Why to Flatten the Curve

The best article I have found regarding previous evidence for social distancing, school closures, and isolation and quarantine as epidemic control strategies (“nonpharmaceutical interventions”) is this JAMA article on the 1918 influenza epidemic. They report the “weekly excess death rate” during the 6-month study period for 43 US cities during the peak of the epidemic. The term “excess death rate” refers to the number of deaths (and presumably serious cases of illness) that happen due to the pandemic over and above the typical death rate, in this case from seasonal influenza. Rates varied quite a bit across cities. Some cities had more deaths only because of the timing and peak of the epidemic that could otherwise have prevented. They report over 100,000 “excess” deaths during the 6 months they studied, out of 550,000 deaths in the US during the 1918 influenza epidemic. Cities that “implemented nonpharmaceutical interventions earlier had greater delays in reaching peak mortality, lower peak mortality rates, and lower total mortality” — meaning that cities that put more aggressive measures into place earlier had better outcomes for their population.

I took a key figure from the paper and broke it down here. The figure highlights the experience of four different US cities. Here is the one for St Louis, which plots the “excess deaths” over the 24 week period in the study:

Here is New York City for comparison, which did not implement school closures or a gathering ban:

And here is a comparison of four cities together:

You can see the variability in the peak of the epidemic, and in how long it lasted. The study found that the combination of school closure and public gathering bans was the most effective strategy, along with early implementation, although sometimes outcomes still varied across cities with similar strategies.

We know that the current situation with COVID19 does not apply exactly, but the ideas here still seem relevant.

Also of note, the time that the cities implemented nonpharmaceutical interventions ranged from 4 weeks to 24 weeks for an illness where they estimated the time from getting the illness to death was 10 days. The time to death seems more in the range of 4-8 week in the case of COVID19 from what I have seen, although I don’t necessarily think that everyone needs to be out of the hospital and well for it to be clear that the epidemic has been brought under control.