Monday, April 27, 2020

The flawed model behind the shutdowns

How One Model Simulated 2.2 Million U.S. Deaths from COVID-19 | Cato @ Liberty - Alan Reynolds:

April 21, 2020 - When it came to dealing with an unexpected surge in infections and deaths from SARS-CoV-2 (the virus causing COVID-19 symptoms), federal and state policymakers understandably sought guidance from competing epidemiological computer models. On March 16, a 20-page report from Neil Ferguson's team at Imperial College London quickly gathered enormous attention by producing enormous death estimates. Dr. Ferguson had previously publicized almost equally sensational death estimates from mad cow disease, bird flu and swine flu.

The New York Times quickly ran the hot news about this new COVID-19 estimate: 'The report, which warned that an uncontrolled spread of the disease could cause as many as 510,000 deaths in Britain, triggered a sudden shift in the government’s comparatively relaxed response to the virus. American officials said the report, which projected up to 2.2 million deaths in the United States from such a spread, also influenced the White House to strengthen its measures to isolate members of the public.'

A month later that 2.2 million estimate was still being used (without revealing the source) by President Trump and Doctors Fauci and Birx to imply that up to two million lives had been saved by state lockdowns and business closings and/or by federal travel bans....

[Dr. Ferguson's model] came up with 2.2 million deaths by simply assuming that 81% of the population gets infected* – ­268 million people – and that 0.9% of them die.... Neither the high infection rate nor the high fatality rate holds up under scrutiny.

To project that nearly everyone becomes infected the report had to assume that each person infects 2.4 others and those people, in turn, infect 2.4 others and so on, with the result that the number infected doubles roughly every four days. This 2.4 'reproduction number' (R0) was 'based on ... the early growth-rate of the epidemic in Wuhan.' But the reproduction number always appears highest during the early phase of an epidemic (partly due to increased testing) and has now fallen to nearly zero in China....

The worst-case Imperial College estimate of 2.2 million deaths if everyone does 'nothing' did not simply mean no government lockdowns, as a March 31 White House graph with two curves implied. It meant nobody avoids crowded elevators, or wears face masks, washes their hands more often, or buys gloves or hand sanitizer. Everyone does literally nothing to avoid danger. The Ferguson team knew that was unrealistic, yet their phantasmal 2.2 million estimate depended on it.... The obvious reality of voluntary self-protective actions makes it incorrect to allude to the extreme Ferguson death estimate, consciously or not, as evidence that heavy-handed government interventions prevented 'hundreds of thousands' of deaths....

The key premise of 81% of the population being infected should have raised more alarms than it did. Even the deadly 'Spanish Flu' (H1N1) pandemic of 1918-19 infected no more than 28% of the U.S. population. The next H1N1 'Swine Flu' pandemic, in 2009-10, infected 20-24% of Americans.

To push the percentage infected up from 20-28% to an unprecedented 81% for COVID-19 required assuming the number of cases and/or deaths keeps doubling every three or four days for months (deaths were predicted to peak July 20). And that means assuming the estimated reproduction number (R0) of 2.4 remains high, and people keep mingling with different groups....  Long before 8 out of 10 people became infected, however, a larger and larger percentage of the population would have recovered from the disease and become immune, so a smaller and smaller share would still remain susceptible....

In short, the Imperial College projection that 81% of the U.S. population could be infected if everyone just did literally nothing to protect themselves or others is inconsistent with rational risk avoidance, history and recent experience. Even with a much smaller percentage infected, however, deaths could still end up extremely high if nearly 1% of those infected died, as the Ferguson team assumed.

The assumed 0.9% death rate (within a range of 0.4% to 1.4%) was tweaked from a smaller estimate in a study of deaths in China by Robert Verrity and others, which found a 'case fatality rate' (CFR) of 1.38% among known and tested cases. By assuming that such confirmed cases underestimated actual infections by only about half, they inferred an 'infection fatality rate' (IFR) of 0.66%.

Epidemiologists have since found growing evidence that the number of undetected cases with few symptoms or none is much larger than merely doubling the small number of known and tested cases. A review of such research by the Oxford University Centre for Evidence-Based Medicine finds 'a presumed estimate for the COVID-19 IFR somewhere between 0.1% and 0.36%.' A middling estimate of 0.22% would by itself reduce the infamous 2.2 million death estimate to half a million even if 81% were somehow infected....

The trouble with being too easily led by models is we can too easily be misled by models. Epidemic models may seem entirely different from economic models or climate models, but they all make terrible forecasts if filled with wrong assumptions and parameters.

Read more: https://www.cato.org/blog/how-one-model-simulated-22-million-us-deaths-covid-19

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

*Prof. Ferguson's foundation for the 81%  figure was his estimation of the initial reproduction number (Ro, or 'r-nought'): "In such scenarios, given an estimated R0 of 2.4, we predict 81% of the G.B. and U.S. populations would be infected over the course of the epidemic." However, the herd immunity threshhold for the disease is a standard function of Ro: (Ro-1)/Ro ; if Ro=2.4, herd immunity would be reached at 58.33% of the population. -gd

No comments:

Post a Comment