thanks!
So, now the link you posted yourself was posted by you not to support your argument at all. Rather, it was a “hatchet job”, which you posted because… [exercise left to the reader to complete]
And now, marvellously, after a whole thread of posts demanding scientific rigour (of others, naturally), your evidence for the quality of Ioannidis’ covid work is… …the number of up posts on a comment(!)
Wonderful!
Next?
Re: “This comment has 25 likes and 3 dislikes. Sanakan’s long winded comment has only 1 dislike and no likes.”
I wonder why a comment made months after other comments has way less responses. Oh wait, the reason is obvious. And the 1 dislike is clearly from, dpy.
Re: “Studies in less affected areas (e.g. Czech Republic) which indicated very low prevalence.””
That’s hilarious since the Czech Republic study implies an IFR ~4 times larger than what the Santa Clara study originally claimed. Maybe Mazyar Javid should go do their homework?
Santa Clara study:
“A hundred deaths out of 48,000-81,000 infections corresponds to an infection fatality rate of 0.12-0.2%”
https://www.medrxiv.org/content/10.1101/2020.04.14.20062463v1.full.pdf
~0.7% IFR for the Czech Republic:
figure 5: https://www.sciencedirect.com/science/article/pii/S1201971220321809
https://twitter.com/GidMK/status/1285020775892709377
Re: “Here’s a representative comment from the preprint of the Santa Clara study from Mazyar Javid who is an actual scientist and presents a more adult and accurate perspective.”
Clearly has less expertise in this topic than I do. Seroprevalence is under the domain of immunology, epidemiology, and public health.
Anyway, the Santa Clara study is trash (if one treats it as an estimate of population-wide seroprevalence and population-wide IFR in Santa Clara County), and virtually everyone working in this field knows it. That’s why it’s yet to be published in peer-reviewed journal, despite numerous other seroprevalence studies making it into peer-reviewed journals (including studies started months after the 2nd draft of the Santa Clara pre-print was posted). I wouldn’t be surprised if it never enters a peer-reviewed journal, or if it gets into a lower-tier journal, or journal where one of the co-authors serves on the advisory board, etc.
“For now, we do not think that the data support the claim that the number of infections in Santa Clara County was between 50 and 85 times the count of cases reported at the time, or the implied interval for the infection fatality rate of 0.12–0.2%. These numbers are consistent with the data, but the data are also consistent with a nearly zero infection rate in the county. The data of Bendavid et al. (2020a, b) do not provide strong evidence about the number of people infected or the infection fatality ratio; the number of positive tests in the data is just too small, given uncertainty in the specificity of the test”
https://rss.onlinelibrary.wiley.com/doi/epdf/10.1111/rssc.12435
https://twitter.com/AtomsksSanakan/status/1332738821998456841
And just for grins… if we use Arizona’s total case count by testing, multiply it by 5X – the estimated highest multiplier from the random sampling serological test done by CDC and local universities last year – and divide that into the actual death count, we get an IFR of .6% If we use their lower bound, we get an IFR of .9%
Cases: 765083
Deaths:13362
Multiplier:5 (from CDC study)
Note that the CDC study has not yet been released in any form other than press release at this time.
The point is that scientific opinion of the Santa Clara study is mostly positive while acknowledging that its an early study and has limitations. VTG, Sanakan, and Josh are of course outside this scientific range of opinion and have no qualifications to be taken seriously.
If the multiplier is 5, then the population infection rate is around 53%. And it will undoubtedly go at least a hit higher (infection rate dropping fast but still at 4800 a day so the % would likely grow a bit before vaccinations enable the establishment of a real, not model-based theoretical, herd immunity).
Even if it’s 4, the population infection rate is 42%.
Yet another example that shows Nic’s toy modeling had some fundamental problems if applied to the real world.
I love the irony that so many “skeptics” lined up behind flawed theoretical modeling when it fit their political agenda.
Just to correct the record here as the level of disinformation is high in this comments for this post.
The Santa Clara preprint attracted hundreds of comments. This a typical one from a scientist and agrees with my opinion.
“I left a comment for the first version expressing my astonishment on how
many seem to be obsessed with tearing this study apart and discrediting its
findings altogether. I agree that the study has limitations (as a scientist and
a peer reviewer, I am yet to see a “perfect” study). Nonetheless, the authors
have made substantial attempts to address the limitations reasonably and adjust their results accordingly.
Since the publication of the original report, we have seen results of multiple
serologic studies that have largely corroborated these findings: Studies in
less affected areas (e.g. Czech Republic) which indicated very low prevalence.”
This comment had 25 likes and 3 dislikes. Sanakan’s long winded comment has 1 dislikes and 0 likes. I guess his obsession only sells outside of actual scientists.
There’s simply no getting around the fact that the peer-reviewed literature is full of researchers pointing out why the non-peer-reviewed Santa Clara pre-print is garbage. You can’t change that by complaining that comments that come months later have less likes than comments that come months earlier (as if likes in Medrxiv’s comments section is a sound metric of scientific rigor).
“H.3 Studies excluded due to non-representative samples
H.3.1 Active recruitment of participants
[…]
Santa Clara, California, USA[127]
Participants were recruited via social media and needed to drive to the testing site. Stanford Medicine subsequently released a statement indicating that the study was under review due to concerns about potential biases”
https://static-content.springer.com/esm/art%3A10.1007%2Fs10654-020-00698-1/MediaObjects/10654_2020_698_MOESM1_ESM.pdf
Non-targeted, active recruitment of participants (as opposed to randomized sampling) is a bad idea and leads to non-representative samples. That’s just one of the many reasons why the Santa Clara study is trash.
Their own subsequent seroprevalence study of MLB employees came in higher than they expected based on the Santa Clara study. And their rationalization for why it did was funny as all get out.
Ioannidis extrapolated from the Santa Clara study to make projections that were ridiculous. Sad to see a 🦁go down this path.
Er…. lower that is.
-snip-
At-home Covid-19 test to ramp up production with $231.8 million federal contract
(CNN)The Biden administration announced on Monday that the US Department of Defense and US Department of Health and Human Services are working with Australian company Ellume to provide more of its fully at-home Covid-19 tests to the United States.
https://www.cnn.com/2021/02/01/health/at-home-covid-19-test-ellume-contract-bn/index.html
Joshua
Did you see the article today by Nick?
https://lockdownsceptics.org/a-rebuttal-of-claims-by-christopher-snowdon-about-false-positive-covid-19-test-results/
Tonyb
Tony –
Hospitalizations, ICU admissions and deaths all track with positive PCR tests.
In South Wales Australia they’ve given over a hundred thousand PCR tests per week with vanishing few positive tests and those few that came back positive were confirmed to be true positives.
Of course, where the base rate is higher the rate of false positives will be higher due to contamination – but this rhetoric about “casedemic” due to PCR false positives is specious.
How will such tests improve anything practically speaking?
Rob –
Rhetorical question?
Just trying to understand how the effort (in your opinion)will be effective in helping the US economy.
It would depend on the scale.
For me individually, access to such tests would make a huge difference. I could test people who want to visit us (including my 90 year-old mother in law with multiple comorbidities). I could test myself before visiting peope outside my pod (assuming they’ve tested also).
On a larger scale, it could make a big difference to individual employers who could test employees or even individual consumer-fscimf businesses that could test customers, or schools who could test students and teachers and staff.
In a larger scale still it could significantly accelerate the return to a “normal” economy.
This is interesting – not exactly a shock if true:
-snip-
Young and middle-aged adults responsible for most COVID spread
The coronavirus pandemic in the U.S. has been chiefly driven by young and middle-aged people, while killing mostly older people.
Driving the news: Adults aged 20-49 were responsible for the vast majority of virus transmission last year, even after schools reopened in the fall, according to a new study published in Science.
The notion that non-vulnerable people can go about their normal lives, while vulnerable people self-isolate, has not borne out in the U.S.
-snip-
https://www.axios.com/young-adults-coronavirus-spread-a998cf65-fcf3-472b-a1ed-e29938356351.html
How Anti-Trumpers and the media killed hundreds of thousands people in the US!
ndian Council of Medical Research (ICMR) has maintained its recommendation and approved the use of HCQ as prophylaxis based on the studies conducted in India, despite World Health Organisation (WHO) suspending the clinical trials using hydroxychloroquine (HCQ) under its Solidarity Trial.
A recent case-controlled study by ICMR has underlined the benefit of hydroxychloroquine (HCQ) as prophylaxis, showing that the sustained use of the anti-malaria drug along with the use of personal protective equipment (PPE) was associated with a significant decline in risk of Covid-19 infection rate by upto 80% among the healthcare workers.
https://health.economictimes.indiatimes.com/news/pharma/why-icmr-continues-to-stand-firm-on-using-hydroxychloroquine-as-prophylaxis/76172274
Cases in India have plummeted:
https://www.npr.org/sections/goatsandsoda/2021/02/01/962821038/the-mystery-of-indias-plummeting-covid-19-cases
More on countries with frequent use of hydroxychloroquine.
Is there any possible correlation between countries where the drug is routinely administered due to high malaria rates and lower cases of COVID-19? It’s too early to tell, medical professionals say, but it’s part of the investigation process.
“If you look [at the] countries where malaria is more prevalent and countries where COVID-19 infections are prevalent, you will find a striking difference. This correlation needs to be explored further as this is not just a mere coincidence,” Dr. David Nazarian, a Beverly Hills-based physician, diplomate at the American Board of Internal Medicine and founder of My Concierge MD, told Fox News.
https://www.foxnews.com/world/do-countries-with-high-rates-of-malaria-have-fewer-coronavirus-deaths
Models and tests don’t matter so much if you have a cure:
https://judithcurry.com/2021/01/10/covid-19-why-did-a-second-wave-occur-even-in-regions-hit-hard-by-the-first-wave/#comment-942461
We have not only one cure, hydroxychloroquine, but TWO! Ivermectin!
And so, it is with great pride as well as significant optimism, that I am here to report that our group, led by Professor Paul E. Marik, has developed a highly effective protocol for preventing and early treatment of COVID-19. In the last 3-4 months, emerging publications provide conclusive data on the profound efficacy of the anti-parasite, anti-viral drug, antiinflammatory agent called ivermectin in all stages of the disease. Our protocol was created only recently, after we identified these data. Nearly all studies are demonstrating the therapeutic potency and safety of ivermectin in preventing transmission and progression of illness in nearly all who take the drug.
https://www.theautochannel.com/news/2020/12/10/922677-senate-testimony-dr-pierre-kory-ivermectin-miracle-drug-to-treat.html
The Hydroxychloroquine Scandal
Posted on 10:56 am, February 2, 2021 by Rafe Champion
This shameful distortion of medical science was emblematic of
how politics not only crept into the treatment of the CCP virus, it
bludgeoned that treatment and resulted in untold thousands,
perhaps millions, of deaths, while simultaneously making life
unbearable for an even greater number across the globe—in fact,
for practically everyone.
https://catallaxyfiles.com/2021/02/02/the-hydroxychloroquine-scandal/
No antiviral has been shown to reduce mortality in SARS-COV-2 patients to date. In the present observational and retrospective report, 3,099 patients with a definitive or highly probable diagnosis of infection due to COVID-19 were evaluated between May 1st to August 10th, 2020, at Centro Medico Bournigal (CMBO) and Centro Medico Punta Cana (CMPC), and all received compassionate treatment with Ivermectin. A total of 2,706 (87.3%) were discharged for outpatient treatment, all with mild severity of the infection. In 2,688 (99.33%) with outpatient treatment, the disease did not progress to warrant further hospitalization and there were no deaths. In 16 (0.59%) with outpatient treatment, it was necessary their subsequent hospitalization to a room without any death. In 2 (0.08%) with outpatient treatment, it was necessary their admission to the Intensive Care Unit (ICU) and 1 (0.04%) patient died.
https://www.medrxiv.org/content/10.1101/2020.10.29.20222505v1.full
As best I could tell, that article didn’t present either age or comorbidity statistics, other than an average age.
That makes it not at that interesting, unless I am missing something.
There is some age information here:
https://www.medrxiv.org/content/10.1101/2020.10.29.20222505v1.full.pdf+html
The only demographic info I see is:
“The distribution by gender of all patients admitted to the ER was as follows: 50.02% male and 49.98% female. While the distribution in hospitalized patients was: 63.5% male and 36.5% female. (Table 5)
The average age of the patients hospitalized in rooms of the Covid-19 isolation area was 52 years; while in ICU patients it was 58 years.”
It’s over.
In his post above, Nic Lewis approvingly cites Tkachenko et al.’s pre-print:
“The effect of such heterogeneity on transmission of infection and on the HIT can be represented by a single parameter λ, the heterogeneity factor (Tkachenko et al. 2020)[3]”
Lewis has done this before. For example:
https://archive.is/Qunlx#selection-419.0-429.4
https://archive.is/gPC2m#selection-745.0-753.43
Tkachenko et al. recently updated their pre-print to version 4, where they admit they were wrong about HIT. They admit they were instead talking about a transient equilibrium resulting from behavior changes, etc., not herd immunity achieved after reaching HIT:
“We derive a non-linear dependence of the effective reproduction number Re on the susceptible population fraction S. We show that a state of transient collective immunity (TCI) emerges well below the HIT during early, high-paced stages of the epidemic. However, this is a fragile state that wanes over time due to changing levels of social activity, and so the infection peak is not an indication of herd immunity: subsequent waves can and will emerge due to behavioral changes in the population, driven (e.g.) by seasonal factors. Transient and long-term levels of heterogeneity are estimated by using empirical data from the COVID-19 epidemic as well as from real-life face-to-face contact networks. These results suggest that the hardest-hit areas, such as NYC, have achieved TCI following the first wave of the epidemic, but likely remain below the long-term HIT. Thus, in contrast to some previous claims, these reqions can still experience subsequent waves.”
https://www.medrxiv.org/content/10.1101/2020.07.26.20162420v4
I told Lewis this in July, about a month before Tkachenko et al. posted the first version of their article. And I cited sources explaining this ( https://archive.is/basiq#selection-11877.0-11923.159 ):
“Neither Stockholm, nor Sweden in general, hit the herd immunity threshold. You’re conflating the results of interventions, behavioral changes, etc., with herd immunity.
[…]
“These early onset peak rates should arise not because of herd immunity but because of changes in behavior. […]
The peaks occur at levels of infection far from that associated with herd immunity. Post-peak, shoulders and plateaus emerge because of the balance between relaxation of awareness-based distancing (which leads to increases in cases and deaths) and an increase in awareness in response to increases in cases and deaths.”
https://web.archive.org/web/20200701170752/https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7273247/
“This allows us to “bend the curve” and predict temporary equilibrium states, far away from the equilibrium state of herd immunity, but stable under current conditions [26]. Yet, these states can quickly become unstable again once the current regulations change. Our dynamic SEIR model allows us to study precisely these scenarios.”
https://www.medrxiv.org/content/medrxiv/early/2020/05/10/2020.05.01.20088047.full […]”
Lewis also discussed Britton’s work on heterogeneity lowering HIT:
https://archive.is/Qunlx#selection-37367.0-37387.500
Britton later admitted he was wrong on Sweden reaching a low HIT:
“Britton told Foreign Policy that his calculations that Sweden would reach herd immunity turned out to be incorrect. Britton now says that U.S. government officials misinterpreted his study and that using his June research to promote herd immunity was wrong, adding that “too many people will die in order to reach herd immunity.””
https://archive.is/2NJIk#selection-1607.0-1623.302
And others explained the role of interventions, behavior changes, etc. in mitigating Sweden’s outbreak, as opposed to herd immunity. That again contradicts Lewis’ claims on Sweden achieving herd immunity:
“Initially, some local authorities and journalists described this as the herd immunity strategy: Sweden would do its best to protect the most vulnerable, but otherwise aim to see sufficient numbers of citizens become infected with the goal of achieving true infection-based herd immunity. By late March 2020, Sweden abandoned this strategy in favor of active interventions; most universities and high schools were closed to students, travel restrictions were put in place, work from home was encouraged, and bans on groups of more than 50 individuals were enacted. Far from achieving herd immunity, the seroprevalence in Stockholm, Sweden, was reported to be less than 8% in April 2020,7 which is comparable to several other cities (ie, Geneva, Switzerland,8 and Barcelona, Spain9).”
https://jamanetwork.com/journals/jama/fullarticle/2772167
It’s also clear that Gomes’ low HIT work was wrong, since she didn’t adequately account for the mitigating impact of behavior changes, etc., leading to her team conflating their effects with that of herd immunity. That’s telling, since Lewis has been citing Gomes’ low HIT work since at least May:
“The recent publication of the Great Barrington Declaration (GBD), which calls for relaxing all public health interventions on young, healthy individuals, has brought the question of herd immunity to the forefront of COVID-19 policy discussions, and is partially based on unpublished research that suggests low herd immunity thresholds (HITs) of 10-20%. We re-evaluate these findings and correct a flawed assumption leading to COVID-19 HIT estimates of 60-80%.
[…]
While Aguas et al. matched their mitigation shapes to mobility data, these data are not an accurate picture for total transmission mitigation, as decoupling of mobility and transmission has been previously noted as populations adopt precautionary behavior like mask wearing and social distancing “
https://www.medrxiv.org/content/10.1101/2020.12.01.20242289v1.full
I’ve been explaining to Lewis since at least May that HIT is not low:
https://archive.is/Xjyec#selection-15423.0-15747.43
This was not a lucky guess on my part; this is my field.
So again: it’s over.
It’s now just a matter of all the “HIT is low!!” people honestly admitting they were wrong (like Britton and Tkachenko et al. did), and that competent mainstream experts were right. If they don’t admit that, then that will say a lot about their mindset.
https://twitter.com/AtomsksSanakan/status/1332738821998456841
“Given your long track record of distortions and misrepresentations over the last 9 months, I don’t take anything you say seriously.”
You will be pleased to know that few here take you seriously. I certainly don’t. I tire of your constant wrangling, to the point I may add you to my comment notification email filter that already suppresses comments by RE.
It will save me deleting the emails.
You obviously haven’t bothered to read the above Post, which is what this Comment thread relates to.
Tkachenko haven’t “admitted that they were wrong”. They have revised their manuscript in a way that makes it, I imagine, easier to get published. They present no evidence for the new assertion in their revised version that the spread of the epidemic in the first wave was dominated by heterogeneity – ‘bursty social activity’ – of a transient nature.
As for the HIT not being “low”, as I explain in this post the HIT is a function of both R0 and population heterogeneity, both of which are affected by changes in behaviour (forced and/or unforced), and R0 also varies with the season.
No doubt the fact that the epidemic seems to be dying out in India despite there being relatively few restrictions enforced there and people’s behaviour having at least partially normalised won’t cause you to reconsider your position.
Washington post, 4 February: https://www.washingtonpost.com/world/asia_pacific/india-coronavirus-cases/2021/02/04/d7f92f72-6562-11eb-bab8-707f8769d785_story.html
“The apparent retreat of the coronavirus in India, the world’s second-most populous nation, is a mystery that is crucial to the future course of the pandemic.”
“The results of a nationwide antibody survey of 28,600 people by the government released on Thursday indicated that more than 1 in 5 Indians — about 270 million people — had been exposed to the virus as of early January.”
-snip-
[…] We examined seroprevalence of anti-SARS-CoV-2 antibodies in Pune city in India and its implication for protective immunity.
[…]
Findings Seropositivity was extensive (51·3%; 95%CI 39·9-62·4) but varied widely in the five localities tested, ranging from 35·8% to 66·4%
https://www.medrxiv.org/content/10.1101/2020.11.17.20228155v1
Nic wrote: “As for the HIT not being “low”, as I explain in this post the HIT is a function of both R0 and population heterogeneity, both of which are affected by changes in behaviour (forced and/or unforced), and R0 also varies with the season.”
However, from a practical point of view this definition of herd immunity isn’t very useful. For example, when considering a vaccination campaign to completely suppress a pandemic disease, we want the disease suppressed in all seasons and when all people are behaving normally: not wearing masks, not working from home, schools open, and the elderly not cowering at home. The best value of R0 for that purpose is derived from the early exponent stage of the pandemic before people began to respond. In the US, that was the first three weeks of March before lockdowns began. (Actually, it takes roughly a week for changes in behavior and policy to influence the rate at which new cases are detected and perhaps 3.5 weeks if you track deaths.) So I think many use R0 from that period to calculate a HIT that doesn’t take into account heterogeneity. To some extent, I think there is a communication problem.
IMO, the bigger problem is that during a pandemic, when R0 is changing it is difficult to accurately calculate the heterogeneity parameter, lambda. As best I can tell, one must pick a period when R0 is constant and assume that all of the reduction in cases is due to the rising number of resistant individual. If R0 isn’t constant, then you overestimate the importance of the heterogeneity factor. Or, if you estimate the impact of changes in behavior and public policy or estimate those changes from observations (like Google mobility data) you still run the risk of underestimating these factors and overestimating lambda.
Respectfully, Frank
“However, from a practical point of view this definition of herd immunity isn’t very useful. For example, when considering a vaccination campaign to completely suppress a pandemic disease, we want the disease suppressed in all seasons and …”
Franktoo – thanks for that insight. It’s a very good point – and it’s true whether suppressing it with vaccination, mitigation or immunity by infection. You need to use the high end of the HIT spectrum. The lower HIT values are really only of interest if you want to predict a transient maximum in the case rate, and it’s not clear the value of that (although I could imagine some uses).
“IMO, the bigger problem is that during a pandemic, when R0 is changing it is difficult to accurately calculate the heterogeneity parameter, lambda.”
Exactly. The biggest problem I had with the use of heterogeneity was the lack of information for coming up with the value of the parameter (lambda).
Worse, if you think about how you’d build an agent-based models(which I like and did professionally in a different field once upon a time) – you need multiple parameters which include not only heterogeneity of population susceptibility, but also of travel, of behavior, etc.
That leads me to ask of Nic (because I really don’t know): does it work to use single heterogeneity parameter (lambda) to modify R, or do you need to use a more complex formula?
Another problem is that R0 is also sensitive to heterogeneity in the population for which it is measured. That means that using that R0 and then tossing heterogeneity into a model will overestimate the impact of that heterogeneity if you don’t discount that in the R0 situation – i.e. you will get too low an HIT.
It’s certainly an interesting set of problems. I would enjoy studying it more were in not also messing up the whole world, and killing millions of people.
Repeating [with some editing] what I said above. I would love a correction if I’ve got something wrong::
Imagine a scenario where a boat with 1000 people aboard get stranded on a desert isle.
1 has COVID at the time of the shipwreck and then infects three others and then each of those 3 others infect two others after that. So a total of 10 have been infected.
So then everyone isolates and 2 of those infected died and the rest recovered and developed antibodies and immunity and there’s no more virus on the island.
So, according to [Nic’s] thinking, 0.1% of the population on the island has been infected and the island population has reached a “herd immunity threshold.”
Basically, [he’s] equating “herd immunity threshold” with infection level. You could call any prevalence of infection a “herd immunity threshold” depending on the conditions.
Nic stated that Stockholm had (likely, but with high confidence) reached
a “herd immunity threshold,” [8+ months ago] specifically because he had identified the people in Stockholm as engaging in near normal-level behaviors. He has re-stated such in this thread. If you have reached “herd immunity threshold” in near-normal level behaviors, changes in behavior (presumably because of season) would not result in the subsequent level of growth in infection rate that they’ve seen in Sweden since the date when Nic said they’d reached a “herd immunity threshold.” Well, unless those changes in behaviors included something like going around and deliberately infecting people
For the term “herd immunity threshold” to have some meaning it needs to mean more than “any level of prevalence of infection I want it to mean no matter the context [ as in saying they’d reached a “HIT” on that island]. To be meaningful, it needs to have some implication w/r/t the level of confidence people can feel that they won’t likely get infected under normal (or near-normal) conditions (with the understanding of course, that with overshoot there’s still some level of potential to get infected).
Re: “Tkachenko haven’t “admitted that they were wrong”. They have revised their manuscript in a way that makes it, I imagine, easier to get published.”
Their initial draft said this was herd immunity. They admitted they were wrong by switching their claim from herd immunity, to a transient equilibrium they explicitly say isn’t herd immunity. You’ve given no evidence for your speculation on their motives for doing that. I could just as easily presume they are non-experts who honestly realized they made a basic error in a topic outside their field of expertise. After all, none of them have a background in this field (a look at their profiles on Medrxiv shows that), and as an immunologist, I have more expertise in this field than any of them do.
Re: “As for the HIT not being “low”, as I explain in this post the HIT is a function of both R0 and population heterogeneity, both of which are affected by changes in behaviour (forced and/or unforced), and R0 also varies with the season.”
HIT and R0 don’t depend on additional behavior changes (including forced behavior changes) since R0 is, by definition, under conditions of no additional behavior changes and no additional public health interventions, beyond what would be typical for that time of year. In laymen’s terms for the COVID-19 pandemic, that means what conditions would have been at the same time of year in 2019. So if you have a substantial level of additional behavior changes and/or public health interventions beyond what you had at the same time of year in 2019, then you aren’t at baseline conditions and R0 doesn’t apply. For a novel pathogen, baseline conditions are typically present right near the beginning of the outbreak when the pathogen is first introduced, with virtually no one infected and before there’s enough time for behavior + interventions to change much.
You’ve been cited sources on that for months, and again in the comment you’re responding to. Even Britton admitted this in his work, as have others. For instance:
Ferguson et al., March 2020:
“In the (unlikely) absence of any control measures or spontaneous changes in individual behaviour, we would expect a peak in mortality (daily deaths) to occur after approximately 3 months (Figure 1A). In such scenarios, given an estimated R0 of 2.4, […]”
https://www.imperial.ac.uk/mrc-global-infectious-disease-analysis/covid-19/report-9-impact-of-npis-on-covid-19/
Britton:
“Herd immunity is defined as a level of population immunity at which disease spreading will decline and stop even after all preventive measures have been relaxed. If all preventive measures are relaxed when the immunity level from infection is below the herd immunity level, then a second wave of infection may start once restrictions are lifted.”
https://science.sciencemag.org/content/369/6505/846.full
Re: “No doubt the fact that the epidemic seems to be dying out in India despite there being relatively few restrictions enforced there and people’s behaviour having at least partially normalised won’t cause you to reconsider your position.
Washington post, 4 February: https://www.washingtonpost.com/world/asia_pacific/india-coronavirus-cases/2021/02/04/d7f92f72-6562-11eb-bab8-707f8769d785_story.html “
India had a lockdown and a number of additional public health interventions that pushed them from the baseline, HIT-relevant conditions of R0. A number of those interventions were still in place into late January 2021, when the most recent data is available:
https://ourworldindata.org/policy-responses-covid
https://www.mha.gov.in/notifications/circulars-covid-19
Moreover, a number of regions in India with representative sampling (not convenience samples) have seroprevalence greater than your low HIT model would predict, despite India being under non-baseline, non-R0, non-HIT-relevant conditions that helped mitigate infection rates. For instance:
– ~51%
https://www.medrxiv.org/content/10.1101/2020.11.17.20228155v2.full
– ~54% in urban area, ~47% population-wide
https://jamanetwork.com/journals/jama/fullarticle/2776292
– ~55% in slums, ~15% in non-slums
https://www.thelancet.com/journals/langlo/article/PIIS2214-109X(20)30467-8/fulltext
So no, the India example doesn’t help your case at all. And I don’t get my information on science from press sources like the Washington Post. But if you insist on citing press sources, then here’s one for you:
“Survey finds 56% of Delhi has COVID antibodies, adding to India’s declining cases mystery”
https://fortune.com/2021/02/03/delhi-covid-antibodies-india-declining-cases-herd-immunity/
If that sampling is the same as previous rounds of testing in Delhi, then its representative and not a convenience sample:
– round 1: https://pib.gov.in/PressReleaseIframePage.aspx?PRID=1640137
– rounds 2 – 4: https://www.medrxiv.org/content/10.1101/2020.12.13.20248123v1.full
Feel free to explain how 56% seroprevalence under non-baseline non-R0 conditions squares with your low HIT. I assume you’ll set aside any seroreversion or failure of some infected people to seroconvert, since that would make your case even worse by making the infection rate higher than seroprevalence implies.
Your example is extreme and impossible Josh. Complete isolation is impossible because people need to eat and social interaction is essential to mental health.
Defining herd immunity in terms of Rt or R0 makes sense. The problem is that we never know either accurately. Both can change a lot. And “normal conditions” is a vague and uninformative phrase.
From a practical perspective every spring flu goes away and it makes sense to say herd immunity is achieved. Next winter flu comes back with new strains and that herd immunity is not applicable. Or maybe it’s the same strain and R changes with the season.
The problem here is that viral epidemiology is a primitive science largely based on crude mechanistic narratives to explain what is happening that lack quantification. Perhaps there are quantum mechanics like effects that are inherently unpredictable. It is a shocking example that the political doctrines about “following the science” are just emotional appeals.
The article I cited on flu is really immensely good at summarizing how little we know.
Nic and friends: FWIW, ND and SD remain the US states with the highest cumulative number of cases at 12.9% and 12.3% of the population. In Burleigh County (Bismarck), ND 15%. Almost all occurred this fall and winter when testing was not a major factor in detection. The current rate of new detect cases is 12 and 16 per 100,000 per day or barely over 0.01%/day, about 1/3 the average rate in the country, and it has been this low for more than one month. If these states haven’t reached herd immunity by now, they aren’t going to be approaching it anytime soon when it takes about two months to add 1% to cumulative cases. Since we haven’t done any seropositivity studies lately, the best I can tell is that the pandemic has slowed with at least 25% cumulative infections assuming a minimum of one missed case for every detected case. However, cumulative cases could easily be approach 40% of even more if we are missing more.
However, we now have nine other states with cumulative detected infections totaling 10-11% of the population, including many where the rate of detection of new cases is above average. That average has fallen by factor of 2 since the post-holiday peak in early January, when many locations had more than 100 new detected cases/day per 100,000.
Yuma county, AZ pop 200,000 has 16.5% cumulative case and an average rate of new infections 41/day/100,000, which is down from more than 100. I’m not sure locations with small populations are representative of anything, but Chattahoochee County, Ga with 10,000 people living in one city has cumulative infections amounting to 25.5% of the population and the pandemic is still raging there with 223 new case/day/100,000 (but that is only 24 cases/day, 7-day averaging). With only 5 deaths from 2,782 total cases, the data is getting pretty noisy.
“Yuma county, AZ pop 200,000 has 16.5% cumulative case and an average rate of new infections 41/day/100,000, which is down from more than 100. I’m not sure locations with small populations are representative of anything, ”
Beware of Yuma – your caution is warranted. It sits almost on the Mexican border, and Mexico has a real COVID-19 catastrophe in progress (their deaths are way up, their testing results are irrelevant). This means that people are routinely traveling back and forth across the border in that area, not to mention the illegal immigrants (“undocumented welcome future workers” to Joshua).
You might want to look at the epidemic curves and final values (so far) for Arizona overall. Twice we have had the highest rate of infections in the nation, and twice that has turned sharply around (I’m hoping that the current <.80 Rt continues). I estimate that our percent already infected is approaching 50%.
Total cases per testing: 779,000
Missed cases ratio (per CDC report): 4X-5X
Total infected individuals – estimated (million): 3.2 – 3.9
Population (million) 7.5
Percent of population already infected (million): 43% – 52%
Number vaccinated (at least first dose): 740,000 – 10%
Total vaccinated or already infected: 53%-62%
Vaccination rate per week: 3%
And since Curry is blocking comments citing evidence that Lewis is wrong on a low HIT, I’ll end with this [ https://archive.is/gToqx#selection-11545.0-11806.2 ]:
India’s mobility data is not near the baseline conditions of R0, especially in terms of less traffic for retail and recreation. There’s some increased traffic to supermarkets / pharmacies, presumably in large part due to people going to get vaccinated, getting medications for illness, and needing to buy food for home as they eat out less + spend more time at home. Traffic everywhere else is down, with the exception of increased time at home/residences. That’s consistent with non-baseline, non-R0 levels of distancing as people stay home more than usual.
So no, I’m not surprised that India managed to stop reported cases/day from increasing (i.e. get Re below 1) through a combination of additional public health interventions and additional behavior changes, without achieving the non-vaccine-mediated herd immunity of R0. That’s the equilibrium I and others have been talking about for months, and which Tkachenko et al. finally caught on to. Same thing happened in Sweden, and the other places you incorrectly claimed achieved a low HIT when you incorrectly predicted no strong 2nd waves ( https://archive.is/0GLKr#selection-2869.0-7510.0 ). This time, though, the equilibrium may not be transient, if the behavior changes and/or public health interventions hold up long enough for India to vaccinate enough people to reach vaccine-mediated herd immunity.
https://web.archive.org/web/20210206193031/https://www.gstatic.com/covid19/mobility/2021-01-31_IN_Mobility_Report_en-GB.pdf
I block comments that insult Nic Lewis (or anyone else). If you have a justified criticism, it will pass through. I will also start blocking your posts with repetitive links.
Nic: I’m forced to confess that another of my early criticisms of heterogeneity has been proven wrong. The polyclonal antibodies from people who have been infected with COVID turn out to differ greatly in their ability to bind the spike protein from coronavirus variants. With 20/20 hindsight, those who have asymptomatic or mildly symptomatic and don’t get tested are probably those who amount a more effective antibody response. The genes that are recombined to make antibodies aren’t the same in all people. So, while everyone mounts the same defenses – which need to find a sweet spot between protecting from infections and cancer and avoiding autoimmune disease – at the molecular level these responses are heterogeneous.
In fact, I now suspect that more transmissible variants are those that are capable of better dealing with the polyclonal antibodies the average person makes. Perhaps these variants make 75% of people sick enough to get tested while the average variant causes only 50% get tested. That would result in a 50% increase in detected transmissibility. Symptomatic patients are more infectious. The more transmissible variants are also modestly more deadly.
Josh, Your island example is simplistic and completely unrealistic. Complete isolation of everyone is impossible because everyone needs to eat and needs social interactions.
R0, Rt, and HIT are idealized concepts that are hard to define meaningfully and even if estimated may or may not predict anything about the future.
Epidemiology is not a mature science. It relies on crude mechanistic narratives to explain most things about the flu for example. Often these narratives lack real quantification. You should read the flu article I referenced. You would learn a lot of new and interesting information about how ignorant we are.
Just as an example, why are flu rates not decreasing in Italy over the last decade when vaccination levels have gone from 10% to 60%? I think that may be in older populations but can’t recall exactly.
Franktoo, I do think that as more people get vaccinated we may yet determine what the HIT is at least in the spring. At the point where hospitalizations get really small, we I think can conclude that those vaccinated + those already infected will be the HIT, or at least a reasonable estimate.
I also believe that in the Dakotas that may have already been achieved. Winter has not gotten less severe over the last 3 or 4 weeks while new infections seem to be down dramatically and falling.
It will also I think give more information about bounds on IFR. The population fatality rates in the Dakotas I think are around 0.2%.
David –
> Complete isolation of everyone is impossible because everyone needs to eat and needs social interactions.
Lol.
The point of the example was to illustrate the arbitrariness in how some people are defining “herd immunity threshold,” as you do when you say it can move all over the map within a huge range of population infection rates.
I mean sure, one CAN claim that a HIT occurs under non-near normal conditions (such as in my example), or conditions that otherwise would normally change such that the infection rate would then grow exponentially (as happened in Sweden 8+ months after Nic said they’d reached a “herd immunity threshold.” But what is the value in such a definition? I don’t think there is any. And it certainly isn’t the common usage of the term as it has been used during this pandemic.
I wasn’t suggesting that total isolation is a realistic possibility.
There, I hope that clears it up.
Josh, What you say here is not really true. I cited a Nature article where a prominent scientist defined HIT using Rt. The reference ATTP gave says the term is used in many ways.
To you what constitutes “normal conditions?” I don’t think such a thing exists. Is it with reference to winter, spring, summer, or fall? At the very least it will be different in each country and region. So then R0 is highly variable too. That means HIT even using the naive definition is highly variable.
David –
> To you what constitutes “normal conditions?”
Lol. I’m going with Nic’s definition.
I’ll respond to Sanakan to save Nic the time. He’s got vastly more important things to do than wade through another long string of largely baseless assertions. It’s mostly quibbling about definitions and misrepresenting what Nic and others have said.
What’s odd about this latest one is that it ignores everything that has been said previously about HIT, R0, and Rt and then leverages extremely simplistic thinking to show that Nic was “wrong” when the situation is vastly more complex. But then Sanakan’s self declared intent is to be the Antifa of science by “correcting” wrong science and his track record is quite nasty. Let’s review some material from previous comments here.
1. I did a google search on herd immunity threshold and one of the first hits was a Nature piece in which a scientist was quoted saying that HIT varied immensely between countries with a range of 5% to I think 78% and clearly used Rt to define HIT.
2. ATTP cited an entire essay on the definition of HIT. Basically, there are lots of ways people use the term including a level of immunity that would be expected to lead to the epidemic dying out.
3. The definition of R0 is vague, making the term not very useful. For the flu for example since R is so seasonal, which one would one use? Noone here has answered that question. One could use the R for the season when the epidemic “started” whatever that means. Or one could average over all seasons and all countries. Asking for a single R0 is a little like asking for a single number to define the climate of the globe. It’s absurd on its face.
4. Whether Tkachenko admitted they were “wrong” is irrelevant. I think this “controversy” says more about how simplistic and vague things like R0 and HIT really are.
5. The biggest simplistic error Sanakan makes is to assume that HIT is a universal constant that is either “low” or “not low.” This is pseudo-science. HIT as explained by Kwok in Nature is extremely variable and depends on many things. Nic’s post here does an excellent job showing some of the factors that come into play.
6. I’ve cited a paper on the flu that conclusively shows how little we know about the flu. It was shocking to me. At the end, I’ll reproduce the full abstract.
My conclusion (which I was surprised by) is that even compared to other fields of medicine, viral epidemiology is an immature science dominated by crude mechanistic narratives that lack rigorous quantification and very simplistic models that are ill-posed and based on poorly defined and uninformative constants like HIT and R0. The fact that Sanakan claims (with no evidence) that he is an epidemiologist is all anyone needs to know. This state of affairs also makes Sanakan’s style (similar to St Thomas Aquinas’ of stating a thesis and then using proof texts to prove it) all the more unhelpful. The issues are complex and subtle. Sanakan in trying to prove people who don’t fit his narrative wrong, abuses science and rhetoric. It’s online bullying. And the bullies get to wear white hoods and robes.
By contrast, Nic’s venture into this field has been detailed and pushed forward the frontier by considering vastly more complex models that have a far better chance of being informative. The contrast couldn’t be more stark.
“The epidemiology of influenza swarms with incongruities, incongruities exhaustively detailed by the late British epidemiologist, Edgar Hope-Simpson. He was the first to propose a parsimonious theory explaining why influenza is, as Gregg said, “seemingly unmindful of traditional infectious disease behavioral patterns.” Recent discoveries indicate vitamin D upregulates the endogenous antibiotics of innate immunity and suggest that the incongruities explored by Hope-Simpson may be secondary to the epidemiology of vitamin D deficiency. We identify – and attempt to explain – nine influenza conundrums: (1) Why is influenza both seasonal and ubiquitous and where is the virus between epidemics? (2) Why are the epidemics so explosive? (3) Why do they end so abruptly? (4) What explains the frequent coincidental timing of epidemics in countries of similar latitude? (5) Why is the serial interval obscure? (6) Why is the secondary attack rate so low? (7) Why did epidemics in previous ages spread so rapidly, despite the lack of modern transport? (8) Why does experimental inoculation of seronegative humans fail to cause illness in all the volunteers? (9) Why has influenza mortality of the aged not declined as their vaccination rates increased? We review recent discoveries about vitamin D’s effects on innate immunity, human studies attempting sick-to-well transmission, naturalistic reports of human transmission, studies of serial interval, secondary attack rates, and relevant animal studies. We hypothesize that two factors explain the nine conundrums: vitamin D’s seasonal and population effects on innate immunity, and the presence of a subpopulation of “good infectors.” If true, our revision of Edgar Hope-Simpson’s theory has profound implications for the prevention of influenza.”
Sanakan’s first comment is mostly quibbling about definitions and peddling simple minded and uninformative ideas such as that R0 and HIT are universal constants, but his second is actually worse. It’s about the situation in India and quotes some seroprevalence studies in Indian cities and slums where the seroprevalence was around 50%. This in no way disproves what Nic said and in fact supports his and Gomez’s analysis of the impact of heterogeneity. I would expect Rt to be quite high in a slum and much higher than in rural or suburban areas. This means that variable heterogeneity absolutely must be taken into account to get a realistic HIT value for the country.
The simple minded assertion that these studies contradict Nic’s assertion is quite wrong.
I’ll just add that Sanakan’s oft repeated spiel on IFR suffers from the same simplistic pseudo-scientific approach that sees IFR as a universal constant.
Ioannidis never fell into this trap and always had a broad confidence interval. It shows us disrespect to repeat this spiel over and over again, both here and at other blogs.
1. Sanakan misrepresented what Ioannidis’ did with the Diamond Princess dataset. Actually Ioannidis’ confidence interval for IFR was from near 0 to 0.67% when adjusted for the US age demographics. That’s about what Ioannidis had in his later serological study meta-analysis too. In his Diamond Princess statement, Ioannidis suggests a midrange value of IFR of around 0.3%.
2. Ioannidis did roughly double his IFR estimates to account for the fact that more might die in the future and that cruise ship passengers might be more healthy than the general population. Sanakan misrepresented this by claiming that because 13 or 14 ended up dying that this invalidated Ioannidis’ analysis.
2. Sanakan misrepresented how the Danish blood donor study was used by myself and by Ioannidis. No one ever said it was a representative sample and properly noted that it was an estimate for healthy people under 70 years old.
3. The fundamental error here is to pretend that the large number of studies showing widely varying IFR from almost zero to over 1.4% somehow can be used to show that Ioannidis was “wrong.” No, it shows that IFR is highly variable because of the strong age dependent IFR for this disease. In addition, local factors such as variable pre-existing immunity will cause large variations between locations.
4. Ioannidis has been very prolific in the covid area. He has had scores of coauthors. The likely alternative explanation is that Ioannidis is widely respected and scientists compete to help him with his research. That anonymous nonscientist attack him is of no consequence.
5. Referring to the Santa Clara study as “garbage” is rhetorical and shows a crude understanding of how science often uses the best data available and that better data develops over time. It is analogous to referring to HADCRUT as “garbage” because later versions corrected some obvious data quality problems. In this Sanakan falls into the same trap as some nonscientist climate skeptics who he has repeated attacked in the most vile tterms.
In short Sanakan’s “work product” on IFR is pseudo-science with an obvious political motivation. It’s a complex situation and he does us a disservice by trying to weaponize it to attack people he doesn’t like.
I’ll just note that population fatality rates in the US in most states are between 0.1% and 0.2%. The pattern in the Dakotas for example indicates that herd immunity may be coming into play with case rates of over 10% of the population. Using the CDC’s estimate that comes to well north of 30% infected. We will see what happens and conclusions can not yet be dwawn.
Who to vaccinate first.
https://advances.sciencemag.org/content/7/6/eabf1374?utm_campaign=toc_advances_2021-02-05&et_rid=267930073&et_cid=3656961
Ceresokid: Thanks for the link, but I disagree with their conclusions. It is hard to tell from these sophisticated models why they predict what they predict.
I did some simple calculations. Those over 65 are about 10 times more likely to be occupying a hospital bed than adults under 65. If you want to stop hospitals from overflowing (our biggest problem IMO), vaccinate the oldest people first. If lockdowns are being caused by fear of overflowing hospitals, vaccinate the oldest first! This seems like a no-brainer.
Someone over 85 has a 21-fold greater chance of dying of COVID than a relatively old “essential worker” (50-64) and a life expectancy that is 5-fold shorter (5 vs 26 years). So vaccination the elderly first saves 4 times as many person-years as vaccinating an older essential worker. (If the elderly’s remaining years are unproductive retirement years, then remember that most of the older worker’s remaining years are retirement years too.) From the person-years saved perspective, vaccinating those under 50 first is really insane, because their risk of dying from COVID drops much more rapidly as you consider younger workers than life expectancy increases. Vaccinating someone over 85 saves 58 times as many person years as vaccinating someone 18-29. Vaccinate the oldest first, even if you want to vaccinate some “essential workers” first!
The working age population does test positive more than the elderly and therefore are contributing more to transmission than the elderly, but the difference is only a factor of 2-3. You could vaccinate essential workers who are frequently in contact with the public first, but those are the people who are already most likely to be immune from having had the infection! Cumulative cases in the US are approaching 10% of the population and have reached 10% of non-children (who are less likely to be infected and transmit). Assuming 1-2 undetected cases for every detected case, 20-30% of the population is already immune. A reasonable estimate might be that 40-60% of those in frequent contact with the public are already immune – and more are getting immune every day without vaccination and without dying or filling hospitals. Vaccinating this group first is inefficient. Vaccinating medical staff so older people can get necessary health care without fear makes sense, but half of those eligible have turned down vaccination!
(When I looked nursing home residents in September, they were 6-fold more likely to test positive than the average person. Vaccinating them first was the right thing to do.)
NIH now “neither for nor against” ivermectin for covid.
https://covid19criticalcare.com/wp-content/uploads/2021/01/FLCCC-PressRelease-NIH-Ivermectin-in-C19-Recommendation-Change-Jan15.2021-final.pdf
Recommendation
The COVID-19 Treatment Guidelines Panel (the Panel) has determined that currently there are insufficient data to recommend either for or against the use of ivermectin for the treatment of COVID-19. Results from adequately powered, well-designed, and well-conducted clinical trials are needed to provide more specific, evidence-based guidance on the role of ivermectin for the treatment of COVID-19.
https://www.covid19treatmentguidelines.nih.gov/statement-on-ivermectin/
In March 2020, the Front Line COVID-19 Critical Care Alliance (FLCCC) was created and led by
Professor Paul E. Marik to continuously review the rapidly emerging basic science, translational, and clinical data to develop a treatment protocol for COVID-19. The FLCCC then recently discovered that ivermectin, an anti-parasitic medicine, has highly potent anti-viral and anti-inflammatory properties against COVID-19. They then identified repeated, consistent, large magnitude improvements in clinical outcomes in multiple, large, randomized and observational controlled trials in both prophylaxis and treatment of COVID-19. Further, data showing impacts on population wide health outcomes have resulted from multiple, large “natural experiments” that occurred when various city mayors and regional health ministries within South American countries initiated “ivermectin distribution” campaigns to their citizen populations in the hopes the drug would prove effective
https://www.frontiersin.org/articles/10.3389/fphar.2021.643369/abstract
Has anyone else noticed how extremely repetitious Sanakan’s comments are. I was just reviewing some of them and saw the same spiel on IFR at least 4 times. I doubt that anyone else read the comments though. Judith, it really does clutter the comments. I think it likely that Sanakan has spent a lot of time preparing comments and documents (perhaps he gets paid to do this) and then when a topic comes up he just cuts and pastes, maybe adding a few new items or buried insults. The comments uniformly conform almost exactly to partisan talking points. This is confirmed in that he posts almost the same thing at other blogs and on twitter.
I did find this a couple of weeks ago. Ioannidis and several co-authors have re-analyzed the Santa Clara data. It’s very dense statistics and I can’t judge the details. It looks to me that with the improved methods, the results are pretty much the same as before.
https://arxiv.org/pdf/2011.14423.pdf
Deletions are the key to covid19 mutations. This Science Daily article reports research from Pittsburg University by Paul Duprex and colleagues which shows why deletions are the most likely sort of covid19 mutation to occur. That’s because even a coronavirus has a gene “proof-reader” that corrects mutations during replication. However a deletion can’t be repaired, there’s nothing to replace.
https://www.sciencedaily.com/releases/2021/02/210203144533.htm
Of course, randomly snipping off a part of the covid19 genome most of the time will render the virus dysfunctional and it will fail to replicate and spread. A dead end. But here and there in the genome there are short bits of RNA whose removal does not impair the virus’ vitality. Some may even enhance it. Such survivable deletions include ones that change the virus’ spike proteins, evading the immune response and – in some cases – the effectiveness of vaccines.
There are a limited number of such survivable and advantageous deletions in the viral genome and this is why the mutant strains emerging, from southern England, Brazil and South Africa for instance, are similar to each other sometimes involving the same deletion.
Finally another factor driving evolution of resistant covid19 strains is the severity of the infection in many patients and the length of time the person is sick with the virus. For weeks or even months. This gives time for the “cat and mouse” game between immune system and virus out of which successful mutant strains can arise. That could be bad news if severity of the infection carries survival advantage for the virus.
It’s not just these deletions that may be important. Recombination with other strains can make large changes suddenly. This sort of recombination is part of the evolution of some new viruses, but in this case, the paper addresses it as a way of acquiring traits from other strains.
“This report provides the first evidence for genetic recombination—a new way of evolution besides mutation in SARS-CoV-2. The existence of genetic recombination has the following implications: 2 different SARS-CoV-2 strains (here, hap_048 and hap_050) should have coinfected the same cell; a SARS-CoV-2 strain might acquire new traits such as virulence and drug susceptibility directly from other strains [12, 13]; the adaptability of SARS-CoV-2 to the human immune system might be significantly strengthened through genetic recombination; the accuracy of diagnosis based on serologic and molecular biology assays might be compromised [14]; and the transmission tracking based on phylogenetic tree could be misleading since the topology of the mutation route is a network rather than a tree.”
https://academic.oup.com/cid/article/71/15/884/5781085
I found Ioannidis’ article in Stat about the Diamond princess from March. It’s mostly about our lack of knowledge about COVID-19. All kinds of possibilities are mentioned. Sanakan and Josh rely on a few out of context sentences.
After rereading it I’m frankly sickened by how anonymous hacks on the internet can totally twist that article in what amounts to a political smear.
dpy6629: Somewhere above you commented on my data that you think ND and SD may have reached “herd immunity” with cumulative infections totaling almost 13% of the population (or possibly at 16% of some counties). I think the correct conclusion should be that we don’t know why the pandemic slowed in ND and SD, because it is impossible to distinguish between approaching herd immunity and changes in behavior and/or public health policy. As I discussed above, many of us intuitively believe herd immunity means an outbreak will die out no matter was season it is, whether or not people are taking precautions like social distancing or masks. However some people say the HIT can be influenced by these factors.
What I noticed is that the states neighboring ND and SD peaked at about the same time and have seen a huge drop-off in new cases – even though the cumulative percentage of the population infected was under 10%. ND hospitals were pronounced at “100% of capacity” in early November while cases were still rising, and I suspect neighboring states began falling about the same time as ND because of fear that they were on the same path changed behavior. Meanwhile, you can find some counties where cumulative cases are above 16%. So I don’t think approaching herd immunity played a major role in slowing in neighboring states.
Until some seropositivity studies are done in the places like ND, I don’t think we will know. Another surge might occur in ND, but soon vaccination hopefully will make this impossible.
FWIW, I think the early seropositivity studies done this spring could have been biased by a higher than expected false positive rate in the field compared with the false positive rate determined in the laboratory.
I agree I’m not sure about the herd immunity. I am assuming that no changes in behavior have occurred over the winter. I do think that seroprevalence studies are also not definitive because of pre-existing immunity which seems to be common especially among younger people.
This mess of ignorance is in my view a nearly constant feature of viral epidemiology. I’m even beginning to question the value of concepts like herd immunity for viral epidemics. The only hope is better technology that can develop vaccines and treatments much faster than at present.
The other thing I’m increasingly certain of is that policy responses have been driven by pseudoscientific and political narratives and guesses by “experts.” The modern West I fear is drifting down a corrupt path not seen since the Guilded Age.
dpy6629: The infection detection rate links the IFR and the CFR. At one point this fall, NJ, NY and MA had roughly the same number of deaths/100,000 as ND and SD, but the latter had twice as many cumulative cases. In the former states, most of the deaths were in spring (their winter surge had barely begun) and almost all of the ND pandemic occurred in the fall when testing was more widely available. Assuming their were no improvement in treatment, this suggests the detection rate was roughly two fold better in fall compared with spring. However, treatment has improved modestly and those getting infected and going to the hospital have become younger as the elderly learned to protect themselves. If I’ve got this right, there hasn’t been a big change in the detection rate from spring to fall.
Steve McIntyre at Climate audit dot org has an insightful comment on Sweden and Canada
Steve M comment – “I agree that comparisons between Sweden and other Nordic countries are tainted because of relative “downstream-ness” as noted in blog. Canada was less impacted than northern tier US states. COVID arrived here mostly via US, so we had more advance warning and foothold less.”
Second comment – “In other words, if Sweden had more COVID-19 deaths that its neighbors, it’s not so much because it dediced not to go on lockdown, but mostly because,by the time people realized that the pandemic had reached Europe and governments started to act, for whatever reasons, the epidemic was already more advanced in Sweden than in other Nordic countries.”
Steve M does a good job of presenting the broader perspective while paying attention to the minute details.
Note that numerous commentators have bashed Sweden’s policy compared to its nordic neighbors. While failing to realized that Norway & Finland were the outliers and Sweden fallen much closer to the middle of the rest of europe.
I wish Steve would return to blogging. I hate Twitter with a passion. Even though no else cares, I will point out that this Nordic countries thing is Josh’s favorite talking point. McIntyre is vastly more qualified and credible on this than our local anonymous undisciplined intellects
The Chamie-Quintero study from Peru shows the West could be only weeks away from reducing the Covid death toll if we used the safe cheap sheep-dip, lice-killing Ivermectin like less wealthy countries do.
A study across the states of Peru found that after Ivermectin was introduced, deaths started to fall about 11 days later, and within a month after that, deaths were down around 75%.
…
Ivermectin first showed promise against Covid in a lab dish back in April. It wasn’t clear if it would work as well in people, but it was so cheap and safe that the rich world …o, pretended it didn’t exist and tested expensive drugs instead. (Big-Pharma don’t make big profits from old cheap drugs that are out of patent.) Meanwhile less wealthy countries were desperate enough to try it en masse. And Ivermectin has had remarkable success.
https://joannenova.com.au/2021/02/in-peru-ivermectin-cut-covid-deaths-by-75-in-6-weeks-cheap-safe-and-quite-ignored/
Jim2: https://joannenova.com.au/2021/02/in-peru-ivermectin-cut-covid-deaths-by-75-in-6-weeks-cheap-safe-and-quite-ignored/
Thanks for the link
https://www.medrxiv.org/content/10.1101/2020.11.11.20229708v1
This is an interesting analysis of mortality in Sweden and Norway.
Peer reviewed critique of “Lion’s” study of lockdown effectiveness published.
https://onlinelibrary.wiley.com/doi/abs/10.1111/eci.13518
Twitter summary here from author for those who can’t access paywall.
the conclusions cannot possibly be drawn from the methodology and design, which themselves tell us very little about whether harsh restrictions are or are not effective
Ouch.
https://mobile.twitter.com/lonnibesancon/status/1361921174473736192
> 4/ The analysis does not correct for the relationship between interventions and case counts.
So totally obvious, it’s really shocking that it got past 🦁’s
Clearly, the precipitating conditions would predict the severity of the interventions (on average) and also the kinds of outcomes independent, to some degree, of the efficacy of the interventions. To not control for that is mind-blowing.
Sad to see 🦁’s reduced to 🐈⬛’ s.
This gets old very fast. There is a big literature on NPI effectiveness and I’ve cited quite a few papers above including a much more brutal takedown of Flaxman’s paper. Quite frankly I don’t care what you think as its just biased activists posting random stuff. I’m sure Ioannidis and coauthors will have a response. It’s how science works. Teenagers post sarcastic meaningless comments on blogs. Grown ups do science. That you value your time so little is a sure sign its not worth much.
dpy
“That you value your time so little is a sure sign its not worth much.”
the irony. It burns
I’m skeptical, but at least it’s logical:
https://twitter.com/davidwdowdy/status/1358992725677395973?s=20
I read it, and don’t really know what to think. The guy is a smart data scientist whom I’ve followed to some extent for months. But I really don’t know how good his formula is. At first glance, his estimate of total number infected for the one data point I have – Arizona CDC survey last fall – is maybe 15% high, which is not too bad.
I don’t have time to pull his GitHub archive and really dig into it. Perhaps some experts such as Atomsk might comment.
He’s doing curve fitting, which can be dangerous – as we’ve seen in our discussions. He is humble about his results – he’s trying to figure this out, not push a viewpoint, which I appreciate.
A couple of links to his work that some might find interesting.
First – his comments on herd immunity seeking on Twitter. Not to surprising, although I think he misunderstands to some extend how epidemiologists probably think about this: https://twitter.com/youyanggu/status/1359941463757516802
Second, his modeling site. He abandoned this last November, but then re-invigorated it. He has graphs for the US, and for each state: https://covid19-projections.com/
I’m assuming this was mis-nested.
The link to his comments on herd immunity didn’t work
It works for me. Here it is again – and again, with the caveat that I don’t agree with all of it.
https://twitter.com/youyanggu/status/1359941463757516802
and in HTML: twitter comments click here
Interesting – this time, WordPress went out and grabbed the first tweet of the thread and put it inline.
> Interesting – this time, WordPress went out and grabbed the first tweet of the thread and put it inline.
I’m thinking that might have something to do with why the link didn’t work for me. I find that the manner in which I put in a Twitter link determines whether it shows inline or just as a URL. One thing I’ve found is that it depends on whether “mobile” is included in the URL.
https://twitter.com/numbers_truth/status/1354915510400339969
Keep in mind, COVID deaths are likely under-reported.
You leftest most favorite pandemic is slowly coming to a close. I’m sure you’ll miss it greatly.
Yes. No doubt. I really enjoy not going shopping, not eating out, not hearing live music, not going to the movies, not traveling, not seeing family and friends, hundreds of thousands of people dying, many more getting seriously ill, and having so many people unemployed.
You have such powers of insight.
Hi Joshua, did you notice that your chart claims flu deaths (as in influenza, not Covid) are up year over year? The year you claim the flu “disappeared?”
The “unclassified” is also up 30,000+ year over year.
My suspicion is that the numbers for 2020 will be out of whack for a while. First, Covid was big starting in March 2020. Winter kills, so you have January and February and March 2020 – the peak of flu season, peak cold weather – killing people pre-Covid for the most part. This would partially explain why you think the flu “disappeared” thanks to Covid even though flu deaths were up year over year. The 2020 flu season, just like all those before it, was really January-March. The 2021 flu season will be as well.
The flu is spread by kids who get it at school. Schools are open in competent states and even the incompetent ones are slowly trying. So 2021 could still be a bad year for the flu, just like 2020 was.
That “not-classified?” A whole lot of people couldn’t go to their doctor and were afraid to go to the hospital.
> Hi Joshua, did you notice that your chart claims flu deaths (as in influenza, not Covid) are up year over year? The year you claim the flu “disappeared?”
(1) It’s not “my” chart.
(2) I never claimed the flu “disappeared.”
(3) I don’t see seasonal flu on the chart I posted.
> even though flu deaths were up year over year.
Sorry – but there is tons of data that seasonal flu deaths have been waaaaaaay down during the COVID pandemic. That’s one way to know “it’s just because of less testing” is nonsense.
> So 2021 could still be a bad year for the flu, just like 2020 was.
Wow. Ok.
Here’s a good link re 2019-20 flu season in the US. It seems that tragically, it was a bad year for children’s deaths from the flu:
https://www.advisory.com/daily-briefing/2020/05/05/flu-update
“(1) It’s not “my” chart.
(2) I never claimed the flu “disappeared.”
(3) I don’t see seasonal flu on the chart I posted.”
1- you posted the chart (which is CDC data by the way- now you “deny” the CDC?)
2- You did repeatedly.
3- read the chart you like. It’s the line item right below diabetes.
The CDC, again. Flu season peak is January, February, March:
https://www.cdc.gov/flu/about/season/flu-season.htm
That’s why 2020 can be both worse than 2019 in flu cases and you still think the case counts are “down” (so far). The flu peak in 2020 was pre-Covid and killed more than 2019 season. And we don’t yet know the results of the next season because we’re in the middle of the peak right this minute. At a time when there are a whole lot of sick people testing negative for COVID.
> 1- you posted the chart (which is CDC data by the way- now you “deny” the CDC?)
It’s not “my” chart. I don’t own anything that I post or link to. I own my words. Which is why…
‘2- You did repeatedly.
Is ridiculous. Not only did I not do it repeatedly, I didn’t do it even once. Not Once. Not once.
> 3- read the chart you like. It’s the line item right below diabetes.
On that chart, flu isn’t differentiated from pneumonia. I posted a good link regarding data on 19-20 seasonal flu.
> And we don’t yet know the results of the next season because we’re in the middle of the peak right this minute.
There’s tons o’ data that deaths this season are anomalously low by historic standards – not just in comparison to last year (which was bad).
Ah, it was VTG who said it “disappeared,” you simply attacked and maligned anyone who disagreed with him, noting that it was, in fact, gone as your friend in Boston told you.
Anyway, to recap: VTG/Joshua: the flu has “disappeared” in 2020 or otherwise “way down” in 2020 or “something-something.”
CDC says: flu deaths in 2020 are higher than 2019.
To which Joshua says the CDC must not be tracking the flu because they lump it in with pneumonia (non-covid). “Pneumonia,” this year only!, is defined as an illness unrelated to flu even though the CDC says it is, which affects people who must not actually be sick, because how could they be sick in a world where lockdowns work (except for their intended virus)?
Jeff gives Joshua an “out” from painting himself in the corner- Joshua could note that the CDC is correct, flu is “up” because the 2020 flu season was pre-COVID- Jan-March. Of course that means he may be correct- the lockdowns will reduce the flu during the season we’re currently in, because the flu season is on-going. But that also means the flu season could just be beginning. So, Josh doesn’t take the out.
Frank, meso –
If you see this, I’d like to read your thoughts:
https://covid19-projections.com/estimating-true-infections-revisited/
-snip-
We present a simple nowcasting model that 1) computes a standardized test positivity rate for every state in the United States and 2) uses the adjusted test positivity rate and confirmed cases to estimate the true prevalence of COVID-19 infections for every US state and county. The heuristics we present are computable using simple arithmetic and are hence easily accessible.
Joshua: https://covid19-projections.com/estimating-true-infections-revisited/
Thank you for the link.
They estimate the case count in the US is a bit over twice the reported.
They did the estimates in Europe too, and it’s worse:
https://www.degruyter.com/document/doi/10.1515/em-2020-0024/html
Estimated cases are about 8 times higher in France than reported counts. Just under five times more in Germany. etc etc. Five times worse than reported seems to be normal for EU countries.