Friday, July 03, 2020

Part of the Covid Mortality Puzzle

We would like to know  the mortality rate, the percentage of people who get the disease who die as a result. That requires two numbers: number infected and number dead as a result.

The best estimates of number infected come from seroprevalence studies, testing a random sample of people for evidence that they have had the disease. There has been a good deal of discussion of the results, which generally indicate a much larger number, up to ten times as large, as the number of recorded cases.

I have not seen a careful discussion of problems with the other number. Under current circumstances, I think it unlikely that many deaths in developed countries will be missed, since the symptoms that suggest Covid are obvious and tests available. But we might err in the other direction, by counting deaths due to other causes but occurring to people infected with Covid. As a number of people have pointed out hospitals have a financial incentive to classify a death whose cause is ambiguous as due to Covid.

How large could that effect be?

I start with a high value of number infected in the U.S. — ten times the number of known cases, or 28 million. Assume, for simplicity, that those infections were spread out evenly over a period of four months and that each would test positive for a month. Then the number who would test positive at any given time was 7 million, or about 2% of the population.

The  U.S. mortality rate from all causes is 2,813,503 per year. I want to know how many of those would have died from causes other than Covid, over a four month period, while infected, so I multiply by .02 for number infected, divide by 3 for a third of a year of mortality. The result comes to about 19,000.

That should be an upper bound on by how much we are overcounting covid mortality by classifying deaths of people infected with Covid but due to other causes as due to Covid. The current count for total deaths in the U.S. due to Covid is 131,000. Assuming I haven't made any mistakes in my very rough calculation, that might be too high, but not by very much.

Readers are welcome to point out errors in either the logic or the numbers I am using, remembering that what I am looking for is an upper bound on the overcount.


Christoffer Hansen Nilsen said...

Are you not assuming that lockdowns wouldn't lower the mortality of everything else, for example the flu? If we use your method here in Norway, if I understood it correctly, we would conclude that corona has had a negative death count. We in Norway would probably see a mortality spike when the corona is over as the elderly start getting infected by other diseases when the lockdown is done.

Rebecca J said...

Not only are most past cases never recognized during the infection period, but also most (probably 80%) of these recovered infections are not identified by the current antibody tests. See this article:

Tom Hudson said...

At my workplace, the response to your "I think it unlikely that many deaths in developed countries will be missed, since the symptoms that suggest Covid are obvious and tests available" would be "Citation Needed".

That claim is inconsistent with my observations of the US, the UK, and the news I read from overseas; if it's incorrect, I believe the rest of your logic is undermined.

Excess deaths are high: Santa Clara County, which I believe is near you, had a 20% excess death rate in March. Pneumonia deaths are high, several times normal rates in many impacted areas. Unexplained deaths alone at home are high in impacted areas. Most of these deaths are not being attributed to COVID but there's a really strong correlation and I've not seen another explanation proffered.

SB said...

I'm not convinced by "the symptoms that suggest Covid are obvious and tests available" either.

First, "the symptoms that suggest Covid" have been changing (mostly expanding) from month to month: at first it was viewed as a respiratory disease, then people noticed kidney complications, then blood clotting and vascular damage, then whole-body general inflammation (in children), then nerve damage. It seems to me quite possible that a good fraction of the people who have died of Covid were never recognized as such at the time. As Tom points out, there have been a lot of excess deaths from other causes in heavily-impacted areas: some of these were probably undiagnosed Covid cases, while others were probably people without Covid going untreated because of Covid hospital overload or people reluctant to go to the hospital for fear of Covid.

Second, tests were decidedly not available in the US for the first few months, and are still scarce in some places. If a doctor has the choice between testing a current patient or a recently-deceased one, the former provides more actionable information. Other developed nations with greater test availability may provide better data in that sense, but those nations have had far fewer cases in total so their N is much lower.

John Schilling said...

"Are you not assuming that lockdowns wouldn't lower the mortality of everything else, for example the flu?"

Not sure about Norway, but here in the United States influenza is responsible for less than one percent of deaths in a normal year. You could reduce that to literally zero, and in this pandemic we wouldn't notice. Non-COVID infectious diseases in total, are responsible for maybe 5% of US deaths, and a good chunk of that is "very frail old person goes to hospital, contracts an infection that kills him faster than whatever he went to the hospital for".

Most US deaths are due to heart disease, cancer, strokes, Alzheimer's, and diabetes. The pandemic and lockdown aren't going to reduce those. Accidents are about six percent of the total; there will probably be some reduction there, but tiny compared to COVID mortality. Suicide is about 2% of the total in a normal year, and that's going up.

Seriously, I expect that fear of and response to the pandemic is going to increase, not decrease, non-COVID mortality. I've seen reports that suggest a very substantial increase in suicide rates in the Bay Area at least; I was trying to get better numbers or at least field reports on that from SSC before it went down. Opioid overdose deaths were up 42% in May, but on an apparent exponential curve with a roughly two-month doubling time. And mortality from boring old heart attacks, etc, is almost certainly increasing as victims are afraid to go to the hospital.

How many people are going to die of these secondary effects, I don't know. But they're dying of things that we normally take a year or so to collect and distribute statistics on, whereas we're now getting daily updates on COVID-19, as if nothing else matters. So I expect people to look at the excess mortality and say "that must be undiagnosed COVID-19; clearly we are not doing enough to fight COVID-19", rather than "maybe those people are dying because of the things we are doing to fight COVID-19".

Douglas Knight said...

Seroprevalence and excess death studies have high background rates and are only possible in the worst disasters, like Lombardy and NYC.

You assume that obvious cases will be counted. This is manifestly false. NYC had 18.5k deaths with a positive Covid test and another 4.5k that were obvious but not tested. But this latter category was tabulated and included only after tremendous pressure because all eyes were on NYC. As I understand it, similar deaths are excluded in the rest of the country. Thus I think -20% is a safe upper bound on the overcount.

Then we can turn to excess deaths. For March and April, the ratio of probable to confirmed deaths was even higher and the total confirmed+probable Covid was 78% of excess deaths. I believe that the excess deaths are due to Covid and thus the undercount is even higher.

michael choe said...

Almost half of deaths are nursing home residents. Majority of deaths are age 80 or higher with a preexisting condition. How many would have died this year it covid didn’t exist?

mister bitcoin said...

Diamond Princess covid data from March 16

20% infection rate
1% case fatality rate

Most of the passengers were in the 60s and 70s.

Median age in USA is 38.

Median age on Diamond Princess was 65.