Wednesday, April 8, 2020

Great, now we can’t trust the death numbers, either



All numbers below are based on yesterday’s reported figures of total confirmed cases, total deaths and total recoveries, published on Worldometers.info. I’m happy to send my spreadsheet that calculates all of these to anyone who wants to check the calculations, although they follow the assumptions described below.

I. Numbers based on total cases, actual and projected
Total US confirmed cases: 400,549
Increase in cases since previous day: 32,890 (vs. 30,818 increase yesterday)
Percent increase in cases since yesterday: 9% (vs. 9% yesterday)
Percent increase in cases since 3 days previous: 29% (vs. 32% yesterday)

“Set in stone” US deaths* over course of pandemic:  187,673 (based on 4% case mortality rate)
*This number assumes a) Total cases grow by 29% into the future (= yesterday’s 3-day growth rate in cases); b) We impose a massive lockdown, with prohibition of all non-essential travel, today; c) New cases drop to zero in 28 days, because of the lockdown, but they continue to grow at the current projected rate up to the 28th day; d) testing is widely available by the 28-day mark; and e) case mortality rate = 4%. To consider a 6% mortality rate, multiply each projection by 1.5. For 8%, double it. For comparison, Italy’s case mortality rate is currently 11.75%.

Projected as of April 15 (7 days from today):
These numbers answer the question: What would happen if we wait seven days to totally lock down the US, based on the assumptions below (which frankly are themselves wildly optimistic).
Total expected cases*: 789,250 (vs. 774,747 expected yesterday)
Total expected deaths set in stone* over course of pandemic: 340,245 (vs. 411,521 expected yesterday)
* The expected cases and deaths set in stone numbers assume a) Total cases grow by 29% into the future (= yesterday’s 3-day growth rate in cases); b) We impose a massive lockdown, with prohibition of all non-essential travel, on April 15; c) New cases drop to zero 28 days later, because of the lockdown, but they continue to grow at the current projected rate up to the 28th day; d) Testing is widely available by the 28-day mark; and e) case mortality rate = 4%. To consider a 6% mortality rate, multiply each projection by 1.5. For 8%, double it. For comparison, Italy’s case mortality rate is currently

Projected as of April 22 (14 days from today):
These numbers answer the question: What would happen if we wait 14 days to totally lock down the US, based on the assumptions below (which frankly are themselves very optimistic)?
Total expected cases*: 1,430,883 (vs. 1,473,380 projected yesterday)
Total deaths set in stone* over course of pandemic: 1,248,242 (vs.  614,248 projected yesterday. Note the number I actually published yesterday was 1,248,242. I’m not sure where I got that, but I apologize)
* The expected cases and deaths set in stone numbers assume a) Total cases grow by 29% into the future (= yesterday’s 3-day growth rate in cases); b) We impose a massive lockdown, with prohibition of all non-essential travel, on April 21; c) New cases drop to zero in 28 days, because of the lockdown, but they continue to grow at the current projected rate up to the 28th day;; d) Testing is widely available by the 28-day mark; and e) case mortality rate = 4%. To consider a 6% mortality rate, multiply each projection by 1.5. For 8%, double it. For comparison, Italy’s case mortality rate is currently 11.75%.

Date on which 500,000 deaths will be set in stone: April 20  
Date on which 1 million total pandemic deaths will be set in stone: April 28
Number of deaths set in stone on April 30:  (vs. 1,215,460 estimated yesterday)


II. Numbers based on total deaths, reported and projected
Total US deaths as of yesterday: 12,857
Increase in deaths since previous day: 1,914 (vs. 1,323 yesterday)
Percent increase in deaths since previous day: 17% (vs. 14% yesterday)
* Projected deaths = previous day’s new deaths number, grown by that day’s 3-day percentage growth rate – equal to 52%. Note this is calculated completely differently from the deaths set in stone, which refers to projected deaths over the entire pandemic, and is calculated by multiplying expected cases by the mortality rate

Date on which the number of new deaths on that day will probably exceed the toll of Sept. 11: April 12
Projected* number of actual deaths on 4/15 alone: 4,646
Projected* number of actual deaths on 4/22 alone: 15,530
Number of deaths on that day on April 30: 37,967


III. Reported case mortality rate so far in the pandemic in the US:
These numbers are also unaffected by the error I made.
Total Recoveries in US as of yesterday: 21,711
Total Deaths as of yesterday: 12,857
Deaths so far as percentage of closed cases (=deaths + recoveries): 37% (vs. 36% yesterday) Let’s be clear. This means that, of all the coronavirus cases that have been closed so far in the US, 37% of them have resulted in death (compare that to the 4% mortality rate I’ve been using to calculate total pandemic deaths, based on total cases). Of course, this number will come down as time goes on and more cases are closed in which the victim recovered. But it’s only come down by about 4 percentage points since Worldometers started publishing the recovery rate on March 26, and on about half the days, it’s gone up; there is still no sign of a downward trend. I’d say it’s much more likely my 4% mortality rate will turn out to be too low, after the pandemic’s over and all of the bodies have been counted, than it will be too high.


You may have noticed that the 3-day rate of increase in total cases (which drives total expected pandemic deaths) fell from 32% to 29% yesterday. That’s of course good news, but it only results in the milestone dates for total pandemic deaths being pushed back a few days. For example, the date on which one million deaths will be set in stone is just 4 days later than the one I calculated yesterday.

But you may have noticed that the 3-day rate of increase in reported total deaths rose from 48% to 52%. This resulted in daily deaths milestones being moved forward a day or so, and it also increased the estimate for total deaths on April 30 alone to 37,697. Of course, since the daily deaths numbers are estimated completely differently from the total pandemic deaths set in stone numbers, there’s no reason why the two should move together.

More importantly, there’s every reason to believe that the total cases number (and therefore the total pandemic deaths set in stone number) is a woefully low estimate, because of the continued drastic lack of availability of tests, and the long wait times to get a result once the test is administered. But the coronavirus doesn’t just kill people who have been tested and counted as infected – so deaths provide a much better “count” of actual infections than the reported numbers do.

However, it’s likely that even the official total deaths figure is an undercount of actual Covid-19 deaths because, given the shortage of tests, it’s highly unlikely that somebody who dies at home, and was never tested, will be tested after death. So their death won’t be counted as due to Covid-19.

But the really big problem with the total deaths numbers is that total deaths identified as due to Covid-19 is almost certainly a serious undercount of total deaths caused so far by the pandemic, for several reasons:

  1. There’s the well-known fear that there will be a big shortage of ventilators and hospital beds, which is already beginning to be realized in New York City and will almost certainly become much more critical in the coming two or three week (although once we get to tens of thousands of deaths nationwide every day, which will happen in a couple of weeks, it will almost certainly begin to manifest itself in other cities as well).
  2. For at least the last few weeks, EMT’s in NY City have been triaging patients at home. People were very worried about hospitals having to reject patients because they were full, but the reason this isn’t happening (much) yet is that the EMT’s – who of course aren’t doctors – are making decisions on the fly as to which person might be saved and which is too far gone to save. So they never even reach the hospital. And this is true for all diseases.
  3. For heart attacks specifically, for a couple weeks in NYC there’s been an explicit policy that if an EMT can’t detect a pulse in someone who is suspected of having a heart attack, they aren’t to take them to the ER. Normally, the person would be rushed in, while the EMTs tried to resuscitate the person in the ambulance, and the doctors would continue to do that in the ER, until they were satisfied there was no hope of reviving the person. The fact that non-doctors are being forced to make instant life-or-death decisions every day is just horrible to contemplate.

Perhaps worse than all of these reasons, there was a story in the NY Times today (and by the way, they are doing an outstanding job of covering the pandemic, perhaps because NY is the epicenter now, both for the US and the world). The article (I read it online a couple hours ago, but I can’t find it now) was by a doctor in a hospital in I believe NYC, who was mystified by the fact that patients with non-Covid-19 problems, especially cardiac patients, were very hard to find at his hospital, even though they’re not full yet.  He talked to colleagues in other hospitals and locations like New Jersey, and they had observed the same phenomenon.

Obviously, he doesn’t know the reason, but he could easily guess at it: Even though a patient might suspect he or she is going to have a heart attack or a stroke, they’re now very reluctant to go into a hospital, since they believe it’s very likely they’ll end up acquiring Covid-19 there, and perhaps dying as a result of that. They would rather stay home and hope for the best. The doctor points out that this is definitely the wrong attitude, since the hospitals are able to protect the patients – assuming they finally have enough surgical masks and gowns, which as we all know might not be the case in some places.

Of course, the problem is that, if the person who stayed home loses the gamble and their situation rapidly worsens, they will die too quickly for them to reach the ER. But their deaths will never be counted in the total coronavirus deaths numbers, because that’s not what they died of. Still, daily deaths from the coronavirus is a much better predictor of the ultimate death toll than are figures that are based on total cases, like the set in stone numbers above.

I’ve got a simple statistical test I’m going to perform tomorrow (I hope), that might help quantify how big the delta is between the two numbers. But what I’ve discussed above shows that even the daily deaths numbers are already underestimating the true count of deaths caused by the pandemic.

No comments:

Post a Comment