Hey, USA! How are those death Projections?

Well, despite my dramatic and terrifying underestimate of how US residents would pull together and alter their behavior to slow the spread of SARS-CoV-2, there are still many COVID-19 deaths happening around the USA. Looking at graphs, the growth of the disease looks a lot more linear than exponential lately, and that’s a very, very good thing. With linear growth, each doubling takes twice as long as the last doubling, which means that we buy ourselves lots and lots of time to make vaccines and take more preventative actions before reaching megadeath-scale numbers. Of course, this is still much worse than negative growth rate in COVID-19 deaths or just stopping the disease in its tracks, but when the choice is between linear growth and exponential growth, as it was for the USA at the end of last month, you have to choose the former.

So where are we? In my first post on this subject, regardless of how wrong I was on the important question of whether infections and deaths would continue doubling at the same rate as in the recent past, the math on how many doublings it takes to reach megadeaths is the same: it would require 8 doublings. I thought COVID-19 related deaths might get as many as 6 doublings before finally slowing dramatically. But it only got two, it looks like, before we hit linear growth. (This was, if you remember, foreshadowed by someone who had better data and better thinking on this than I, militantagnostic, who told us then that we were less than two weeks out from killing the exponential growth curve.)

Still, that first look at the data indicated 6,593 deaths. And if I’m calculating correctly, that means the most recent doubling, which has taken 9 days, brings us to 2.5 doublings, total. Those doublings represent 45,000 human beings who have lost their lives to this disease, as you can see:

ECDC data on deaths and time-to-double deaths related to COVID-19.

If growth stays linear, we will reach a total of 3 doublings at about 52,000 persons dead, which is about 3 days from now. This is bad. This is tens of thousands of tragedies playing out across the country that must be partly attributed to our government’s response. Still, I did say:

There really is very little hope at all that deaths will be less than 100k on Earth Day this month.

and yet here we are, and we’re at less than half that number. For those of us not grieving someone, we might take a moment to reflect on how much better things are than they might have been if we hadn’t gotten good leadership from our governors and good cooperation from our neighbors.

This is a good thing, yet it is also not time to relax. With 2500 people dying in the last 24 hour reporting period, that means that there are many, many infected persons in the USA right now, and “reopening the economy” won’t place us back where we were in mid-to-late January with a single hot spot. There are tens of thousands of infections ready and able to touch off tens of thousands of hot spots if people begin mingling again. And yet, that’s exactly what we seem set to do with beaches reopening in Florida and churches holding more services again this weekend in Texas and Georgia. There are many places where this is going to get bad again.

I won’t tempt the wrong again by making more projections, but we are barely more one doubling away from 100,000 persons to mourn. Clearly some people are going to show up for demonstrations and church services and days at the beach. Clearly more people will get sick and some of those will die because of this behavior. I hope that the people of the USA continue to show more wisdom than I thought likely at the beginning of this month, and I’ll continue periodically checking the statistics to see whether or not that comes to pass.

Stay safe everyone.


  1. says

    Based on the Worldometers numbers for the world, roughly every six days the infected increase by 500,000, and the dead increase by about 150%. March 20 was when Iran, Italy and Spain started piling up numbers.

    April 21 @ 16:00 UTC: 2,500,000 infected, 174,000 dead
    April 15 @ 09:00 UTC: 2,000.000 infected, 126,800 dead
    April 8 @ 20:30 UTC: 1,500,000 infected, 87,800 dead
    April 2 @ 18:00 UTC: 1,000,000 infected, 51,000 dead
    March 26 @ **:** UTC: 500,000 infected, 34,000 dead
    March 20 @ **:** UTC: 275,000 infected, 11,400 dead

    Noticeably, if you take the US’s numbers out of the equation, the world’s rate of increase is slowing. Most of the increase since April 8 has been from one country. However, the increase in number of infected (122,000) in the last 26 hours is one of the biggest yet.

    If the world had barricaded China in January (no one comes out, only supplies go in), we might have stopped it. Now barricading the US might be the answer.

  2. says

    About a Vietnam war’s worth (in terms of American casualties only) this should be hung around congress’ necks and carved in stone, then chained to Trump’s ankle and drop the whole mass off a bridge.

  3. says

    The number of deaths is likely larger than the number of confirmed deaths, as the NYT has been reporting. By comparing the recent mortality rate to historical averages, they estimate there are 4k unconfirmed deaths in New York City alone. That also includes people who died simply because hospitals were overwhelmed, but in my book those people count as victims too.

    Speaking as a non-epidemiologist, I suspect that “linear growth” just means that the rate of exponential growth has slowed down. I expect that it will continue to grow exponentially, but now at a slower rate. Probably it’s more complicated though.


    Exponential growth was never going to continue… 11 or 12 more doublings after 100,000 and *everyone* in the USA would be dead.

  5. says

    No, but everyone would have been infected. At the mortality rate quoted by WHO a while back that would have meant a bit over 10,000,000 dead.

    If you read the original article, I thought near-exponential growth might continue until ~250,000 dead (about 8 million infected) and then arithmetic growth might carry us up to 500,000 to 2,000,000 dead (about 15 million to 60 million infections) over a much longer period of time. Why did I think that? Well, mostly because I thought a large number of people wouldn’t take social distancing seriously until people they knew started dying. 250k dead is still only 75% of 1% of the population, but the most connected people in our society would easily know **someone** who died, and start taking preventative measures. The more socially isolated people might not know anyone who died, but they are by definition more isolated to begin with.

    Of course, I admitted up front this was a guess, and I was clearly wrong, but 15 million infections isn’t an unreasonable or impossible number. It just didn’t come to pass because people in the USA turned out to have more wisdom than I credited them.

  6. publicola says

    The death numbers may represent an undercount. We know now that the virus was here earlier than we thought, and Covid deaths early on may have been attributed to the flu.

Leave a Reply

Your email address will not be published. Required fields are marked *