On July 6th, eight of the models on FiveThirtyEight’s COVID-19 Forecast Tracker made projections for what the total death count in America would be on August 15th. The projections ranged from heartily optimistic (135,000 total deaths, or just five thousand or so more over the 40-day stretch) to cautious (173,000 total deaths), with a median projection of 153,000, which was not quite six hundred more per day (a lot of death, but on a consistent basis, the best number seen in the U.S. since March).
On July 13th, those same eight models made projections ranging from 153,000 total deaths (the previous median projection) to 168,000. The new median was 159,000, which might not seem like a huge increase, but represents a jump from projecting not quite 600 new deaths per day over the forecasting period to projecting nearly 1,000 new deaths per day.
Today, as you may have guessed, the numbers are worse. A range from 156,000 to 194,000 (!!), with an average of 162,000. The new daily average? Over 1,100.
That’s a lot of death. More than a 9/11 every three days. Roughly two-thirds the respective daily tolls from cancer and heart disease. Worse, the number keeps rising.
These models are not all blind to changes in human behavior. A few are, but the majority factor in either state reopenings, anonymized mobile phone data, or both. This goes to say that the models are making the best prediction they can given the facts of the moment, and the die is not fully cast for these next three and a half weeks (though it’s worth remembering that there’s a lag between infection and death, so some of the die is certainly cast, and we’re already seeing daily new death totals back over one thousand). Beyond August 15th, there’s a lot that can be changed through social distancing, the effect of which has been demonstrated here in North America over the months in which the death counts dropped, and mask-wearing, the efficacy of which is widely believed to be high.
The question, then, is whether we’ll do it—whether we’ll take these measures, and what hit we’re going to take from our collective, aggregate decision to less fully take these measures over the back half of June. In positive news, if you weight the state-by-state transmission rates from rt.live and compare them between a month ago and now, the national rate of transmission has dropped from 1.06 to 1.04. In negative news, any rate of transmission over 1.00 means the virus’s spread is accelerating (in an exponential manner, in which 1.04 on a larger base of infections can lead to more death than 1.06 on a smaller base, if the bases are different enough in size).
Hopefully, we can get the rate of transmission continues to decrease and find its way below 1.00. Hopefully, our healthcare system and the dynamics of infections continue to lower the rate of deaths per confirmed infection. At best, though, with hospitals stressed in multiple areas of the country, it seems we’re hoping for a death plateau, and a plateau that requires significant sacrifice to be reached.
