What Is The True Level Of Mortality Caused By The Covid-19 Virus?

Several commenters on yesterday’s post express a strong interest in learning the true level of mortality from the Chinese Virus. Are the numbers reported at sites like Worldometers as “Coronavirus Deaths” accurate and reliable, or are they inflated? Or, for that matter, could it be that the number of “Coronavirus Deaths” is under-reported for reasons that could include people dying at home without ever being tested?

My answer is that it is impossible to know at this point. The best indication we will get will come when the CDC issues final data for deaths from all causes in the U.S. for the month of April. When we get that number, we can subtract from it the approximate “normal” number of deaths that would have occurred anyway during April. The difference will be a good estimate of the number of excess deaths attributable to the virus. My prediction is that that number will be far less than the number of “Coronavirus Deaths” being officially reported. I’ll bet on about half or even less; but I’m the first to admit that I could be proved wrong.

The problem with attributing deaths to the virus begins with the fundamental problem of all scientific endeavor, which is that events in the real world have not just one but many causes. As a common example closely related to the subject of this post, many patients with cancer get pneumonia at the final stage of their disease. Did they die from the cancer or from the pneumonia? Or from both? When you look at the CDC statistics, you will see that each death has been assigned uniquely to just one of the major categories. In this case, each death has been assigned either to cancer or to pneumonia, not both. But how did someone decide that death A was from cancer, but death B was from pneumonia, when the patients had both and were in terminal condition? Generally, there is not much riding on the decision, and if it is made arbitrarily — half to one, half to the other — that’s probably fine.

This time, people very justifiably want to know if some $3 to 5 trillion (or more) of our economy has been lost for a good reason, versus from panic stemming in large part from artificially inflated statistics.

With Covid-19, the problem of multi-causation deaths is particularly severe because the virus is known to be associated with many so-called “co-morbidities,” including cancer, heart disease, hypertension, diabetes, and others. Several commenters yesterday noted that the government has put in place strong financial incentives for ambiguous or multi-cause deaths to be assigned to the Covid-19 virus as “the” cause. Commenter mc states:

 “$$$ 39,000 to hospital if listed as CV death or died with CV”

If true, that would be a very powerful incentive, to say the least, to assign every ambiguous or multi-cause death to the virus. Is there really such a powerful incentive? Apparently, some YouTube videos have been circulating touting the $39,000 figure. FactCheck.org (a project of the Annenberg Center) has a piece on April 21 on this issue by Angelo Fichera. Here is the question and answer:

Q: Are hospitals inflating the number of COVID-19 cases and deaths so they can be paid more?

A: Recent legislation pays hospitals higher Medicare rates for COVID-19 patients and treatment, but there is no evidence of fraudulent reporting.

Even the phrasing of the question and answer shows that the people at Annenberg either don’t understand the issue or are intentionally misconstruing it in order to misdirect the audience. The issue here is not really whether there is “fraudulent reporting.” There may well be some of that, but even if there is none of it, there remains the question of whether some thousands of deaths that involve other concurrent causes and would have occurred anyway are getting reported as deaths “from Covid-19” because Covid-19 was present and the hospital got a big check for reporting in that category, when it would not have gotten the check if it picked some other category. Reporting of deaths as “caused by” Covid-19 when co-morbidities are also present could be completely legitimate in the sense that the virus contributed at least in some degree to the death; but at the same time, the financial incentive could lead many multi-cause deaths to get the Covid-19 label, thereby hugely swinging the statistics to make it falsely appear that the virus was much more dangerous than it actually was.

When FactCheck.org gets to the simple question of whether there is a strong financial incentive from the federal government to the hospital for reporting a death in the category of Covid-19, they basically confirm that the incentive is present:

It is true . . . that the government will pay more to hospitals for COVID-19 cases in two senses: By paying an additional 20% on top of traditional Medicare rates for COVID-19 patients during the public health emergency, and by reimbursing hospitals for treating the uninsured patients with the disease (at that enhanced Medicare rate . . . The CARES Act created the 20% add-on to be paid for Medicare patients with COVID-19. The act further created a $100 billion fund that is being used to financially assist hospitals — a “portion” of which will be “used to reimburse healthcare providers, at Medicare rates, for COVID-related treatment of the uninsured,” according to the U.S. Department of Health and Human Services.

This is not the same thing as the purported $39,000 premium for reporting a death as caused by “Covid-19,” as cited by commenter cm. (The $39,000 figure appears to derive from average Medicare reimbursement for a patient who goes on a respirator in a hospital; but that is irrespective of whether the patient does or does not have Covid-19.). Still, the premium payments confirmed by FactCheck constitute a strong financial incentive that could very well significantly swing the statistics as to causes of death.

So what should we be looking for to get a handle on excess deaths, which is the real measure of the impact of the disease? According to CDC data here from 2019, the U.S. averages about 7,700 deaths per day. There are more deaths per day in the winter, and fewer in the summer. April is right about at the average. That means that in April, a normal level of deaths for the month would be about 231,000.

As of yesterday (April 26), Worldometers is reporting that the U.S. has had 56,527 deaths from the virus, of which 5,210 were in March. That means about 51,000 Covid-19 deaths in April, with four days to go. Figure several thousand more deaths before the month ends. So the official reporting will say that some 55,000 to 60,000 deaths in April were “caused” by the virus. But will the U.S. really have 286,000 - 291,000 deaths in April, in other words, 55,000 - 60,000 more deaths than last April, when there was no such virus and only about 231,000 deaths?

I’m putting my bet on a bump of additional deaths of around 20,000, and no more than 30,000. Let’s see if I’m right. 250,000 deaths for April is my number. Readers are invited to place their own bets.

UPDATE, April 28: Here is a link to the most recent data for total deaths in the U.S. by week, rather than by month, from the CDC. The most recent week for which they are reporting complete data is what they call 2020 Week 15, which appears to be the week of April 5 to 11. They report 55,905 total deaths that week. If the normal number of non-coronavirus deaths per day in April is 7700, that would mean that a normal non-coronavirus week would be 53,900. So the number of “excess deaths” in that week is only about 2000. Yet Worldometers reports the number of Covid-19 deaths that week as 13,678.

For the previous week (2020 Week 14, corresponding to March 29 - April 4) CDC reports 60,324 deaths, or about 6,400 above the norm of 53,900. The Worldometers figure for Covid-19 deaths for that week is 8,350.