Here’s a two-futures forecast issued jointly by the CDC and Homeland Security. Dated April 9 — two days ago — the joint Task Force report estimates the likely adverse consequences of relaxing current social isolation measures in part or altogether. Not surprisingly, infections and resource needs and deaths mount under both of these suboptimal alternate scenarios. Under the “steady state” scenario — schools closed until the fall, 25% continue working from home, some relaxed social distancing — 195 million Americans will be infected, or about 60 percent of the population, 300 thousand of whom will die from the disease. In what seems to be Trump’s preferred future — “steady state” plus shelter-in-place for another month — 160 million, or half the population, will be infected, and 200 thousand will die.
It’s no surprise that relaxing restrictions would come at a cost in human lives. What’s striking though are the mortality rate estimates. 200K deaths/160M infections = 0.13% mortality rate. That’s only marginally higher than the death rate for seasonal flu.
What’s up with this estimate? In yesterday’s post I explored a hypothetical model in which the infection rate is far higher and the mortality rate far lower than most publicized models have inferred from the data. These new US government estimates take things even further. Do the Feds have access to confidential data sources indicating that the virus is both more virulent and less lethal than has been supposed?
I look at the Task Force graph. It shows a small blip in ventilator demand 30 to 60 days from day zero, which I presume to be the date of the first documented case in the US. The small blip is followed by a steeper rise, extending from about day 70 through day 365, requiring far more ventilators. Presumably then this model assumes that the current wave of viral spread, the one that’s cresting today, will hit only a small fraction of the US population. That’s what the widely cited IHME model assumes. But the present wave is projected to kill more than 60 thousand people. It’s already killed nearly 20 thousand.
Why then, with contagion jumping from 3 percent to 50 percent of the population — a sixteen-fold increase in infections — wouldn’t the Fed Task Force estimate that deaths would increase at the same rate? Extrapolating from the present data, why wouldn’t they estimate the corona death toll under the preferred Trump future to total something more like 1 million?
Maybe the NYT copied the numbers wrong when they released this report. Or maybe the Task Force knows what its client wants to hear. Either way, I’ve got a sinking feeling that I know what’s coming…