By Nic Lewis
Key points about the recent Nature paper by Flaxman and other Imperial College modellers
1) The transition from rising to declining recorded COVID-19 deaths in the in 11 European countries that they studied imply that transmission of COVID-19 must have reduced substantially.
The study was bound to find that together the five government non-pharmaceutical interventions (NPI) they considered contributed essentially 100% of the reduction in COVID-19 transmission, since in their model there is nothing else that could cause it.
2) The prior distribution they used for the effects of NPIs on transmission in their subjective Bayesian statistical method hugely favours finding that almost all the reduction in transmission is due to one, or possibly two, NPIs with all the others having a negligible effect.
The probability density of the prior distribution at their median estimates of the effect on transmission of each type of NPI, which allocate essentially all the reduction in transmission to lockdowns, was many billion times greater than it would have been if the same total estimated reduction had been spread evenly across the types of NPI.
3) Which intervention(s) is/are found to be important depends critically on the assumptions regarding the delay from infection to death. When using their probabilistic assumptions regarding the delay from infection to death, a huge (and highly improbable given other assumptions they made) country-specific effect is required to explain the reduction in transmission in Sweden, where no lockdown occurred. If delays from infection to death are increased by just three days, their model no longer finds lockdowns to have the largest effect, and a more moderate country-specific effect is required to explain the reduction in transmission in Sweden.
4)The estimated relative strengths of different NPIs are also considerably affected by the use of an alternative prior distribution for their effects on transmission that does not strongly bias the estimation of most of them towards a negligible level. They are also considerably affected by phasing in over a few days the effects of the two NPIs that seem unlikely to have had their full effect on their date of implementation.
5) It follows from the above that that study provides no information whatsoever as to the actual contribution from all NPI combined to the reduction in transmission, and nor does it provide robust estimates of relative effects of different NPI.
Introduction
On 8 June 2020, Nature published a paper (Flaxman et al. 2020[1]) by modellers in the Imperial College OCIVD-19 response team. Its abstract ends with:
Our results show that major non-pharmaceutical interventions and lockdown in particular have had a large effect on reducing transmission. Continued intervention should be considered to keep transmission of SARS-CoV-2 under control.
Using a counterfactual model, the paper also estimated the impact of interventions on deaths from COVID-19 in the 11 European countries studied, saying:
We find that, across 11 countries, since the beginning of the epidemic, 3,100,000 [2,800,000 – 3,500,000] deaths have been averted due to interventions.
The mainstream media publicised the ‘3 million deaths saved’ claim, without critically appraising the paper or, generally, mentioning the relevant caveat in the paper:
The counterfactual model without interventions is illustrative only and reflects our model assumptions.
In Imperial College’s press release Dr Flaxman ignored his own caveat, saying
Using a model based on data from the number of deaths in 11 European countries, it is clear to us that non-pharmaceutical interventions– such as lockdown and school closures, have saved about 3.1 million lives in these countries
In this article I examine the main claim – that major non-pharmaceutical interventions (NPI) have had a large effect on reducing transmission of COVID-19, to which the inferred reduction in deaths is attributable, with almost all the reduction due to lockdowns. I show that this claim is strongly dependent on the assumptions made and is highly dubious.
The case of Sweden, where the authors find the reduction in transmission to have been only moderately weaker than in other countries despite no lockdown having occurred, is prima facie evidence against the paper’s main claim.
How the effects of lockdowns and other interventions were estimated
Flaxman et al. employ a ‘hierarchical Bayesian’ statistical model. It uses data on daily deaths (up to 5 May 2020, when two countries relaxed their lockdowns), the dates of imposition of five types of NPI (school or university closure, case-based self isolation, public events banned, lockdown ordered and social distancing encouraged), and estimates of the infection fatality rate, for each of 11 European countries.[2] Using these data, the model infers what time profiles of the effective reproduction number (Rt, the number of people whom an infected person in turn infects) – and hence of new infections – would produce the best match between projected and recorded deaths for each country. To do so it uses a simple model of epidemic growth and probabilistic estimates, common to all countries, of the time from infection to death and of the generation time (that from a person becoming infected to them infecting others). The assumed infection fatality rate (IFR) is common between countries for each age band, but reflects the age-structure of each country’s population. It averages slightly over 1%.
A separate initial value, R0 (the basic reproduction number), of the reproduction number Rt is inferred for each country. Rt then changes from R0 in stepwise fashion at the date of each NPI, which act multiplicatively with an equally strong inferred effect for all countries. Each country’s epidemic is seeded by a series of infections starting 30 days prior to a total of 10 recorded deaths.[3]
The model is described in more detail here, and is illustrated in Figure 1, taken from Flaxman et al.
Fig. 1. Reproduction of Flaxman et al. Extended Data Fig. 3: Summary of model components
The treatment of interventions
The model uses no information on NPI’s except their type and their implementation date in each country. NPI of each type are treated as having the same (multiplicative) effect on Rt in each country. Each type of NPI is treated identically. As well as the five types of actual interventions, all first interventions (whatever type) are treated as an extra type of intervention, for each country occurring on the date of implementation of its first actual NPI (almost always either self isolation or public events ban, and never lockdown). Hence there are six NPIs with shared values for all countries.
In addition, a pseudo-NPI with a strength that is estimated separately for each country is treated as taking place on the same date as the last actual NPI. These country-specific pseudo-NPIs allow for variation between countries in the effectiveness of the implementation of their NPI. They are probabilistically constrained to be relatively small, making a country-specific effect large enough to cause a halving of Rt exceedingly improbable.
In all 11 countries the exponential growth in infections and deaths experienced early in the epidemics slowed and then turned negative, with infections and deaths decreasing. This implies that in all 11 countries Rt decreased very substantially, to below one, since the start of their epidemics.
In the Flaxman et al. model the only factor that can cause Rt to decrease significantly is the effect of each NPI. Therefore, the estimated overall effect of the NPIs in reducing Rt, and hence deaths resulting from COVID-19 disease, is bound to be very strong.
The only non-NPI factor that affects Rt in the Flaxman et al. model is the reduction arising from the proportion of the population susceptible to infection (set at 100% initially) gradually diminishing over time due to individuals already infected by COVID-19 becoming immune to it. This reduction is very small in their model, for two reasons:
- they make the very unrealistic assumption that all individuals in a country are equally susceptible to COVID-19 and, if infected, are equally likely to infect others.
- the relatively high infection fatality rates they assume result in only very small proportions of countries’ populations becoming infected in their model.
Therefore, their model has to attribute almost all the overall reduction in Rt to government interventions.
Factors not considered by Flaxman et al., all of which are highly likely to have caused some reduction in COVID-19 transmission, and which between them may well have caused substantial reductions in Rt in all 11 countries, include:
- population heterogeneity in social connectivity – which generates highly correlated heterogeneity in both susceptibility and infectivity – and in other factors determining susceptibility to COVID-19
- unforced changes in the behaviour of individuals as they adjust it to reflect COVID-19 risk
- seasonal factors: infections by common coronaviruses peak in the winter and diminish greatly as spring progresses.
As is well known by competent epidemiologists, the first of the above-mentioned factors causes Rt to diminish faster, potentially much faster, with the number of people who have been infected than if it were proportional to the number of people remaining uninfected, as assumed by Flaxman et al. The other factors directly reduce Rt.
If follows that Flaxman et al.’s counterfactual case, which predicts ~3,200,000 deaths in the absence of any NPIs (their ‘counterfactual model’), is completely unrealistic, as therefore is their estimate of 3,100,000 lives saved by interventions.
It also follows that Flaxman et al.’s claim:
Our estimates imply that the populations in Europe are not close to herd immunity (~70% if R0 is 3.8)
may be invalid. As shown here, due to population heterogeneity in susceptibility and infectivity the herd immunity threshold it is bound to be lower – quite possibly very substantially so – than if, as required for it to be ~70% at an R0 of 3.8, populations are homogeneous.
Flaxman et al.’s assertion that all the reduction in transmission (i.e., the reduction in Rt) was due to NPIs, other than very small reduction as more people have been infected and become immune, is unsound. Nevertheless, it seems quite likely that NPIs have had a significant, perhaps substantial, effect on Rt. However, given the confounding effects of the other factors mentioned it is impossible reliably to estimate the total effect of NPIs on Rt and hence on deaths.
Even when making the unrealistic assumption that almost all the reduction in Rt was due to interventions, any allocation of that reduction between the NPIs is very fragile. Flaxman et al. accept this in relation to NPIs other than lockdown, writing:
Most interventions were implemented in rapid succession in many countries, and as such it is difficult to disentangle individual effect sizes of each intervention. In our analysis we find that only the effect of lockdown is identifiable, …
On their median estimates, lockdown caused an 82% reduction in Rt, whereas no other NPI caused as much as a 1% reduction in Rt. While it would not be particularly surprising if such a drastic intervention as lockdown had had stronger effects than other NPIs, even if lockdown had a strong effect one would expect some other NPIs to have had a significant effect. So how did Flaxman et al. find that, remarkably, almost the entire effect of interventions was due to lockdown? The answer, which turns out to be two-fold, shows that their finding is not credible.
Why Flaxman et al. found almost all reduction in COVID-19 transmission to be attributable to a single intervention
Flaxman et al. use a subjective Bayesian statistical method. I have repeatedly criticised this type of Bayesian method in the climate science field, but – probably due to its ease of use – it remains standard practice there and in many other fields.
A subjective Bayesian method requires prior probability distributions to be assigned for each unknown parameter whose value is to be inferred. These prior distributions are then modified by the likelihood function, which reflects how well the modelled deaths fit the daily deaths data at varying values of the parameters, in order to arrive at a ‘posterior’ probability distribution for the parameter values. They use a common method of achieving this that results in a large number of quasi-random draws (‘posterior draws’) from the derived posterior probability distribution.
They represent the strength of interventions by a six dimensional parameter alpha (five actual NPIs plus the synthetic first intervention NPI), with the corresponding effect of intervention i (i being 1, 2,3, 4, 5 or 6)[4] on Rt being to multiply it by exp(-alpha[i]).
The combined effect of all interventions is then to multiply Rt by exp[-(alpha[1] + alpha[2] + alpha[3] + alpha[4] + alpha[5] + alpha[6])][5], which depends only on the sum of the individual alpha values. Their own posterior draws show a median value of the sum of the alphas of 1.75, which corresponds to an 83% reduction in transmission (1 – e−1.75 = 0.83).
The prior distribution assigned by the authors to the strength of the reduction in Rt caused by each intervention is of particular concern. Each of the six alpha values is assigned a gamma-distributed prior probability distribution; a small offset is applied, so that the gamma-distributed values inferred initially are marginally higher, but that is a cosmetic feature.[6] The authors write:
The intuition behind this prior is that it encodes our null belief that interventions could equally increase or decrease Rt, and the data should inform which.
That is not in fact true. As the left hand panel of Figure 2 shows, their prior allows each intervention to decrease Rt by up to 100%, but only to increase it by less than 1%. And the combined effect on transmission of all interventions (right hand panel) can only vary between –100% and + 5%. However, since the trajectory of the deaths data is, on their assumptions, bound to result in all interventions combined being found to strongly reduce transmission, the +5% limit is of no real consequence.
Fig. 2. Reproduction of the upper panels of Flaxman et al. Supplementary Fig. 3: Cumulative distribution function F(x) of the prior for one intervention’s multiplicative effect x (= e–α) on transmission (left) or for the effect of all interventions combined (= e–Σα) (right).
On the face of it, the combined effect of the six-dimensional joint alpha prior distribution looks fairly uniform over the range in which the estimated reduction in Rt could fall; it assigns a similar probability to a reduction in the range 40% to 50% and in the range 80% to 90%, for example. However, that only looks at one aspect of the six-dimensional prior distribution.
If I take the sum of the six alphas to be 1.75 (the median sum from their posterior draws) and set them to be all equal, at 1.75/6, their joint prior probability density is 0.0023. But if I set one of the alpha values to 1.70 and the remaining five to 0.01, giving the same overall reduction in transmission, the prior probability density is 64.3. That means their prior distribution assigns a 28,000 times higher prior probability assumption to this case, where one type of intervention has a completely dominating effect relative to all the others, than to a case where the same overall reduction in transmission is caused equally by all types of intervention. The reason is that the offset-gamma distribution used assigns a strongly increasing probability density as an alpha value decreases towards −0.008, its lowest permitted level, favouring cases where the effect of all but one or two NPIs is estimated to be almost zero.
So it is unsurprising that they found a single intervention to be totally dominant.
The median individual alpha values in their 2,000 archived posterior draws are −0.007, −0.007, −0.007, −0.007, 1.699 and −0.006. So all interventions except lockdown were estimated to have a completely negligible effect.
The median ratio, across their own posterior draws for alpha, of the actual prior probability to what it would have been if in each draw the total effect of the intervention had been spread evenly across them, was in fact 392 billion to one!
It is not clear that the authors realised that the prior distribution they used very strongly favoured finding that most interventions had a negligible effect, and I very much doubt that any of the peer reviewers appreciated that this was the case.
The Sweden problem
Using the code and data accompanying the Nature paper as is, except with the 8,000 draws split between 4 not 5 chains to better match my computer, I can accurately replicate Flaxman et al.’s findings, with lockdown accounting for almost the entire reduction in Rt (Figure 3).
Fig. 3. Effect of interventions on Rt in the base case, with all aspects of the model as per the original version (that archived for the Nature paper). The red First intervention estimate includes the effect of the synthetic first intervention NPI and so only applies for countries where the NPI concerned was the first to be implemented; it should be ignored in all other cases. Mean relative percentage reduction in Rt is shown for each NPI (filled circle) together with the 95% posterior credible intervals (line). If 100% reduction is achieved, Rt = 0 and there is no more transmission of COVID-19.
Sweden did not have a lockdown, but it still had a large reduction in Rt, albeit one not quite as large as the average for other countries. So how did the model account for that? This is where the country specific factors, which are treated as occurring on the date of the last actual intervention and in effect are an addition to its alpha, come in.
The country specific factors are given an apparently small influence, being zero-mean normally distributed with a standard deviation that is itself zero mean normal+ distributed[7] with a standard deviation of 0.2. But for Sweden a value of 1.27, in the far tail of the resulting distribution, was inferred. The probability of such a large country factor arising by chance appears to be about 1 in 2,000. That in itself implies that their model does not adequately represent reality.
Using a less informative prior
I investigated use of a prior distribution for the six alpha parameters that was essentially flat over the alpha parameter range relevant for NPI, both for each parameter separately and for the six-dimensional joint alpha parameter. For technical reasons, rather than using a uniform distribution I chose an independent zero mean normal distribution with a standard deviation of 10 as the prior distribution for each parameter. I hereafter refer to this as the ‘flat prior distribution’, even though it is not quite flat over the parameter range of interest (approximately 0 to 2).
I then ran the model using the same assumptions, but using the flat prior distribution rather than the original offset-gamma prior distribution. Doing so should eliminate the previous strong bias towards finding that most interventions had almost no effect.
The resulting estimates of the effect of each intervention were as shown in Figure 4. The estimated effects of NPI other than lockdown all increase markedly from their near zero values when using the original prior, but the contribution of lockdown remains dominant.
Fig. 4. Effect of interventions on Rt : as in Fig. 3, but with the flat prior distribution for alpha substituted for the offset-gamma prior distribution in the original model..
The country specific factor for Sweden was slightly less high than before, at 1.12. The probability of such a large country factor arising by chance appears to be about 1 in 900; still minute.
So, even when using the flat prior, the Flaxman et al. model does not adequately fit reality. The problem is that, as it still estimates lockdown to account for the vast bulk of the total reduction in Rt, it cannot adequately account for the reduction in Rt that occurred in Sweden, where there was no lockdown.
Why Flaxman et al. found lockdown was the intervention that dominated the reduction in COVID-19 transmission
I have explained why it to be expected, given Flaxman et al.’s choice of prior distribution for the effect of interventions on the transmission of COVID-19, that a single type of intervention (or at most two types) would account for the vast bulk of the reduction in Rt. But why lockdown?
The key here seems to be that lockdown was, other than in Sweden, on average imposed at a point in time that, allowing for the assumed probabilistic delay between infection and death, would result in deaths peaking at about the time that they actually peaked. Also, the timing of lockdown, relative to the peak in recorded deaths, differed slightly less between countries that locked-down than was the case for most other interventions.
Flaxman et al. took probabilistic estimates of the delay from infection to symptoms appearing and from symptoms appearing until death, with assumed mean values of 5.1 and 17.8 days respectively, and added them to obtain the infection to death delay values. The 5.1 day delay from infection to onset of symptoms seems reasonable. But the 17.8 days mean from onset of symptoms until death looks as if it may be on the short side for European countries. Ideally, a separate onset of symptoms to death delay distribution would have been estimated for each country. However, the authors may well have been unable to find suitable European data. They actually used a value estimated by Verity et al.[8] (also members of the Imperial College COVID-19 modelling team) from just 24 cases in mainland China.
One of the peer reviewers suggested that the value Flaxman et al. were using for the delay from onset of symptoms until death of (in the originally-submitted manuscript[9] being reviewed)18.8 days, not 17.8 days, was rather short, writing:
it is smaller than preliminary estimates available from hospitalization data in Europe (about 5-6 days from onset to hospitalization, at least 2 weeks in the hospital)
I therefore increased the average delay from onset of symptoms to death slightly.
I also took the opportunity to correct the dates used in the model inputs for school/university closure in Sweden and for self-isolation in Spain to those given in Flaxman et al. Extended Data Figure 4, which agree to those in their Supplementary Table 2.
I found that adding 3 days to the infection to death delay, bringing the average onset of symptoms to death delay to ~21 days (median 19.6 days) – which is fully consistent with the peer reviewer’s comment – was adequate to reduce the problem of Sweden needing a very large country-specific factor. That factor was then estimated at ~0.4, to match the reduction in transmission in Sweden – still over twice as large as for any other country, but no longer statistically-inconsistent with their assumptions.
The resulting estimated effectiveness of the various interventions, using the authors’ original prior distribution for alpha, is shown in Figure 5.
Fig. 5. Effect of interventions on Rt : as in Fig. 3 (original prior) but with the infection to death delay increased by 3 days, and one intervention date corrected for each of Spain and Sweden (see text).
School closure is now found to have a slightly stronger effect on transmission than lockdown. This may seem rather unlikely in reality, but the model has no information to go on regarding the likely relative strengths of each type of intervention – it just knows when they were implemented in each country. Other interventions are found to have almost zero mean effect, as is to be expected given the nature of the original prior distribution.
Using instead the flat prior gives slightly different estimates of the effectiveness of the various interventions (Figure 6), with school closure not having quite as strong an effect as when using the original prior. The effects of social distancing, and to a slightly lesser extent public events ban and self isolation (one of which is generally the first intervention, so the red line applies to it), all cease to be negligible.
Fig. 6. Effect of interventions on Rt : as in Fig. 5, with the infection to death delay increased by 3 days, but using the flat prior distribution instead of the original prior distribution.
If the infection to death delay is increased by 5 rather than 3 days from Flaxman et al.’s assumed probabilistic magnitude – arguably still as reasonable as Flaxman et al.’s assumption – and the original prior used, the changes in the relative effectiveness of different interventions become even more marked (Figure 7). Lockdown is now estimated to have far less effect than school closure, while social distancing now has a significant effect. The country-specific factor for Sweden becomes small.
Fig. 7. Effect of interventions on Rt : as in Fig. 5 (original prior) but with the infection to death delay increased by 5 days not 3 days.
When the flat prior is used instead, the estimated effect of school closure reduces while that of all other interventions increases (Figure 8).
Fig. 8. Effect of interventions on Rt : as in Fig. 6 (flat prior) but with the infection to death delay increased by 5 days not 3 days.
Finally, I investigated the effects of phasing in certain of the interventions. Flaxman et al.’s assumption that all interventions immediately have their full effect on their date of implementation is questionable. It may not be too unrealistic for closing schools, banning public events and decreeing a lockdown, all of which it is feasible to enforce. However, responses to self isolation advice and social distancing encouragement (which both generally preceded a lockdown) are more within the discretion of the individuals concerned, and very arguably would take a little time to reach their final strength.
I examined phasing in over four days the effects of just those two NPIs, with their strength increasing evenly from 25% on the date of implementation to 100% three days later. The result, using the original prior distribution for alpha and making a ~3 day increase in the delay from symptoms to death, is shown in Figure 9. The strength of the reduction in transmission attributed to lockdown reduces slightly compared with the no phase-in case, while than attributed to social distancing increases.
Fig. 9. Effect of interventions on Rt : as in Fig. 5 (original prior), but with the effects of self isolation and social distancing phased in over 4 days and the infection to death delay increased by 3.2 days.
Finally, I repeated this experiment using the flat prior (Figure 10). The strength of the reduction in transmission attributed to lockdown reduces noticeably compared with the no phase-in case, although it is still larger than that of school closure (the estimated effect of which reduces only marginally), while the estimated effects of banning public events and (particularly) social distancing increase markedly.
Fig.10. Effect of interventions on Rt : as in Fig. 6 (flat prior), but with the effects of self isolation and social distancing phased in over 4 days and the infection to death delay increased by 3.2 days.
Conclusions
First and foremost, the failure of Flaxman et al.’s model to consider other possible causes apart from NPI of the large reductions in COVID-19 transmission that have occurred makes it conclusions as to the overall effect of NPI unscientific and unsupportable. That is because the model is bound to find that NPI together account for the entire reduction in transmission that has evidently occurred.
Secondly, their finding that almost all the large reductions in transmission that the model infers occurred were due to lockdowns, with other interventions having almost no effect, has been shown to be unsupportable, for two reasons:
- the prior distribution that they used for the strength of NPI effects is hugely biased towards finding that most interventions had essentially zero effect on transmission, with almost the entire reduction being caused by just one or two NPI.
- the relative strength of different interventions inferred by the model is extremely sensitive to the assumptions made regarding the average delay from infection to death, and to a lesser extent to whether self isolation and social distancing are taken to exert their full strength immediately upon implementation or are phased in over a few days.
It seems likely that the inferred relative strengths of the various NPIs are also highly sensitive to other assumptions made by Flaxman et al., and to structural features of their model. For instance, their assumption that the effect of different interventions on transmission is multiplicative rather than additive will have affected the estimated relative strengths of different types of NPI, maybe substantially so. The basic problem is that simply knowing the dates of implementation of the various NPI in each country does not provide sufficient information to enable robust estimation of their relative effects on transmission, given the many sources of uncertainty and the differences in multiple regards between the various countries.
Nicholas Lewis
[1] Flaxman, S., Mishra, S., Gandy, A. et al. Estimating the effects of non-pharmaceutical interventions on COVID-19 in Europe. Nature (2020). https://doi.org/10.1038/s41586-020-2405-7
[2] Denmark, Italy, Germany, Spain, United Kingdom, France, Norway, Belgium, Austria, Sweden and Switzerland.
[3] The seeding continues for 6 days, with the average number of seed infections per day being inferred by the model.
[4] The numbering of interventions used in their code is 1. school (and/or university) closure ordered; 2. case-based self isolation mandated; 3. public events banned; 4. first intervention; 5. lockdown ordered; and 6. social distancing encouraged.
[5] In mathematical notation, exp[-(alpha[1] + alpha[2] + alpha[3] + alpha[4] + alpha[5] + alpha[6])] is written e–Σα.
[6] The alpha distributions are defined by αi ~ Gamma( shape=1/6, scale=1) − loge(1.05)/6. Hence alpha can range between −loge(1.05)/6 (approximately −0.008) and plus infinity.
[7] “Normal+” means a normal distribution with the negative part of the distribution excluded.
[8] Verity R, Okell LC, Dorigatti I, et al. Estimates of the severity of COVID-19 disease. medRxiv 13 March 2020; https://www.medrxiv.org/content/10.1101/2020.03.09.20033357v1.
[9] The original Flaxman et al. manuscript was submitted on 30 March 2020, the same date as Imperial College published “Report 13: Estimating the number of infections and the impact of non-pharmaceutical interventions on COVID-19 in 11 European countries.”, by the same (or almost the same) authors: https://spiral.imperial.ac.uk/bitstream/10044/1/77731/9/2020-03-30-COVID19-Report-13.pdf . From the referencing of comments in the Nature peer review file, it appears that the original Flaxman et al. manuscript was almost identical to Report 13.
Originally posted here, where a pdf copy is also available
Thanks for the analysis.
It is surprising that anyone would attribute all the reductions in Rt to lockdowns, since other methods clearly would have some effect on Rt. In fact, it is unreasonable to attribute all the reductions to governmental actions, because it became clear that people would take NPI’s without the actions. Given those factors, the modelers should have thrown out their results based on common sense.
But, it appears to me that governmental actions do have an impact in reducing Rt. We can look at Arizona, where I live. We had Rt clearly below one after a stay-at-home order (a pretty limited lockdown the way it was written) went into effect. But, it would be hard to disentangle that effect from independent actions by individuals, given the timing.
However, when the lockdowns were lifted, behavior of many became quite careless – per observations by myself and friends. Government messaging was way too weak at that point.
Now, we are seeing what is likely the result of that – our rate of case increase is the highest in the nation on a per-capita basis, and out percent of positive PCR tests has gone way up, with the inflection point in the curve corresponding roughly to the end of stay-at-home.
While increased testing and more targeted testing explain some of the numbers, they don’t explain the increases in COVID19 emergency visits, hospitalizations and ICU usage. Some is explained by imported cases – our border counties have really high rates recently. But, relaxed precautions by individuals clearly had an deleterious effect.
As a result of many people taking “end of lockdown” to mean “we won, let’s party,” the government has had to step in again, with mask mandates (this time, with enforcement), and restrictions on businesses.
Trying to sort this out with a sophisticated model would probably be folly. But, the trends, plus observed behavior, are significant.
Also, as long as people are hesitant to interact, economic impacts will continue. There is absolutely no path to “life as normal” except a vaccine, or herd immunity acquired at huge cost by not containing the epidemic.
meso –
> But, it would be hard to disentangle that effect from independent actions by individuals, given the timing.
Indeed. It’s amazing how many people overlook that aspect.
Along similar lines…
> However, when the lockdowns were lifted, behavior of many became quite careless – per observations by myself and friends.
What isn’t known, and will likely never be known, is whether there would have been a similar pattern of increasing carelessness absent government mandated shelter in place orders.
Nic:
We are indebted to you for your usual thorough and insightful analysis.
Thanks!
Nic,
Maybe I missed it, but do you have an estimate for how many lives the lock-downs saved from your analysis?
Pingback: America is Giving Up on Lockdowns | al fin next level
Michigan seems to have a downward trajectory in new cases/deaths, but when observing actual behavior, it’s pretty close to “we won” parties.
One has to wonder if the massive protests with very crowded conditions, shouting, and a large portion of the participants not wearing masks will have any perceptible impact on infection rates in those areas. If not, then maybe “we won” is closer to the truth than “second wave”.
The lockdowns were sold to us as a means of flattening the curve; no one ever said they would reduce infections or deaths at that time.So any claim that the lockdowns reduced deaths is a bald face admission that they were completely wrong or that any benefit was due to other causes.
https://judithcurry.com/2020/06/21/did-lockdowns-really-save-3-million-covid-19-deaths-as-flaxman-et-al-claim/#comment-919549
You are taking *one* reason for actions and assuming that it was the only reason. Another reason was to suppress the virus – drive it down to where it could be contained with standard methods – testing, tracing, isolation. We failed. New Zealand succeeded – they just had their first soccer match with no limits on audience. Taiwan succeeded. Japan – no lockdowns – is well on its way to success, as are South Korea and Australia.
New Zealand is living in terror of cases emerging and will be isolating itself from the rest of the world forever because they have been so shortsighted. The effect on their tourism and export dependent economy and quality of life is obvious. Sooner or later you are going to take the pain, which turns out not to be that great after all. Australia just said no tourists til at least 2021. Yep, really smart, farsighted policy to deal with a pathogen that causes serious illness in a very, very small portion of the population.
> The effect on their tourism and export dependent economy and quality of life is obvious.
The impact of the pandemic has been obvious everywhere.
We don’t yet know which countries will have which impacts long term as the result of which policy differences. Those who think they know are fooling themselves because we don’t know yet the future course of the virus And we don’t know what might have happened had things been different (i.e., had government responses been different). Counterfactual reasoning requires a high level of evidence because it is by definition entirely hypothetical. That includes your hypothesis about the long term impact on New Zealand of their COVID policy response. They may be in a much better situation to open up in the future than other countries, precisely because of the polices they’ve implemented thus far.
Ultimately, the people of New Zealand have the right to determine what they think is best for themselves. Now you may think you’re better able to judge than they what is good for them, but point of fact in New Zealand is that what they have done to react to the pandemic is the reflection of what the people of New Zealand think was the best approach. Would you have it any other way?
Right now, heading towards reelection, New Zealand’s PM has historic popularity ratings. Maybe in a few years the people in New Zealand will wish they had listened to people of your opinion. Or perhaps they’ll be glad they didn’t.
New Zealand at least has industry. Their counterpart in the US is Hawaii, which also “succeeded” by completely isolating itself. The result is unemployment that’s been variously reported as between 25-35% because the state is almost entirely dependent on tourism.
They’re stuck. Now what? Nobody on the island has been exposed, they can’t “re-open” without undoing everything they’ve done. Ironically I’ve read that one idea the state is floating is to allow only tourists from New Zealand and Australia.
In one respect the island nations and states had one idea researchers will be examining- their policy was nobody in … or out.
I think research will show the big mistakes in lockdowns were the weird effort to make them one-size fits all, and to ignore people who moved. We should have locked New York City residents inside New York City (they actually fled all over the country and took the virus with them). And we never should have prohibited lunch in Kansas.
Science means pandemics will be very unpleasant in big cities until they burn out. You can close all the restaurants you want in Georgia, but if you let hundreds of New Yorkers go to Atlanta the state will have an epidemic.
I doubt the “terror” in New Zealand is anything compared to that in countries where the virus is running rampant. Since they have such a low prevalence, any new infections will be contained – standard epidemic protocols will work.
Their tourism will suffer. But tourism world-wide is suffering, because people are avoiding COVID19 dangers, and governments are restricting travel. A number of states in the US now require 14 day quarantine just to enter them, for example.
You are grossly underestimating the impact of this disease, if measures are not taken. It has a gross case fatality rate of around 5% in the US, currently (it was higher earlier). It is far more deadly than seasonal influenza. It overwhelms hospitals when the infection prevalence is well below 5% of the population.
There’s a reason that governments world wide are taking measures to mitigate, or even better, contain the infection.
There’s no question this is lethal virus, and contagious. How lethal we don’t actually know. The number for “the US” isn’t consistent.
7.6% of confirmed cases in New York resulted in death.
3.13% in Florida
1.9% in Texas
Numbers are what’s reported in the Washington Post this morning.
And this is where the partisanship makes knowing even harder.
Politics demands that you cannot say Texas did a better job than New York at anything. In fact, politics demands the reverse- you must state New York handled the virus much more competently than Texas.
Which means- You cannot say the disparity is due to testing (ie they simply confirmed more cases than New York did), because that would imply Texas tested more than New York. You may not say there were any policies or hospital protocols that reduced fatalities, because that would mean New York did something wrong. And, in fact, you cannot acknowledge any difference between New York and Texas at all even though New York is a smaller state by population and had 13 times more people die than Texas. Why can’t you acknowledge this? Simple, Trump is the president of both states and since politics demands we ignore the fact that America is a federalist republic and assert the president is uniquely and singularly responsible for the results in both state, the results must be assumed to be the same.
Which means not only can we not study why the virus Texas was four times less lethal than New York, we’re absolutely forbidden to notice that the virus was twice as lethal in France than New York. Because obviously the EU is better at everything.
It is far too late in this pandemic to still be claiming an IFR of 5%. The CDCs best estimate is currently 0.26%. The IFR was inevitably going to be far less than the CFR based on identified cases, and this was pointed out months ago. A 0.26% IFR would be similar to the mid century flu pandemics. The 2009 pandemic had an early estimated fatality rate of between 0.1% and 5%, with one study putting it at 13.5%! It is now believed to be about 0.02%.
Reblogged this on Utopia, you are standing in it!.
> The lockdowns were sold to us as a means of flattening the curve; no one ever said they would reduce infections or deaths at that time
What?
If you flatten the curve, hospitals don’t become overwhelmed. You reduce shortfall of PPE in healthcare settings, you avoid inefficiencies resulting from overwhelmed caregivers, etc.
That’s why you want to flatten the curve.
Preprint – to be ingested with a grain of salt:
–snip–
Full genome sequences are increasingly used to track the geographic spread and transmission dynamics of viral pathogens. Here, with a focus on Israel, we sequenced 212 SARS-CoV-2 sequences and use them to perform a comprehensive analysis to trace the origins and spread of the virus. A phylogenetic analysis including thousands of globally sampled sequences allowed us to infer multiple independent introductions into Israel, followed by local transmission. Returning travelers from the U.S. contributed dramatically more to viral spread relative to their proportion in incoming infected travelers. Using phylodynamic analysis, we estimated that the basic reproduction number of the virus was initially around ~2.0-2.6, dropping by two-thirds following the implementation of social distancing measures. A comparison between reported and model-estimated case numbers indicated high levels of transmission heterogeneity in SARS-CoV-2 spread, with between 1-10% of infected individuals resulting in 80% of secondary infections. Overall, our findings underscore the ability of this virus to efficiently transmit between and within countries, as well as demonstrate the effectiveness of social distancing measures for reducing its spread.
–snip–
https://www.medrxiv.org/content/10.1101/2020.05.21.20104521v1
Still waiting for someone to show that the relatively limitated shelter in place orders in Sweden had a marked effect of limiting economic and/or other damages compared to Denmark, Finland, and Norway.
One would think that with all the cries of “draconian” we would have seen convincing evidence of such.
And of course, in such an analysis, accounting for the markedly fewer deaths and less illness in those other countries relative to Sweden, as an economic consideration, would necessarily need to be included to be comprehensive.
Imagine if, with all those many more deaths and much greater illness in Sweden, it turns out a vaccine is distributed on a relatively short time frame and countries can open up safely, and no one actually shows a marked economic advantage to a more-death-and-more-llness-more-rapidly policy, such as that followed in Sweden.
And then project such a greater impact in Sweden in mortality and morbidity (should a vaccine prevent reaching an equilibrium point over time whereby other similar countries reach a similar magnitude of spread) to scale in a population as large as the UK or the U.S.
What will those who recommended the Sweden approach say then, I wonder? I hope they’ll at least be forward in being thankful that no one followed their advice.
Sweden is a country of about 10 million people within the EU. It could no more fare significantly better than Michigan could have fared without a lockdown while the world around it is locked down. There inevitably will be some advantages to local business but to expect a huge difference in economic impacts is economically illiterate. The argument is that we would have all fared better had no widespread lockdowns occurred.
I wouldn’t rely on a vaccine to get us out of this. IMO this virus will run its course through society as others before it have for millennia. Revisiting periodically as others have for millennia. The severe flu season of 2017/2018 was caused by the genetic descendant of the H3N2 virus that caused the 1968 pandemic that killed over 100000 people in the USA.
Doug –
> The argument is that we would have all fared better had no widespread lockdowns occurred.
Two problems related to that. The first is that people are making that argument without evidence to back it up. They ASSUME that we would all have fared better absent government mandated shelter in place orders without that evidence. My point is that people should be careful about making such assumptions. Especially when their assumptions align with their political orientation.
The second problem is that actually people are making that argument as if they have evidence. They point to Sweden as if it proves their point about the benefits of not issuing government mandated shelter in place orders. Yet as you acknowledge, that argument is only speculative.
Some people have described the government mandated shelter in place orders as “draconian.” and that they have caused great economic hardship. Well, compared to what? If you dont know what the differential effect is, then you should just say that you don’t know, imo. And as I mentioned elsewhere, the problem is that different localities responded differently because the conditions they were responding to were different, and their ability to respond in various ways were different. Sweden was, in many ways, ideally situated to respond as it did. As such it isn’t a very good model for countries that faced very different conditions with very different resources to bring to bear.
> I wouldn’t rely on a vaccine to get us out of this. IMO this virus will run its course through society as others before it have for millennia. Revisiting periodically as others have for millennia. The severe flu season of 2017/2018 was caused by the genetic descendant of the H3N2 virus that caused the 1968 pandemic that killed over 100000 people in the USA.
I wouldn’t suggest “rely[ing] on a vaccine to get us out of this.” Not by a long shot. I think that it is much better to rely on practices such a comprehensive testing, robust contact tracing, and devoting resources for quarantining people who test positive.
But my point is that relying on a fast/wide spread policy is relying on speculation that a vaccine WON’T be made available on a widespread basis in the near future. The odds may well be in that direction. But it’s a gamble. If a vaccine does become available in a short time frame it is probably a losing bet. And I question whether it was a worthwhile bet in withe cass because it isn’t clear to me that there is any differential payoff.
The burden of proof is on the advocates of lockdown to show benefits exceed the very high costs. In fact, there is no evidence I’m aware of that lockdown actually saved net lies. Unemployment has a well proven correlation with early death and I found an economist saying (from memory) that at least 60K deaths will direcly result for a month long lockdown if unemployment reaches 30%. Delaying cancer screenings. Josh, you are the one who should show some evidence.
Swiss and German health authorities have done a careful analysis of the data showing that R was already reduced to almost 1 before lockdown.
> The burden of proof is on the advocates of lockdown to show benefits exceed the very high costs.
I always love “burden if proof” arguments because they’re usually such a good example of weak, self-sealing logic. Some people think that those advocating for shelter in place orders have the “burden of proof.” And others think that those advocating for “immunity herd” strategies have the burden of proof.
Who’s right?
If we only went with the logic of “cost” as you are asserting, then maybe you have the burden of proof to show that there is a greater differential cost from shelter in place orders? But I have yet to see anyone do so, (although it is often assumed or asserted absent proof) .
But the bottom line is that your assumption that you have the privilege of determining who has what burden (unless Trump has made some degree that I don’t know of) just doesn’t pan out in the real world. Many countries, including your own, have determined no such burden exists as you declare. They (including your president) have implemented (or supported) such orders without “proving” that they reduce or even justify costs (in part because you can’t prove the this costs exist, differentially). So in reality, no such burden exists.
But don’t let that get in the way of you assuming you have the privilege to assert it, if it works for you.
giving you the privilege of determining
Joshua: I always love “burden if proof” arguments because they’re usually such a good example of weak, self-sealing logic.
In a free society the burden of proof is on the government to show that depriving citizens of their rights is in their interest.
Matthew, you’re debating the point with a person of the political persuasion that believes the burden of proof is on you to prove how much of your own income you get to keep.
Unfortunately this is the result of generations of gradual shifting of the burden of proof from government to individuals. Forgive my grumpiness but after 40 years of boating, I recently had to prove to the state of Virginia that they should grant me a piece of paper allowing me permission to boat. The rationale for this is, of course, “safety” and the proponents are, of course, the same people who will happily tell you that people living near the riots should have no expectation that the state has any interest in their safety. But you might get the opportunity to prove to the state that you have a life or property worth defending.
Jeff –
> Matthew, you’re debating the point with a person of the political persuasion that believes the burden of proof is on you to prove how much of your own income you get to keep.
Anytime you want to know what I actually believe feel free to ask. Until that time, go ahead and indulge your fantasies about what I believe if that makes you feel better about yourself.
You been here what , joshie? Seems like at least a decade. We know all we need to know.
Joshua: Anytime you want to know what I actually believe feel free to ask. Until that time, go ahead and indulge your fantasies about what I believe if that makes you feel better about yourself.
You did not dispute his claim. That’s a “non-denial denial”.
Reblogged this on uwerolandgross.
Nic, thanks as always for your clear thoughts on this matter. I take a different path to the same conclusion.
There are a bunch of countries that did the worst. They are all grouped up at the top of this graph:
https://i2.wp.com/wattsupwiththat.com/wp-content/uploads/2020/06/corona-deaths-20200616.png
A few things worth noting. First, the different things those countries did … they didn’t make much difference. They’re all in a group with Sweden, who did very little.
Next, NO COUNTRY ON EARTH has had even 0.085% of their population die from this disease. Even countries that did nothing. Even developing-world countries. Nobody. That’s the worst we’ve seen. Well, except for New York, where Governor Cuomo ordered the nursing home to admit all covid cases transferred from the hospitals …
So I would consider 0.085% of the population dying to be a hard upper limit on what the disease does when you do nothing. No country to date has gotten there, and there is no sign that any country will get there after the virus subsides.
Now, the population of the 11 countries that were studied is 374,207,000 people.
And that would mean that if the 11 countries had done nothing, the maximum number of deaths from the virus would be about 320,000.
As of today, the 21st of July, those 11 countries have had ~ 160,000 deaths.
So IF the deaths stopped today, the maximum number saved MIGHT be ~ 160,000. But of course, the deaths haven’t stopped. In fact, the actions taken were NOT generally aimed at REDUCING deaths, merely at POSTPONING deaths.
So as you’d expect, although the countries hardest hit have seen a steep peak, their deaths quickly dropped near zero. On the other hand, the countries that locked down have seen their deaths extend out well beyond the peak.
Which means that there are still more deaths to come … and every one of them is a death that was NOT avoided by the lockdown.
How many more deaths? Hang on … OK, here’s my source:
https://covid19.healthdata.org/
Their best estimate is that those 11 countries will see another 12,000 deaths. Seems low, but if so, total deaths saved by everyone not getting max mortality is about 150,000.
Of course, that does not allow for the fact that different groups of people will end up with different death rates. But it does indicate that 150,000 is a much more believable estimate of deaths that MIGHT have been saved by the various measures … or not …
My best to you and your good lady,
w.
> A few things worth noting. First, the different things those countries did … they didn’t make much difference.
Note Willis includes no sensitivity analysis of the different interventions included.
All you need to kniw about Willis’ analysis there is that at the time he did it he pointed to similar death rates in Switzerland and Sweden as evidence that lockdowns did nothing. This he did even while ignoring important factors such as a much closer proximity to Lombardy in the case of Switzerland and a higher prevalence of single-person households in Sweden.
But he has ignored the dramatic divergence in death rates subsequently. Indeed, now in Sweden it is more than double the per capita rate of deaths (500 per million) than it is in Switzeand (225 per million).
And you can also consider that along with that posting that analysis, he attributed lower death rates to mask-wearing in some countries… only to later put up a post where he changed his mind – with the explanation that he actually looked at the evidence and didn’t think it showed a benefit to mask-wearing.
Imagine reaching a conclusion in the first place and putting up a post without looking at the evidence!
(That said, he hadn’t even done a thorough examination of the evidence with his 2nd post).
Also notice which countries Willis leaves off his chart: the majority of those with drastically lower death rates.
He picks a group of countries with similar death rates, leaves off most of those with much lower death rates, and then notes that those with similar death rates have similar death rates (except those that don’t, of course like Germany and South Korea).
Joshua, when you are looking for what the worst case might be, you do NOT look at places like South Korea and Germany.
You look at the top of the list, the countries with the worst death rates, duh. And no, I didn’t pick “countries with similar death rates”. I simply noted that the worst off countries are in a fairly tight group.
But then, you knew all of that.
You really should stop posting about my my claims and statements, Joshua—your rampant hatred of me warps your mental abilities so badly that it makes your comments not just laughable, but totally ludicrous …
w.
Willis –
> Joshua, when you are looking for what the worst case might be, you do NOT look at places like South Korea and Germany.
When you’re comparing cause and effect of different policies (and conditions) in different countries, if makes no sense to just consider those groups by similar effects.
This problem is wll illustrated by you the comparison of Sweden to Switzerland. You argued that the similar outcomes in those countries indicated that the different policies had no differ risk effect. Except you just took a cross-sectional snapshot. If you compare the situations longitudinally, you will see dramatically contesting outcomea.
Now personally, I don’t think that comparing across counties where the conditions are so different yields much insight – but if you’re going to do it you should do it properly: Control for confounding variables, do a sensitivity analysis, use longitudinal data, consider why starting conditions (which have a signal in outcomes) help to explain why different countries took different approaches to begin with.
> your rampant hatred of me warps your mental abilities so badly that it makes your comments not just laughable, but totally ludicrous …
I don’t hate you Willis. I don’t even know you. I don’t hate peope I don’t know. But you shouldn’t conflate criticism of what you post online – when it is poorly done – with hatred. It’s just criticism of your work. Nothing more.
Should be…
This problem is well illustrated by your comparison of Sweden to Switzerland. You argued that the similar outcomes in those countries indicated that the different policies had no differential effect. Except you just took a cross-sectional snapshot. If you compare the situations longitudinally, you will see dramatically contrasting outcomes.
“Except you just took a cross-sectional snapshot. If you compare the situations longitudinally, you will see dramatically contrasting outcomes.” Please explain.
George –
> Please explain.
I don’t like relying on copies definitions in blog comments because they often miss important context. But in this case I think they might do. From Wikipedia:
–snip–
Cross-sectional data, or a cross section of a study population, in statistics and econometrics is a type of data collected by observing many subjects (such as individuals, firms, countries, or regions) at the one point or period of time. The analysis might also have no regard to differences in time.
Longitudinal data, sometimes referred to as panel data, track the same sample at different points in time. The sample can consist of individuals, households, establishments, and so on. In contrast, repeated cross-sectional data, which also provides long-term data, gives the same survey to different samples over time.
-snip–
Seems to me that it is very important when examining for cause and effect in terms of the outcome of interventions to address the pandemic to look at changes and trends over time. In this case, Switzerland faced an initial situation of a high rate of infection, I assume largely due to proximity to Lombardy and probably travel to/from China in comparison to Sweden. Thus, initially, the government interventions were dealing with worse conditions and you can’t judge their impact until they’ve had time to mitigate the starting conditions. Sweden also has other factors that affected outcome differently than Switzerland, such as a higher % of people who could work from home, who live in single-person households, etc. So if you used a snapshot of data when the pandemic first got under way, you get one picture. But if you track what happens over time a different picture develops.
There are factors that run in the other direction over time as well. For example, the robustness of the testing paradigm. Or apparently in Sweden they’ve taken a casual attitude about treating older people (Iow, they aren’t providing oxygen therapy in nursing homes).
Joshua: Note Willis includes no sensitivity analysis of the different interventions included.
…
Imagine reaching a conclusion in the first place and putting up a post without looking at the evidence!
You totally missed the point of Willis Eschenbach’s elegant analysis. Instead of trying to compute, as Flaxman et al did so inelegantly (as shown by careful reading of their manuscript and Nic Lewis’ critique), the exact effects of the lockdown orders in 11 more or less comparable countries, Willis estimated an upper bound on the maximum number of lives that might possibly have been saved by the lockdown orders. The number of lives saved by the lockdowns can’t have been greater than the total number of lives that would have been lost without the lockdowns. Working with the evidence available to date, he shows that the maximum number of lives that might have been saved by the lockdowns is not very great.
The disparaging tone of your posts is not supported by any thought or calculation. Do you have an actual criticism that might be clearly expressed?
J:
“…but if you’re going to do it you should do it properly: Control for confounding variables, do a sensitivity analysis, use longitudinal data, consider why starting conditions (which have a signal in outcomes) help to explain why different countries took different approaches to begin with.”
Today’s word: Resolution.
You’re turning it up too high and losing sight of what’s important. What is failure? We can see it. How is failure prevented? It’s not. There is failure with X and without X. X is not the control variable.
Don’t study success. We are trying to prevent failure.
The more steps you add, the less you can say. We are swirling in unknowns. Keep it simple.
Matthew –
> […] Willis estimated an upper bound on the maximum number of lives that might possibly have been saved by the lockdown orders.
How so? By looking at some cherry-picked selection of countries that did issue orders, without actually defining what a “lockdown” comprises, and a cherry picked country that didn’t? Obviously, there is a possibility that the countries that issued orders reduced the number of lives lost by doing so. The countries that didn’t issue orders didn’t issue orders because (1) they had less of a pressing need to do so and (2) they had more resources to bring to bear to help prevent loss of lives than many other countries. We’ve been over the basic logic problem there. Comparing countries that issued orders with those that didn’t is inherently problematic because the starting conditions were different. Let’s look at again at what Willis wrote:
–snip–
Next, NO COUNTRY ON EARTH has had even 0.085% of their population die from this disease. Even countries that did nothing. Even developing-world countries. Nobody. That’s the worst we’ve seen. Well, except for New York, where Governor Cuomo ordered the nursing home to admit all covid cases transferred from the hospitals
–snip–
Elegant? Really?
Take Lombardy – the death rate there was on the order of 0.16% – and that was in a developed country, where strict shelter in place orders were issued. Do you really think that absent those orders, not more deaths would have occurred?
Bergamo – a fairly wealthy town in Lombardy? The most recent data I could find with numbers showed a death rate of 1.8%. Do you really think that absent those orders, not more deaths would have occurred?
NYC – close to 0.3%. With shelter in place orders and incredible resources to bring to bear.
And Willis says that the most deaths that could possibly occur is 0.085%.
Now of course, those regions could suffer a higher fatality rate than an entire country would. Infection mortality rates differ by a lot of factors – primarily age stratification. But at some point infection fatality rates tell you about the virulence of the disease. At some point, it tells you on average how many people who get infected, die. Most estimates have it at between 1.0 and 0.5%, but even if you go with the lower estimates, say 0.3%, clearly, Willis is wrong. In Bergamo, seroprevalence studies show @ 60% of the population was infected. If we assume that’s accurate, with an infection fatality rate of 0.3%, that’s more than double Willis’ putative worst case scenario – in a fairly wealthy town with fairly good healthcare resources. Even if we were to assume that prevalence would top out at @30%, with an infection mortality rate of 0.3%, Willis’ number is low – in a fairly wealthy country with reasonably good healthcare resources (compared to developing countries).
Here, check this out:
https://www.nature.com/articles/d41586-020-01738-2
> The number of lives saved by the lockdowns can’t have been greater than the total number of lives that would have been lost without the lockdowns.
If you’re using the number who died without shelter in place orders by looking at a place like Sweden, a wealthy country, with many people who live in single family households, and great healthcare systems, and many people with the ability to work from home, and relatively sparse population density, where they did issue some shelter in place-type orders (closing some schools. limiting crowd sizes), to say that’s the worst that could possibly happen anywhere…well, I guess there’s really nothing that I could say to dissuade you because it’s obviously nonsense and you apparently are willing to accept nonsense.
Willis never did a sensitivity analysis to determine whether some interventions might have been more explanatory than others. Major components of “lockdowns” were closing schools and limiting crowd sizes. These are the sorts of things that happened in Sweden. Ordering people to work at home was another part – something that many people in Sweden did without the orders being necessary because they had the infrastructure to allow for that. Would that have happened in other places like the U.S.? Would so many people have worked from home, or not gone into work without the government mandated shelter in place orders? Would that have resulted in more deaths and disease? Do you have any idea?
Ragnaar –
> We are swirling in unknowns. Keep it simple.
Well, I do agree with that. Lot’s o’ unknowns, and it’s too early to even get a firm grasp on many of them.
“Except you just took a cross-sectional snapshot. If you compare the situations longitudinally, you will see dramatically contrasting outcomes.”
Thank you for the definition of snapshots. What contrasting outcomes will we see? And when? Isn’t it too early for a longitudinal snapshot – unless you extend it with a crystal ball? Please explain.
Willis pointed to Sweden and Switzerland a while back and saw similar death rates and said it showed interventions don’t work. No difference in association with the interventions.
He didn’t take into consideration possible confounding variables.
Now Sweden death rate is double that of Switzerland. What do you infer, about the effects of interventions on outcomes, from the dramatic difference in changes over time?
What does that say about his comparison from months ago, other than it was facile?
Or to you infer that the interventions work?
Joshua, thanks. I did not understand that you were disagreeing with something that Willis wrote elsewhere. Should today’s total death rates be considered lateral or longitudinal?
George –
> Should today’s total death rates be considered lateral or longitudinal?
Aye. That’s the crux of the biscuit.
Ideally, we could be very careful of cross-country comparisons. For example, in Sweden it seems they might not be providing oxygen therapy to infected people in nursing homes. How much of their increased death rate is explained by that policy? We don’t know.
But if we have enough evidence to control for confounding variables – such as thst one – even then we should be waiting.
What we don’t know about a “herd immunity” approach is whether in the long run the same % of the populafkon will get infected either way. That information is crucial for evaluating the comparative merits of a fast/deep spread versus a slow/shallow spread. If the sme number gets infected either way the comparison looks very different than if a significantly higher % get infected in Sweden over time, after controlling for confounding variables. And we necessarily need to wait to know if that will be the case. Any fair analysis might pivot diametrically depending on that outcome.
So we can learn from Willis’ mistake of jumping to a conclusion based on cross-sectional rather than longitudinal evidence. We can learn to be more circumspect, and to have more respect for uncertainty.
I will note, that even at that time Willis might have looked at the trends in the data over time, rather than drawing conclusions from a snapshot. But he didn’t take the trends into account when he drew his conclusion.
The same rule applies now. We have limited data on trends compared to what we will have in the future. But with some humility we can acknowledge that reality even as we consider the tends at this point. The worst thing to do from an analytical perspective is to consider the data from this point in time only. Data from this point in time has relevance, but it needs to be grounded in the context of change over time.
Joshua: Elegant? Really?
Yes, especially by contrast with the Flaxman et al mess. I think that point-by-point comparisons of critiques, yours vs Willis, Nic vs Flaxman et al clearly show that Willis achieved a more reliable result.
Little nitpicking joshie accuses Willis of cherry picking. Then joshie picks Lombardy and NYC the biggest Wuhan virus cherries in the world. Actually, they are the same banana.
Lombardy was overwhelmed with disease from the round trip direct flights to and from Wuhan of tens of thousands of imported Chinese workers in Lombardy garment industry sweat shops. After the infection came to town, local officials told their citizens to hug a Chinese person. Lombardy spreads the Wuhan virus to the rest of Europe and to NYC.
NYC officials say to their folks don’t worry go out and have fun. Celebrate at the Chinese New Year’s parade and show the rest of the xenophobic country how woke you are. And don’t worry about the old folks, they are safely tucked away in nursing homes. Hey, if anything goes wrong we will pin it on Trump and the NYPD.
Don –
> NYC officials say to their folks don’t worry go out and have fun. Celebrate at the Chinese New Year’s parade and show the rest of the xenophobic country how woke you are.
People’s behavior changed prior the issuance of the SIP orders. Public officials in NYC encouraging public interaction at that point was a mistake, but you don’t have any data on what the actual impact was, just as you don’t hand any data in how much disease was spread as the result of republican politicians encouraging people to go out to bars and restaurants then, and even later in the pandemic. Arguably, NYC officials’ bigger mistake was the lateness of closing schools and the lateness of issuing SIP orders, but we don’t really have hard data on that either and it’s intersting to note that many here are arguing that such interventions have no effect and/or are “TYRANNY!!11!!11!!!!
> And don’t worry about the old folks, they are safely tucked away in nursing homes.
Do you happen to know where NY ranks among states in % of covid deaths that occurred in nursing homes? If you don’t know, give me your guess. I’m curious to know what you think the ranking is?
And what would you have done with infected people who had been in hospital and recovered, and ready to be released (perhaps some of not most who were no longer contagious)? Where would you have suggested putting them? Or maybe you think they should have just remained in the hard-pressed hospitals that were dealing with a pandemic, and which had a lack of PPE (thanks Donald)? And why was sending them to nursing homes the federal guideline?
>Lombardy was overwhelmed with disease from the round trip direct flights to and from Wuhan of tens of thousands of imported Chinese workers in Lombardy garment industry sweat shops. After the infection came to town, local officials told their citizens to hug a Chinese person. Lombardy spreads the Wuhan virus to the rest of Europe and to NYC.
I addressed above that Lombardy and NYC were obviously hit particularly hard – but at some point the overall infection rate has a general limit absent interventions (and some argue regardless of interventions) and the infection fatality rate has some average level (even if that level isn’t very instructive relative to how to react to the pandemic because policies should be differentiated by specific populations). If we half the infection rate in Lombardy and go with a much lower infection fatality rate than what occurred in Lombardy then we will have a much higher death rate than Willis’ 0.085%. Same with NYC.
Where is your sensitivity analysis, joshie? Why did you ignore the divergence in death rates subsequent to the subsequent. blah blah blah What about the longitudinal thang? Cross sectional snapshot this…..
It’s blatantly obvious that you are a silly cherry picking nitpicker.
Whatever certain people say, you say the opposite. It’s just mindless attention seeking agitprop. Judith won’t allow you to incessantly yap at her heals, as you have done in the past. So, you yap at Judith’s deplorable band of skeptics. It’s really tedious.
Don –
I take it you don’t know where NY ranks in states in terms of % of Covid deaths in nursing homes, and that’s the reason why you didn’t answer?
We won’t know how many old folks died in nursing homes under the regime of the capo di tuti Cuomo and his underboss De Blasio, until there is a credible investigation:
https://nypost.com/2020/06/20/cuomo-takes-a-bow-after-his-deadly-nursing-home-decision-goodwin/
“Officially, New York says the coronavirus claimed 6,200 lives in nursing homes, or about 25 percent of the state total of nearly 25,000 fatalities, but the actual total is certainly higher. Some estimate that nursing home deaths are closer to 12,000.”
Is 6,200 enough for you, joshie? How many states had total deaths from the virus lower than that? I’ll give you a little help:
CA 5,565
TX 2,222
FLA 3,256
Those are the most populous states ahead of NY. The three don’t have total deaths that add up to 12,000. Don’t you feel just a little bit foolish, joshie? You would, if you were just a little bit self-aware. Carry on with your nitpicking and distorting of facts and logic, little lefty. We know the routine.
Question asked twice, Don. And you still haven’t answered.
Methinks there’s prolly a reason.
But do keep up with your cheering for old people in NY to die so you can hate on Cuomo.
And why was that the federal recommendation, after all? And what were you recommending, when decisions had to be made, what to do with older people who were being dismissed from crowded hospitals? Was it your recommendation to just keep them in hospitals, where they could be added to the list of people that hospital staff had to care for?
Please document where you made your recommendations at the time. Or where anyone objected to the federal policy at the time.
Or are you just playing Monday morning QB after the fact to score cheap political points by exploiting the deaths of old people?
That couldn’t be it, now could it?
Why cherry pick %, joshie? The real minimum number of dead old folks not meaningful to you?
The feds did not tell any state to willy nilly force nursing homes to accept corona virus infected old folks. That’s a fake story the left loons made up to cover for negligent homicide. You are not fooling anybody here.
“Arguably, NYC officials’ bigger mistake was the lateness of closing schools and the lateness of issuing SIP orders, but we don’t really have hard data on that either and it’s intersting to note that many here are arguing that such interventions have no effect and/or are “TYRANNY!!11!!11!!!!
The net effect of the shelter in place order was that any NYC resident with money, a second home, or parents with a spare bedroom fled the city and took the virus with them. Even the NYC dwelling CNN anchors were congratulating themselves for sheltering in place, but not, you know, in the city they live in.
This was when the media and ACLU were attacking the awful government of Rhode Island, who said they didn’t want infected New Yorkers to flood their state and kill everyone and dared- dared! – to stop NY cars and tell them about self-quarantine orders. Little Rhode Island was ordered to welcome New Yorkers, drop the warnings, and is currently the fifth worst state in terms of deaths per capita.
“TYRANNY!!!!” is apparently momentarily inconveniencing a blue state resident with good reason- a New Yorker under a shelter in place order that they weren’t following.
But never use the word “TYRANNY!!!” to refer to ordering the permanent bankruptcy of a restaurant in a town with no cases.
Don –
Since you’re cheering for old people to die, you should move on to other states since the majority of them have a higher % dying in nursing homes.
Now back to my other points.
What were you recommending at the time as the alternative policy? Please document.
Should they have just kept them in the hospitals? Please answer.
Why was it the federal guideline?
https://www.cms.gov/files/document/3-13-2020-nursing-home-guidance-covid-19.pdf
Look – I don’t think it was the best policy. In hindsight, it was a mistake. People made mistakes during this process, and they resulted in deaths. But that isn’t a partisan feature. Trump and his administration made tons o’ mistakes. But it was a difficult situation. You had concerns about hospitals being overtaxed, and healthcare workers running out of equipment and you had people who had recovered and the idea of just keeping them in hospitals wasn’t a very viable option either. Ideally they should have provided them housing in some other kind of skilled facility – but that’s easy to say in hindsight and it would have been quite costly.
But that isn’t the reason why the death count in NY was so high. And many other states had similar policies. And the % of deaths in nursing homes was lower in NY than in the majority of states.
So why don’t you just stop cheering for people to die in NYC nursing homes so that you can score cheap political points?
You will never engage in honest discussion, joshie. You are an agitator. Spoiler. Pain in the …. Everybody knows it. You are just digging your credibility hole deeper.
Sorry you feel that way, Don.
I think that people on all sides have made mistakes, but that it would be impossible to not make mistakes.
I think that people on all sides have made mistakes and avoided accountability. – and for me that is the most important element. Accountability is key, IMO.
And Jeff is wrong about his belief about what I believe.
Joshua,
“Bergamo – a fairly wealthy town in Lombardy? The most recent data I could find with numbers showed a death rate of 1.8%. ”
If the information you are citing came from the InTwig survey, then you are confusing an estimate of IFR with population mortality rate.
Jon –
It was based on thr raw numbers of deaths that I divided by population. The number of deaths I got from this Guardian article which gives an official number from April :
https://www.google.com/amp/s/amp.theguardian.com/world/commentisfree/2020/apr/06/coronavirus-bergamo-work-ethic-lockdown
But now I see that the the Guardian article referenced another article (which in turn referenced In Twig) that has @double the deaths I used for the calculation – but indicates that the number was for all the provinces of Bergamo with a much higher population, while I calculated with the official number thinking it was for the city only.
https://www.ecodibergamo.it/stories/bergamo-citta/coronavirus-the-real-death-tool-4500-victims-in-one-month-in-the-province-of_1347414_11/
Using that lower official number for that larger population nets a much lower rate of 0.37%, while a rate calculated with the InTwig investigation numbers for that larger calculation working out to 0.74%.
But keep in mind, that was from numbers back in April. Surely the number of deaths is higher now. Also interesting that the Ecodibergamo article reports population infection rates above 45% in some areas – again, back all the way to April.
Jon –
Here’s an estimate based on excess mortality of a 0.58% population fatality rate for Bergamo as of May 9.
https://www.medrxiv.org/content/10.1101/2020.04.15.20067074v3
Note, that study estimates an infection rate for Bergamo of 69%, which is just a tad higher than when Nic estimates the herd immunity threshold kicks in 😉
Yikes. Correction for NYC also: should be 0.094%
“And Jeff is wrong about his belief about what I believe.”
This under a string where you suggested only wingnuts (I paraphrase) screech “tyranny!” at lockdown orders that caused real economic damage.
And I gave you an example where the ACLU and New York Times screamed “tyranny!” at a state’s order that really could have saved lives for what really was a momentary minor inconvenience. But they didn’t like it because it implied the truth- New Yorkers “sheltered in place” only on paper, in reality they did no such thing.
This is what educated people call an irrational double-standard – in our media culture anyone who objects to damaging edicts that have flimsy backing is partisan and wrong, yet anyone who objects to rational edicts that inconvenience the wrong tribe is a patriot.
Double standards do real damage to credibility, damage that is compounded when they are celebrated, laughably justified, or worse: denied even though they are transparently obvious.
My Facebook feed this morning greeted me with the assertion (from a Yale-educated man no less) that all of Europe wears mandatory masks and nobody in the US does because Trump. Which isn’t true of course as anyone with access to Google can see- according to Statista about 2% of the populations of Denmark, Finland, Norway, Sweden wear masks and over 80% never do. Just a few months ago, these nations were the gold-standard of US progressives.
Here’s how damaging that is- I now know not to trust anything the guy posts, knowing him well I assume he just embarrassed himself by trusting a news source that used to be reliable. But people who don’t know him well may assume a Yale educated man knows exactly what the truth is about something he feels strongly enough about to share and he is being deliberately polarizing. He just trained people to ignore him at best, be angry with him at worst. All without any help from Fox News.
Joshua: But do keep up with your cheering for old people in NY to die so you can hate on Cuomo.
You believe that Don Monfort is cheering for old people in NY to die?
And that absolves Cuomo of responsibility for a truly bad decision?
What do you believe about these assertions?
Matthew –
> You believe that Don Monfort is cheering for old people in NY to die?
No, of course not. I’m just saying that on the off chance that in realizing how absurd it is when I say that about him, Don might realize how absurd it is when he constantly demonizes anyone who criticizes Trump’s handling of the pandemic as “cheering for the virus” or whatever nonsensical way that he describes it.
> And that absolves Cuomo of responsibility for a truly bad decision?
??? No, nothing absolves Cuomo for the mistakes that he made. A mistake is a mistake. However, some mistakes are understandable. It doesn’t absolve them if it is understandable. For example, Trump has made many, many mistakes. Most of them are understandable. That doesn’t absolve him from making those mistakes. Others, however, aren’t understandable – such as when he lies over and over about the testing.
> What do you believe about these assertions?
??? I described that above.
Willis Eschenbach,
O/T:
Can you give me the sources for the USA and Australia data points you included in this chart?
https://4k4oijnpiu3l4c3h-zippykid.netdna-ssl.com/wp-content/uploads/2015/08/rstudioscreensnapz027.png
Source: Eschenbach, 2015, https://wattsupwiththat.com/2015/08/03/obama-may-finally-succeed/
Also, is there more recent data for Europe, US and Australia than you used for the chart?
Source for the data for Europe data: https://strom-report.de/strompreise/#strompreise-europa
Try link to chart again:
https://wattsupwiththat.files.wordpress.com/2015/08/rstudioscreensnapz027.png
Otherwise, see Figure 1 here: https://wattsupwiththat.com/2015/08/03/obama-may-finally-succeed/
If you want the worst case, you have to do two things: wait until the epidemic ends or is contained, and look at countries where there were no or minimal interventions – i.e. third or fourth world countries where the population has no access to masks, or even sanitation, and cannot isolate well due to poverty. Then you need to look at the total death rate and adjust, since they will not be able to test most who die.
Every country in your graph fails to meet the second criterion. All of them had substantial NPI’s, especially voluntarily by individuals, plus they have good sanitation and good economies.
Right now, Arizona is experiencing its worst case rises in counties bordering Mexico. Do we have good case or death rates from Mexico? Of course not.
The Navajo Nation, also mostly in Arizona, had our previous highest rates – far higher than any county in the state. It also is very poor, resulting in 30% of homes with no running water, and extended families living in the same quarters. And, until the problem was recognized, had no masks and no interventions. Fortunately, that has changed, so the total fatality rate for the population is relatively low. It currently has a .2% death rate of the entire population, and that is with access to good hospitals for critical care.
Navajo Nation numbers are not included in Arizona totals.
I should add… the epidemic is ongoing on the Navajo reservation. The tribe has been enforcing very strict measures – weekend long curfews, roadblocks, etc.
Willis Eschenbach: I take a different path to the same conclusion.
Good job. Thank you.
Willis, thank you for your comment and informative graph.
I agree that Flaxman et al’s estimate of 3.2 million deaths in the 11 European countries in the absence of government interventions is way too high, and the convergence in your graph of lines approaching the 0.085% of population deaths level is certainly suggestive of somewhere near that being a possible upper limit.
My best to you and your ex-fiancee,
N
> the 0.085% of population deaths level is certainly suggestive of somewhere near that being a possible upper limit.
So Belgium should just let it rip, since there pretty much there already. No more social distancing. Don’t need to worry about PPE. Who needs to wash hands or wear masks? Sell the ventilators. Open up international travel. Play soccer games in front of full stadia. About 150 more deaths and they’re done. No need to worry about a 2nd wave for Belgium.
@33% infected HIT and 0.255% IFR are about your limit. Of course the HIT could be a bit higher or lower and the IFR shift accordingly.
At any rate, you’re on record.
Joshua: So Belgium should just let it rip, since there pretty much there already.
Nothing supports that conclusion. How do you arrive at it?
Matthew –
> Nothing supports that conclusion. How do you arrive at it?
?
I multiplied the population of Belgium by 0.085%: Mr. Excel said 9850. I looked up the # of deaths in Belgium: Mrs. Worldometers said 9713.
You’re smarter than I am. What am I missing?
Joshua: What am I missing?
1. What you said: So Belgium should just let it rip, since there pretty much there already.
2. what I said: Nothing supports that conclusion. How do you arrive at it?
There is no implication of what Belgium ought to do.
I multiplied the population of Belgium by 0.085%: Mr. Excel said 9850. I looked up the # of deaths in Belgium: Mrs. Worldometers said 9713.
So what? Perhaps Willis Eshenbach’s estimate of maximum fatality rate is a little low? Or estimates based on a few small areas are too unreliable to serve as global estimates?
Matthew –
I was being sarcastic because 0.085% as some kind of upper limit is obviously wrong.
Joshua: I was being sarcastic because 0.085% as some kind of upper limit is obviously wrong.
No it isn’t. Are you choking on “suggestive of somewhere near that being a possible upper”?
Willis,
I don’t think you can use national statistics to estimate how bad fatality rates could be without social intervention. You need a much more granular analysis.
In the absence of proven therapies, Population Mortality Rate (PMR) is a strong function of:-
(a) Age and health stats
(b) The degree of social connectedness of the population in the context of viral transmission.
Social intervention can only influence (b), obviously.
Intuitively, we would expect that the degree of social connectedness in cities should be much higher than in rural environments, and this is generally borne out by the available information.
The closest analogue we have to COVID 19 hitting an unprepared population in a mixed urban and rural setting is the Lombardy region in Italy. Some cities there had population mortality rates in excess of 0.4% (Codogno, Bergamo). The PFR for the region as a whole is 0.16%, similar to New York state. This value probably represents our best indicator of how bad it can get without social intervention.
In contrast, the fatality statistics for Italy as a whole (at just 0.06% PMR) are considerably diluted by the absence of high fatality rates in regions outside Lombardy. Why was that? I would particularly draw your attention to two populous regions – Lanzio and Campania, which contain Rome and Naples, respectively, with many similarities to Milan in the Lombardy region. By the time community transmission became important in these regions, fear induced by the dire events in Lombardy had already led to the implementation of draconian social measures. The PMR values for these two regions are 0.012% and 0.007% – one to two orders of magnitude better than the PMR for Lombardy.
So, while I would agree that the Flaxman paper is pessimistic on PMR with no intervention, (no doubt to defend IC-inspired policy measures retrospectively), I do not believe that we should take too much comfort from national statistics, since in all instances death rates have been influenced by social intervention to a greater or lesser extent. My best estimate for any western nation would start with the Lombardy PMR value of 0.16%, followed by some attempt to normalise for age and health and the ratio of urban to rural population.
kribaez –
Sorry for calling you Jon below in this comment:
https://judithcurry.com/2020/06/21/did-lockdowns-really-save-3-million-covid-19-deaths-as-flaxman-et-al-claim/#comment-919738
Where did you get the population fatality rates you listed?
Joshua,
For deaths by Italian region:-
https://www.statista.com/statistics/1099389/coronavirus-deaths-by-region-in-italy/
“But for Sweden a value of 1.27, in the far tail of the resulting distribution, was inferred.”
Maybe I missed it but, inferred based on what?
The Flaxman hierarchical Bayesian model puts prior probability distributions on many uncertain variables, not just on the alpha values representing the effects of interventions. Some of those prior distributions, including that for country-specific effects on transmission, also in turn have prior distributions put on their standard deviation.
The model then infers posterior probability distributions for all the parameters that have been given prior distributions, not just the alpha parameters, based on how well the resulting simulated death profiles for each country match the recorded deaths.
So the short answer to your question is: based on recorded deaths.
Thank you
That is rather circular logic on their part.
It’s not strictly circular, but IMO they are trying to infer the values of too many unknown variables with just one set of (flawed) death data.
The Flaxman paper seems to be a classic case of the invention of values for a multivariate analysis where even the main variables cannot be precisely defined, let alone quantified. Thank you, Nic, for your analysis that shows this.
One much-discussed variable is the face mask. Values placed on its effects are inventions because there has been no adequate experimental study. Further, no adequate values can be assumed for lockdown variables, because they different from country to country, region to region down to family to family. You simply cannot do a meaningful analysis with invented numbers in place of measured variables. This is so basic a guideline. – why do studies like this Flaxman et al work make it past peer review? Geoff S
why do studies like this Flaxman et al work make it past peer review?
If results of a study will scare people, it will pass.
If results of a study will not scare people, it will not pass.
This does not always work, but most often, it does.
“why do studies like this Flaxman et al work make it past peer review?”
Because Peer Review is part of the same fallacy of competence as the rest of the setup.
The analogy I use is this. If you wrote a paper about Papal Infallibility and put it up in a well respected journal that distributed it reviews only to tenured Catholic Priests at well respected Catholic Churches, what would you expect the outcome of the review to be.
Peer review has become part of the Groupthink as the toleration of differences of opinion has decreased in institutions.
Nicholas Lewis,
Thank you for this analytical essay.
That is quite long an complicated. But say the US lockdown {by itself}
saved 2 to 4 million lives.
And I would say China’s failure to control the China virus, has cost the world at least 1/2 million lives. And far more damage than 1/2 million deaths.
In same way China cost the world in terms of deaths and in other significant ways, harm to to entire world. US saved the world from more damage that China government could have caused.
The primary way US lockdown saved the world, was it stopped further air travel from China. Which by simply stopping air travel to US it saved the rest of world. Or if it didn’t one could blame the US for causing to stop it’s air travel from China.
Or in simple terms, World Health Organization committed war crimes, and US could have been more of a party involved in WHO’s war crimes.
So not only did the world’s superpower stop air travel from China, it by it’s action, encouraged other nation to stop air travel from China. And US was not a second party to China spreading it’s virus from US to rest of the World.
Now, we always wish, that US stopped air travel from China, quicker than it did. Second guessing after the fact is easy and it’s something idiots do.
So saying US should have stopped air travel sooner, when it faced hostility for doing as soon as it did {btw resistance and hostility from WHO, among others parties}. So, if want to be fool and blame US for not stopping air travel sooner, I suggest that such fool should blame it all on WHO.
As far as other types of lockdown measures within US, it seems some lockdowns may have did more harm than good. But all lockdown measures
are largely about “buying time”. If China had not released this virus upon the world, it would have bought the World, a lot time. Likewise cutting off air travel {though late in game] also bought the world more time. And most time it bought was for countries such countries in South America, which months later, now in middle of hardest part of effects of China virus. And I believe India and Africa have yet to reach same point as South America has.
Buying time, gave us some time to make mistakes, and learn from them.
At this point in time, we make a lot tests, and having access to abundant amounts test, is one best ways to control this sneaky virus.
Of course the common focus of lockdown is of “total lockdown variety” and this type of measure, is like last resort type move. And would claim it’s not choice to make. Or this virus would cause a lockdown whether or not a government orders it or not. BUT it’s better for government get ahead it, rather than allowing virus to do it. I would say with China, it was mostly the virus which caused the China lockdown rather than police state actions.
I would also say New York State waited slightly too long to order a lockdown.
And York State’s delay, caused a mini China effect. Or New York City acted like huge bomb. And China was the biggest bomb.
So at moment New York State followed by New Jersey state have highest recorded death per million of anything else in the World.
One find reason for it. You also claim it was different strain of virus and etc and etc. But I would say if New York State had waited 2 more days, before locking down, it would have been much worse.
And same “rule” applies to shutting down air travel, if delayed from several more days or week, US, and globally would have had more more deaths- and we not done yet. And we still have more time, to prevent further death.
But in terms of US and Europe “total lockdown variety” that phase is over, we already blunted the spread from China. And don’t see any other country needing to do such kinds of “total lockdown”. But trace it, and isolated, yeah sure, do that kind of lockdown. Any nation can actually successful do this, now.
Nic,
You and Judith Curry have written here about several different models. Models have enormous impact on public policy, as seen by the Ferguson model of COVID 19, yet few in the media or in the public understand what models really are and how they work. I would like you to consider making another post here commenting not on a specific model but on models in general. Here is my layman’s view of it.
You wrote: “It seems likely that the inferred relative strengths of the various NPIs are also highly sensitive to other assumptions made by Flaxman et al., and to structural features of their model.”
From what I as a non-mathematician have been able to infer from your writings, some general observations might be:
1. All models include assumptions by the modeler about data that is either unavailable or uncertain, and relationships between such data;
2. The output of the model is determined, in part, by these assumptions;
3. The structural features of the model are also chosen by the modeller, with full knowledge of their impact on the output;
4. The assumptions and the structure of the model together determine the output;
4. The output is therefore fully controlled by the modeller;
5. The purpose of the model is not simply to inform the viewer but, Given the modeler’s control over the output, to persuade the viewer to adopt the unstated policy response preferred by the modeller.
Your thoughts?
Andrew,
Thanks for your comment. I will consider what you say. But I think that I find discussing models and their shortcomings easier in the context of specific cases. I would have liked to include more discussion the Flaxman model, but I felt that my article was already rather long.
On your specific points:
1. and 2.: I agree
3. While I obviously agree that the structural features of the model are chosen by the modeller, I’m not sure it is always the case that the modeller has a good grasp of how they affect the output. But in this case it certainly seems probable that Flaxman would have realised that his model set up was such that it was bound to attribute virtually all the reduction in COVID-19 deaths to the government interventions it featured, and to generate a very high level of deaths if none of those interventions had occurred.
4. This is probably overstating the modeller’s control in most cases: the data will usually place some constraint on the range of possible output, although varying the assumptions made and/or the model’s structural features might in some cases effectively determine the model output almost regardless of the data.
5. That may be the modeller’s intent in some cases (perhaps you had in mind the Ferguson / Imperial College Report 9 that effectively argued for imposing lockdowns?), but in many cases it would not be the case, either because the model output had nothing to do with recommending policy actions or because the modeller was a good scientist who did not allow his policy desires to override his scientific objectivity. (In the case of the Flaxman et al study, they aren’t trying to persuade the reader to adopt their preferred policy, but they are trying to justify and extoll the adoption of a policy that they had previously puched for and which had been adopted.)
As always, the garbage-in garbage- our principle applies. A good question is why anyone should consider Imperial College production?
Any skilled virologist/epidiemologist is used to Gaussian curves for this kind of flu and corona virus, with Spring à sort of natural limit.
Also there is the quite obvious point that lockdown at home without separating ill individuals from non-ills triggers much virus spreading within each famille.
Not least, the fact that there is no evidence that lockdown brought less infection. See Spanish study on 60000+ individuals where workers who continued to go to work were less infected than workers remaining at home.
So one more IC bulls… study.
Daniel
Link, please.
Curves without underlying causes are just lines on a screen.
The important issue is *why* curves flatten or bend downward. With an epidemic, unconstrained, the reason is herd immunity – which is what causes a Farr’s Law curve.
But “unconstrained” is a very important qualifier. You can get the same phenomenon with any other cause that reduces the effective reproduction rate Rt, and we have a situation with all sorts of actions that one would expect to have that sort of effect: social distancing, hand washing and other sanitation, mask wearing. For the death curve, you could also get it if you just give the virus to everyone who is going to die of it, more easily predictable with this than other epidemics. Rt could also drop if the virus mutated in a way that led to less transmission – unlikely so far.
Since we appear to be nowhere near herd immunity (unless one hypothesis on very strong effects of non-uniform social networks pans out), and the virus has not mutated in that way, we are left with interventions as the likely cause. And if that is true, these curves cannot be used to find an upper limit on the total population fatality rate that is anywhere close to what is claimed.
“Since we appear to be nowhere near herd immunity ”
There is an assumption that we are all equally susceptible that I don’t think is borne out by the evidence.
If you start the models with varying percentage of the population already immune to the virus then you may get a better fit.
There are many unknowns. But the more you toss in factors that are not understood, the more you get into modeler’s heaven, reality’s hell.
It looks like the virus hasn’t mutated in a way that reduces its lethality or transmission. That’s under constant watch by the viral genetics folks.
Non-uniform social networks certainly impact transmission, but note that the estimated R0 comes from populations that also have non-uniform social networks. The R0 for a population with uniform social networks and uniform susceptibility might be very different, but we cannot measure that.
In other words, R0, by the way it is derived, already takes into account the social networks and susceptibility characteristics of the populations in which it has been measured. I suspect it would be a lot higher in a uniform population.
Since I couldn’t edit the other reply with an afterthought, here it is…
As far as I know, R0 is determined while the epidemic is unconstrained by mitigation. But… it is possible that the effects of non-uniform networks are different early in the epidemic, and thus R0 doesn’t do that good a job of taking into account effect of those networks. It should, I think, take into account non-uniform susceptibility.
“it is possible that the effects of non-uniform networks are different early in the epidemic, and thus R0 doesn’t do that good a job of taking into account effect of those networks. It should, I think, take into account non-uniform susceptibility.”
Yes, indeed so. Variability in people’s social network sizes substantially increases R0, but as then the epidemic progresses transmission falls much faster than (as in simple compartmental models) pro rata to the number of people remaining uninfected, as the more connected people (who are both more susceptible and more infectious once infected) on average get infected earlier. Other sources of variation in susceptibility, such as those that are highly likely to arise from previous infection with common cold coronaviruses, have a similar but less strong effect, as such variation is not linked to differences in infectivity.
This implies that interventions such as banning public events, closing bars, restaurants and other venues, and lockdowns are all likely to have their strongest effect early in an epidemic, when most of the highly connected people (super spreaders) remain uninfected.
But the problem remains that until and unless the herd immunity threshold (which will be lower in the presence of population variability) has been reached, or mass vaccination has taken place, once these social distancing interventions cease so that people can resume normal living, the epidemic will start to grow again.
Hi Nic
thinking about this & climate etc..
has anyone asked to see & duplicate the code/model used?
Nic,
Thanks for your response. It made me think of the criticisms of the data selection and adjustments made to it in preparing the Mann et al hockey stick. So I have revised my points as below.
1. All models require the modeler to select the raw data and decide what judgmental adjustments to make to that data;
2. All models include assumptions by the modeler about data that is either unavailable or uncertain, and relationships between such data and other data;
2. The output of the model is determined, in part, by the data selection, adjustments and assumptions;
3. The structural features of the model are also chosen by the modeller, with full knowledge of their impact on the output;
4. The data selection, data adjustments, assumptions and the structure of the model together largely determine the output;
4. The output is therefore an if… then… statement largely controlled by the modeller;
5. The purpose of many climate and Covid-19 models is not simply to inform the viewer but to persuade the viewer to adopt the model’s output as accurately reflecting the real world — past, present, and, if projected, the future.
I found this discussion could have been more revealing, if the participants had been a little less collegial.
Is the World’s COVID-19 Response All Wrong?
The Agenda with Steve Paikin
https://www.tvo.org/video/is-the-worlds-covid-19-response-all-wrong
4/ the fatality rate for sick LTC residents is really quite astonishing: 31%, but for sick staff only 0.2%.
COVID deaths in Ontario are dominated by very old: 33% of COVID deaths over 90, another 35% 80-89.
https://twitter.com/ClimateAudit/status/1274910496974606336
Unprecedented and Unjustified A Charter Analysis of Ontario’s Response to COVID-19
https://www.jccf.ca/wp-content/uploads/2020/06/2020-06-22-Unprecedented-and-Unjustified-Ontario-Lockdown.pdf
COVID-19: Epidemiologic summaries from Public Health Ontario
https://covid-19.ontario.ca/covid-19-epidemiologic-summaries-public-health-ontario
Pingback: Did lockdowns really save 3 million COVID-19 deaths, as Flaxman et al. claim? |
Pingback: Did lockdowns really save 3 million COVID-19 deaths, as Flaxman et al. claim? – All My Daily News
Nic: Unsure if you are aware of U of Nottingham study on herd immunity referred to in this article – https://medicalxpress.com/news/2020-06-herd-immunity-threshold.html
Keith – thanks. I had read the preprint but not seen that it had now been published. They use a simpler model than that I used in my article: https://judithcurry.com/2020/05/10/why-herd-immunity-to-covid-19-is-reached-much-earlier-than-thought/ , and don’t allow for variability in susceptibilty other than that arising from a different rate of contacts with other people (social connectivity).
Nic,
Thanks for an excellent review.
I would suggest that even before considering the dubious priors underpinning the IC paper, it is worth considering the question of exactly how one does define an exclusive and exhaustive list of NPI.
I do not know the answer to this question, but I am firmly of the belief that no analysis can have any credibility at all without considering transportation within cities and transportation between cities.
Several excellent papers have looked at the role of mass transit systems within cities as a mode of local community transmission. In all cases, front line transport workers are significantly over-represented in exposed/infected cases. They are both victims and superspreaders. With such an infection source in operation, if lockdown is in operation, each exposed/infected commuter goes home to a physically limited family. The papers I have seen all conclude that there is a strong requirement to improve control of disease spread via mass transit systems – fair enough – but fall short of addressing the next obvious question:- if mass transit systems are allowed to operate as uncontrolled sources of transmission, then does lockdown provide a benefit or a disbenefit in checking the spread of the disease within a city? I do not believe that any sensible analysis can ignore the question of policy with respect to mass transit.
Moving on to the question of transportation between cities, I would draw your attention to this recent paper:- https://virological.org/t/preliminary-analysis-of-sars-cov-2-importation-establishment-of-uk-transmission-lineages/507 (More reasonably, I should suggest that you bring it to the attention of the authors.) What it suggests is that most of the UK transmission (genetic) lineages died out towards the end of April. The growth in active cases from before that date are all associated with imported viral strains – explained by the 200000+ passengers perday who were arriving before lockdown measures and the 15000 passengers per day who were still arriving in the UK even after the lockdown measures were introduced. It is unreasonable to suggest that the list of NPI is exhaustive, or even reasonable, given these data.
Paul,
Thanks for your comment. I agree that Flaxman’s last of NPI is far from exhaustive, omitting for instance banning incoming air travel or enforcing quarantines on arrivals. However, I think the omission of non-NPI caused behavioural changes is probably a more critical omission.
Noted re mass transit systems within cities. But, taking London as an example, my understanding was that few people used public transport once the UK lockdown was implemented?
Thanks for the link to the Pybus paper (which I didn’t tak eas relating specifically to transportation between cities). On my reading, it doesn’t actually show that a substantial proportion of UK transmission lineages died out. it doesn’t say how the sampling frequency varied over time nor is it entirely clear many samples were allocated to the ~1400 lineages (I think under 10,000). Assuming some bias in sampling towards the last 4 weeks, such that on average 2 samples per lineage were obtianed, and a reasonable dispersion of average lineage sizes during that time, binomial distributon based estimation suggests that many of the ~400 lineages with no samples in the last 4 weeks were simply the results of random sampling of, in the main, lineagess that were relatively small, but continuing to exist, during that period.
I have seen recently that part of the issue with the high death rate in LTCFs in Sweden is due, to some extent, with their casual attitude towards treating infected older people, i.e., not treating them with relatively uncomplicated oxygen therapy. So that could have nothing to do with the effectiveness of policies to limit spread, per se.
That said, I never thought that their death rate would go above that of France. Not only has it done so, but it’s about to overtake Italy and could conceivably even go higher than in Spain or the UK? Yikes.
Seems to me that there a very long list of reasons why Sweden’s rate should be lower relative to those countries. And the rates relative to other Scandanavian counties also strongly suggests that those who claim that government interventions don’t have any effect on reducing spread – at least over the short term – are clearly blinded by ideological biases. The growth of death rate despite structural advantages for limiting spread and despite the clear advantages early in the pandemic, seem to deliver a pretty powerful and unambiguous conclusion.
Consider how Switzerland at one point had a relatively similar death rate – presumably for reasons like lifestyle predictors of greater spread and proximity to Lombardy and amoint of travel to/from hotspots – and now the death rate there is less than 1/2 that of Sweden.
We need to wait to evaluate the long term scenario, in terms of economic and health outcomes. But I can’t see any reasonable argument that Sweden’s outcomes are consistent with a view that government mandated social interventions don’t have a significant impact on rate of spread (and associated # of deaths, even if that isn’t necessarily directly proportional).
Maybe Sweden is an outlier for some reason with respect to a typical causality between SIP orders and health outcomes – but I haven’t seen an argument that convinces me that is the case. If someone wants to argue that SIP orders are totalitarian, that’s obviously a matter of personal priorities: A whole other category of debate
“That said, I never thought that their death rate would go above that of France. Not only has it done so, but it’s about to overtake Italy and could conceivably even go higher than in Spain or the UK? Yikes.”
In other words a place that didn’t lock down was almost as bad as some places that locked down, about the same or a little worse than some places that locked down, and nowhere near as bad as others that locked down (Belgium – capital of the EU – is currently worst on the planet).
You realize that’s the point some folks are making, right? If lockdowns resulted in the same results (or worse, or only marginally better depending on some other factor), then it really isn’t obvious that lockdowns were effective. If the deaths in Sweden were in LTC then the street cafes weren’t the problem. Unless you have some evidence nonagenarians in Sweden are big partiers.
Now tell us why the newspaper insists the current “surge” is due to the quiet barbecues in the sunlight on May 25, but not the all night street rallies on May 26.
Good luck.
> In other words a place that didn’t lock down was almost as bad as some places that locked down,
A place that has numerous reasons for a structural advantage has had similar outcomes as those that don’t have those structural advantages. That it’s even close given how they all started, is the point. I spelled that out. But you ignored it. Sweden started out way behind all those countries in death rates at the jump. Thst had nothing to do with COVID-targeted policies. It blew its structural advantage as Norway, Finland, and Denmark did not (in terms of health outcomes, at least in the short term; the comparison in terms of economic outcomes is less clear).
The places that are much more similar, with similar structural advantages and more similar starting conditions, have had much more favorable htalth outcomes.
I could list all those factors in more detail yet again, but it would obviously be pointless.
And Sweden compares unfavorably to pretty much ALL countries in terms of how they fared after the beginning of the pandemic, except for other countries that also trended poorly – like the U.S., and Brazil.
You’re not even trying.
So, the “mutation” factor could be a bigger factor than we will ever know– it does look like Italy got the worst of it…
Presumably people have picked upon Karl Friston’s work in this area?
Can you give a link to it?
Pingback: More Articles and Studies – Common Sense for the Common Good ®
The disease is still in progress.
How can anyone jump to any conclusion now?
You would have to know the strength of the seasonal trend and whether there will be another death spike in the fall.
Flu doesn’t tend to spread well outdoors.
In the US that might explain why northern states were hit hard when their weather was colder. Now those people are outside more and the disease doesn’t spread as fast.
People in the southern states had been outside a lot in the spring … until the hot weather drove them indoors more often for AC … making it easier to spread the virus.
Also important are ncoming flights from China and Europe bringing in people with the virus to the US.
“Lockdown”.is a theory — what did people actually do? And how were people on incoming planes locked down? Assuming they were.
Richard
Regarding ‘lockdown’ aka house imprisonment, as we have been pointing out for months, locking sun starved vitamin D lackimg UK people up in their houses was a very foolish thing to do
‘Writing in a major study, the researchers said: ‘Forcing people to remain indoors may have increased or assured contagion of Covid-19 among same household dwellers and among patients and personnel inside the same hospital or geriatric facilities.
‘In contrast, healthy people outdoors receiving sunlight could have been exposed to a lower viral dose with more chances for mounting an efficient immune response. ’
Ironically not only was it very foolish to imprison people inside their houses and deny them the ability to go outside where it is much safer as regards the virus, , but homes themselves are some of the most dangerous places to be.
The royal society for the prevention of accidents reports some 6000 accidental deaths a year in uk homes which includes some 600 stair related deaths.
The Covid 19 data is available for the South West of England and more specifically our county.
6500 deaths were forecast, 356 have occurred. I can find very few if any cases that were spread within the community. 95% of Deaths occurred in care homes and hospitals. The hospitals decanted elderly people without testing into care homes of whom a number had CV and infected the care home
Hospitals themselves remained a centre for the virus with very many of those who entered hospitals for other reasons, then catching the virus and dying with it not OF it.. As my link to a major study above shows it was gross stupidity to lock up sun starved people in well sealed small homes without gardens, where the virus circulated or was reintroduced to the wider indoor world through visits to shops etc. Outside was ALWAYS the best place to be
Our county (which is one of the least affected) is nevertheless very representative of many other smaller communities in the UK , with most deaths being in the big cities and amongst immigrant communities who for a variety of reasons were more susceptible. .
A limited number of people in specific settings were badly affected, of which some of the elderly or the already very ill were in the front line. Instead of separating those categories out, everyone was locked down and our economy trashed.
Many additional deaths have been caused by people not willing to attend hospitals or who did attend and got infected and the fixation on CV has meant many thousands will die of ailments that could have been treated-Cancer charities estimate 38000 more deaths in 2021 and heart organisations a similar number . They were turned away for 12 weeks whilst the newly built covid 19 Nightingale hospitals lay empty.
So of course not so many people need to have died, but that is primarily because we treated the pandemic as the Black death and not vastly reducing numbers at risk by testing and rooting out the virus from care homes and hospitals. The numbers who actually caught it in their own homes through being locked down is as yet unknown.
tonyb
One problem for this sort of analysis, at least in the US, is that “lockdown” is not well defined. Is it true home imprisonment, or is outdoor exercise allowed? Can you go to grocery curbside pickup?
Given how much that varies, I think the term “lockdown” needs to go away.
Here in Arizona, our “lockdown” allowed one to go outside, and even travel, for “essential purposes” which included exercise. The effect, along with the way businesses were required to operate (curbside or mail order only) had to have significantly cut the transmission rate.
Unfortunately, when that ended on May 15, too many people took it as a sign that the virus had been defeated (or wasn’t that bad a threat) and started taking a lot of risks. You can see the result in our numbers. Also, that sort of behavior will vary by population, another problem in trying to analyze the impacts of NPI.
Nic: Here is another point of view.
https://medicalxpress.com/news/2020-06-epidemic-covid-deaths-october.html
Regarding infection estimating fatality rates cross-nationally, a disturbing article:
–snip–
Both demography and weak health systems explain why COVID-19 deaths are more concentrated among younger people in the developing world
Although predicted IFRs display a steep age gradient in all contexts, due to demographic differences the bulk of deaths in low- and lower-middle income countries is predicted to come from middle-aged patients (40-70).
Less obviously, differences in health system capacity are also likely to flatten the age gradient of COVID-19 deaths in developing countries. In Europe, data is consistent with the hypothesis that intensive care saves the lives of a higher proportion of young than elderly COVID-19 patients. Thus when high-quality intensive care is lacking, the advantages of youth are more muted.
https://www.cgdev.org/blog/predicting-covid-19-infection-fatality-rates-around-world
Er… estimating infection fatality rates..
https://www.folkhalsomyndigheten.se/contentassets/53c0dc391be54f5d959ead9131edb771/infection-fatality-rate-covid-19-stockholm-technical-report.pdf
We estimate the infection fatality rate of COVID-19 in the Stockholm region in Sweden, for cases with symptom onset 21–30 March. We estimate the number of deaths, i.e. the numerator, prospectively, using data from an individual-level database of all confirmed cases in Sweden. The number of infections in the denominator is based on an estimate of the total umber of infections (including unreported) per confirmed case. This estimate is based on a survey in which a random sample of the population in the Stockholm region was tested for SARS-CoV-2 by means of a Polymerase Chain Reaction test.
Our point estimate of the infection fatality rate is 0.6%, with a 95% confidence interval of 0.4–1.1%. For the age group 0–69 years, we get an estimate of 0.1% (c.i. 0.1–0.2%), and for those of age 70 years or older our estimate is 4.3% (c.i. 2.7–7.7%).
Most of the uncertainty in our estimations concerns the relationship between the total number of infections and confirmed cases. We assess how the estimate of this
relationship, and thus the infection fatality rate, varies with alternative assumptions about the time window during which an ongoing or previous infection can be
detected with Polymerase Chain Reaction testing. Additional analysis of excess mortality in the Stockholm region during the period studied suggests that our
estimate is likely to be conservative.
Colchicine vs COVID-19:
https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2767593?utm_source=silverchair&utm_medium=email&utm_campaign=article_alert-jamanetworkopen&utm_content=wklyforyou&utm_term=062420
Open label randomized trial showed benefit of chochicine in event-free survival time and survival rate, but not biomarkers of heart disease.
review of ventilation in COVID-19:
https://jamanetwork.com/journals/jama/fullarticle/2765302?widget=personalizedcontent&previousarticle=2767752
Hi Nic,
I was going to start this comment with ‘I haven’t read Flaxman’s paper’, and then thought I’d better read it.
After reading it, I was dismayed. It gives Bayesian analysis a bad name.
Given the age groups and social circumstances of those that died from SARS-CoV2 it is insane that the prior assumptions would not have been updated as this data became available. In fact it was already clear from deaths in France and Italy before it hit the UK. Questions which could have been asked:
What is the probability that I will die from ‘coronavirus’ given that I a over 70 and live in a care home?
What is the probability that I have been infected with ‘coronavirus’ given the prevalence and the sensitivity of the test?
What is the probability that I will be asymptomatic given that I have been infected?
Flaxman’s model uses high probabilities for Social distancing/lockdown which can only be reinforced by new data.
Is it perhaps an attempt to justify Prof. Freewillie’s initial predictions?
And why gaussian distributions? Given the age range of mortality, wouldn’t a beta distribution be better?
Thank you for your insightful critique of Flaxman’s nonsense.
Pingback: Covid19: a «Nature» não demonstra que o Confinamento salvo Vidas | O Economista Português
Dr. Curry,
I’m curious why you allow certain people to use their own version of somebody else’s name in replying to that person in the Comments? It demeans the quality of the interactions, promotes a lame kind of attempted bullying and generally indicates the weakness of the user’s position. Given that there are very few practitioners of this elementary school level of immaturity in your Comments threads it seems like it would easy to rectify. While many places on the Internet give free reign to this pathetic behavior, to find it on a Science blog does a disservice to your readers, the persons being insulted and even the person who does it. .
I agree i don’t like it. But i have a limited time to read comments and moderate the blog. I will try to do better
Y is the USA.EU and UK not bothered,about the COVID deaths in their part of the world ?
Could it be that they want it ? Who are the dead ? The dead are the pensioners, and the persons,who are fatally sick.dindooohindoo
The gainer in every combo,is the West – which makes one wonder,how the COVID magically mutated in its new avatar.
Posit No.1
Assuming that these dead persons in the West,had a residual life of 15 years, and we can assume that,by August,2020,there will be around 600000 dead in the West.
The pension to a pensioner,would not be less than 12,000 USD per annum, on an average,at the minimum.In addition, the medical and other social costs,on an aged pensioner,would be not less than another 8,000 USD per annum.
If they die,then on 6,00,000 people,if the West saves 20,000 USD per annum, you net USD 12 Billion,PER ANNUM – which will be around 200-300 billion for 15 years
One could argue that the US Fed just printed,the USD 12 Billion – but now it need not.The Youth in the west,had to work at high rates of tax and deductions – to finance the aged pension and health care benefits – which ultimately,led to outsourcing.
The scam would be shocking,if the dead,had no insurance ! That would be telling ! If 6,00,000 are dead,with insurance and an average insurance claim,of USD 1,00,000 – then you have a bomb – to wipe out the insurers.
If 10 million die – we are looking at net savings of USD 200 billion per annum and USD 3 Trillion over 15 years.This will also solve the health insurance problems in the US/EU,as the high claim insurers,will cease to exist – and thus lower the insurance costs,for the young,and the cost of labour in manufacturing.
if the aggregate savings on pensions and medical costs are USD 100,000 per annum,then on 10 million dead,we have a saving of ISD 1 trillion per annum,as a perpetual annuity (which is the minimum target – I suspect) – as the strategem ,is to kill people,with co-morbidities – and these are the people,who are a burden on the medical and pension infrastructure.
So the private LIFE insurers,take a 1 time HIT,in terms of claims paid out – and the state,gets a recurring benefit,in terms of pensions and health care costs – of which,some of the gains of the state,are passed back to the insurers,to offset the claim losses (and keep insurance rates low),and some of the gains to the state, are passed back to the residual young population,to reduce the rates of medical and life insurance.
Posit No.2
Large number of services and industries,in the west,will die out.That will release labour and reprice resources and rents – to drastically lower costs – and that will make,”Make in USA”,viable
How will the state finance the loss of tax revenue and GDP.Ultimately,the state will have to demonetise the deposits, in banks, of the westerners.Simple ! The USA will not be able to demonetise the PRC holdings of US T-bills – not even if the PRC sinks a US aircraft carrier in the South China Sea.
Posit No.3
All the nations who borrowed loans from PRC – will now force the PRC to do debt write offs.That will be a huge loss to the PRC,after the manufacturing shift from PRC to West.Post COVID,If 200 million people are unemployed in PRC – then you have Tiananmen – Part 2 – and then a PRC attack,on the Indian weasels, and US satellite states,like Taiwan.and new stooges like Vietnam.
Of Course,the PRC could also force the IMF,and the WB,to waive loans – but the harm to the PRC,will be done 1st.
Posit No.4
Trump postpones the US Polls,as people cannot stand in queues,and no electioneering,is possible – and he has the cure – and by September,the pensioners are dead – death rate and infections rates drops ….. who is the gainer ? If Trump is winning – Putin will stay calm – else,he might attack Eastern EU.If Trump is winning – then it will be the last chance for PRC to annex Taiwan and Vietnam – and make Trump lose face. But the odds of PRC action is medium.
Posit No.5
With massive unemployment in the West – the migrants will exit.Asians were made to clean toilets – that is their worth.They will exit.That will solve the migrants problem,rents and property rates will fall,labour will reprice,and the Westerners,will have to,start to work
The West has to take a BIG PICTURE view.South East Asia and Indian and Nepal ,are over populated,and there is no humanity there.There is no sentience,in the “so called humans”.They are robots – and 80% of them,have to die.Their time is over – they are obsolete, a dead weight,and a burden on earth.This will de-price the resources sector,lower demand,and solve the environment problem,forever.
Africans have been exploited,for at least ,2000 years – and they deserve,many more chances.
There are 3 simple steps
Are the “so-called humans” – having a “sentience” – to be assessed based on their “individual and collective actions”
If not,then they are “robots”
It is time to “terminate the robots”
It is the moral and ethical solution.They are redundant and obsolete,and there is no purpose served,by their existence.Nations in Asia,will not be able to feed or employ these worms,and that will cause strife.hate,violence, genocide,jingoism and the rise of right wing =,demagogic demonic dictators – and then, catacylysmic wars – which will ultimately,harm the West.Anywhich way,the robots will be purged – Virus is better than nukes – for the bots,and the environment.
The COVID antibodies,will ultimately reside in 7 billion people,and those,are the receptors,of the next,”terminal bio-weapon”
survey of COVID-19 in some skilled nursing facilities in the US
https://jamanetwork.com/journals/jama/fullarticle/2767750?guestAccessKey=87eeabaf-ea4f-4439-a2fd-a8c52dd4ae44&utm_source=silverchair&utm_campaign=jama_network&utm_content=covid_weekly_highlights&utm_medium=email
As the plague demographics shift to younger people there are fewer deaths but 1 in 10 have long term health problems. COVID-19 will now be a preexisting condition once the AHC is repealed.
https://www.reuters.com/article/us-health-coronavirus-effects/scientists-just-beginning-to-understand-the-many-health-problems-caused-by-covid-19-idUSKBN23X1BZ
““We thought this was only a respiratory virus. Turns out, it goes after the pancreas. It goes after the heart. It goes after the liver, the brain, the kidney and other organs. We didn’t appreciate that in the beginning,” said Dr. Eric Topol, a cardiologist and director of the Scripps Research Translational Institute in La Jolla, California.
In addition to respiratory distress, patients with COVID-19 can experience blood clotting disorders that can lead to strokes, and extreme inflammation that attacks multiple organ systems. The virus can also cause neurological complications that range from headache, dizziness and loss of taste or smell to seizures and confusion.
And recovery can be slow, incomplete and costly, with a huge impact on quality of life.”
jacksmith4tx: ““We thought this was only a respiratory virus. Turns out, it goes after the pancreas. It goes after the heart. It goes after the liver, the brain, the kidney and other organs. We didn’t appreciate that in the beginning,” said Dr. Eric Topol, a cardiologist and director of the Scripps Research Translational Institute in La Jolla, California.
Is that really news? It seems to me that I read it months ago. But it is worth repeating, as it seems to be slowly sinking in how many organ systems the virus attacks.
Maybe the clinical effects are becoming more prevalent?
Thanks for the link.
So it’s able to move between different body tissues. I wonder why that is? It emerges having learned to do that. It emerges with the strong ability to be transmitted between humans.
I saw that paper and I wonder how reliable it is. What I’ve been noticing both in Europe and the US is that while cases are flat or increasing, hospitalizations and deaths are continuing to decline. The data for Sweden on Wikipaedia are particularly striking. I doubt that people who are not hospitalized would have much lingering disability or organ damage. My suspicion is two fold.
1. Doctors are getting better at treatment.
2. The population of vulnerable people has declined significantlly. Most have already had the disease.
dpy6629, What’s the latest info on post infection immunity. I see a lot of stories but nobody seems to be sure how long the antibodies last.
I would be shocked not to see the death rate drop considering the billions being spent to develop treatments. I don’t think there is a comparable event in human history that has seen such a global response in the amount of money and R&D spent to solve a problem.
I predict we will have a vaccine by 2021 and we will be better prepared for the next pandemic.
Very early on, I read a Chinese paper on all the different systems directly attacked by the virus. However, that doesn’t mean that the effects are permanent. Also, autopsied showed permanent changes, but again, not clear if that happens in people without the severe illness. In severe casesdamage isn’t just from the virus, but also potentially from immune system overreaction, which can damage pretty much everything, including from blood clotting disorders including DIC.
Recently, I saw one where asymptomatic patients had lung changes on CAT scans – again, not clear how long they last.
But, there are certainly plenty of cases in “not at risk” people who nevertheless get seriously ill.
If anyone has a cite to studies on long term effects (other than obvious ones like strokes), I’d be interested.
Reuters again, optimistic review: https://www.reuters.com/article/us-health-coronavirus-lessons-special-re/special-report-as-world-approaches-10-million-coronavirus-cases-doctors-see-hope-in-new-treatments-idUSKBN23X1I5
Should we start do act like this for all the Winters to come?
Should we lockdown in the same fashion in order to save millions of lives from death due to flu, pneumonia or even air pollution (since in the winter we waste more energy for heating, we pollute more)?
You might find this interesting: https://www.medrxiv.org/content/10.1101/2020.06.15.20131953v1
Sue Denim: https://www.medrxiv.org/content/10.1101/2020.06.15.20131953v1
Thank you for the link.
Imperial College modellers again. Why do people assume these people are experts, when all they are experts at is tweaking unmeasurable parameters to fit predetermined narratives?
Of course Imperial College claims that lockdown saved 3 million lives. Anything less than that should see Imperial College sued into oblivion for causing economic armageddon across the UK.
A bigger question is what the blazes is Nature doing publishing such anti-science?
As I said in Posit 4 of MY post above on June 27, 2020 at 9:26 am
https://edition.cnn.com/2020/07/30/politics/trump-delay-election-no-authority/index.html
Trump WILL POSTPONE .It is 100%.dindooohindoo
I had stated in the 1st 3 lines of the posit
“Trump postpones the US Polls,as people cannot stand in queues,and no electioneering,is possible – and he has the cure – and by September,the pensioners are dead – death rate and infections rates drops ….. who is the gainer ?”
I expect sky rocketing death rates and daily cases !
I expect a Bio Terror hit on US Soil in the last 2 months of this calendar.Only Bio Terror – all else is obsolete
I expect a disaster in the South China Sea – INVOLVING the the PLN and the US Navy
And a disaster in West Asia – Saudia/Persia/Yenen/Syria and Israel
All these will hit US assets and all will be false flags.
It is as certain and inevitable , as the Sun Set. (not the Sun rise)