Monday, May 25, 2020

The CDC Lowers Its Estimate of the COVID-19 Infection Fatality Rate

Two months ago--in the third week of March--Trump and many of the state governors ordered a mandatory lockdown of the U.S. economy, while Prime Minister Johnson did the same in the U.K.  They did this because they were persuaded by some epidemiological models that without such a lockdown there would be up to 2.2 million deaths from COVID-19 in the U.S. and 510,000 deaths in the U.K.  These estimated death rates were based on the assumption that COVID-19 would be at least as lethal as the 1918 flu pandemic.

In my posts in March and April, I argued that there were at least three problems with the reasoning for these lockdown policies.  First, the estimates of the infection fatality rate for COVID-19 were unreasonably high.  Second, the assumption that without a mandatory lockdown people would not voluntarily change their behavior to mitigate the pandemic was implausible.  Third, the people advocating the lockdown had not engaged in any ethical analysis that would weigh the likely human benefits of a lockdown against its likely human costs, and so they never considered the possibility that the costs of a lockdown might exceed its benefits.

Now, a few days ago, the CDC has issued a new statement of its "COVID-19 Pandemic Planning Scenarios."  This statement is silent about the second and third problems.  But it implicitly recognizes the first problem.  In March, Dr. Anthony Fauci testified to Congress that about 1% of the people infected with the SARS-CoV-2 virus could die.  At the same time, the Imperial College model in the U.K. estimated that 0.9% of those infected would die.  In the new statement from the CDC, however, the "best estimate" scenario assumes a symptomatic case fatality rate of 0.4%; and since it also assumes that over 35% of the people infected show no symptoms, the total number of infections must be more than 50% higher than the symptomatic cases, which implies an infected fatality rate of around 0.3%.

This 0.3% is an overall fatality rate.  The rate is much lower for younger people and much higher for older people.  The CDC estimates a symptomatic case fatality rate of 0.05 for people younger than 50, 0.2% for people between 50 and 64, and 1.3% for people over 65.  Most of the deaths have been among older people in poor health, particularly those in nursing homes.  This supports a policy of protecting older people with prior medical conditions while leaving all others more free to move around.

The higher estimates of the fatality rate in March predicted millions of Americans would die without a shutdown, the lower estimates today would say that hundreds of thousands might have died without a shutdown.

We also have evidence now confirming the second problem with the lockdown policy.  Studies using various kinds of "big data" show that people in the U.S. started to restrict their travel dramatically in early March to mitigate the pandemic, which began before the mandatory orders were in place.  Now, we have similar evidence that beginning in late April, people have been travelling more even before the state governors began loosening their lockdowns.

We can also see now that people are confronting the third problem--the failure to weigh the costs and benefits of mandatory lockdowns--because the move to lift the lockdowns implicitly recognizes that these lockdowns are unsustainable because their costs exceed their benefits.

The general lesson from all of this is that when people understand what kind of pandemic they are facing, their behavioral immune system will motivate them to spontaneously change their behavior in adaptive ways to mitigate the health costs of the pandemic through social distancing, while also securing the social benefits of freedom of movement.

No comments: