Dissecting COVID19 Statistics: what they really mean.

Everybody in the media seems to have become experts in epidemiology and statistics, talking about cases, R0, and mortality rates.  Here’s what all these numbers mean to you: Not much. Wash your hands, don’t touch your face, don’t freak out.  The best sources of practical information are at the CDC web site, and the DHS/FEMA “ready.gov” site. Essentially, these are common sense actions.  But, since a 100 word post just isn’t in my nature, here are a thousand or so more words on what we seem to know about COVID19 statistics from a public policy, economics, and emergency management perspective.

If you’re curious, my background on this comes from doing post-disaster disease spread and biological warfare studies.  As an aside, I’m perfectly comfortable around nukes or chemical weapons; I’m afraid of bio-agents, if that tells you anything.  I’m not an MD or epidemiologist, but I use data from those fields to create impact estimates. I posed an interesting question to someone who is an expert in those fields.  If we did not have the ability to do genetic testing on this virus and identify it as a new strain, would we have noticed it at this point?  The answer is complex, but it seems maybe not.  Let’s crunch some numbers.

Globally, on average there are 389 thousand deaths due to respiratory complications from influenza.  The 2009 influenza pandemic had an all-cause mortality rate of 9.2 per 100,000 in China, but was several times higher in other countries, 10 to 20 per 100,000.  The current mortality rate for COVID-19 in China is around 5.1 per 100,000 at worst.  So even if they are being deceptive over the total rate, it could be twice as high and still not terribly unusual.  The end result is that while they would never admit it for political reasons, it may well be that by the time the quarantines were put in place, it had already spread, and China damaged their economy (and the world for that matter) for no good reason.  Other countries are reacting to what China did as much as to the direct threat of the disease itself.  And those measures may well have a lot more to do with China’s internal politics than the virus.

What about those mortality rates – WHO says 3.4%, CDC said 16% in “older populations.”  Dr. Anthony Fauci, director of the National Institute of Allergy and Infectious Diseases, and colleagues wrote an editorial in the New England Journal of Medicine and said it was likely “under 1%.”  What should you make of that?  Nothing – see the first paragraph above 😛 !  The bottom line is all of these statistics are based on “cases” – and that definition is really sketchy.  Again, a mortality rate is dependent on two variables: the number of deaths (the “numerator” in math language), and the population at risk (the “denominator”).  We sort of know the numerator because people dying with the right symptoms are being tested.  But we have NO IDEA what the denominator is.  Only a few thousand people in the US have actually been tested.  Right now testing is so confused that we really don’t have good numbers.  But given what we know, and have seen in other countries including China, the rates are well below 1% – and are likely “just” fractions of a percent.

That’s not to say COVID19 isn’t a problem – but it is apparently on the same scale of threat (maybe even a bit less) than influenza.  And given the bias towards more severe respiratory issues for those hit hard by it, the health care community is right to be concerned about ICU resources.  But saying “on par with influenza” raises another issue: we really don’t take influenza as seriously as we should.  It kills tens of thousands of people every year, and we could prevent many of those by taking simple precautions.  Under 40% of people get flu shots.  How hard is it to properly wash your hands?  And don’t get me started on our society that encourages (even demands) that people go to work or send their kids to school sick.   If a normal flu season (much less a pandemic season like 2009 or even 2017) was given the publicity that the COVID19 is being given, people would be in a continual state of panic (or numb to it).

Which leaves me where I was two weeks ago: the panic will likely be far worse than the pandemic.  The economic damage being done, and that will likely be done, by politicians wanting to get out in front of public opinion and “do something” may be turning a manageable situation into a catastrophe.  To paraphrase something I heard astronaut John Young say, “there are very few situations you can’t make worse by doing something.”

Bottom line: We really do need to try to limit the spread, especially to vulnerable populations like those over 65 and those with health problems.  But we know how to do that.  It’s not glamorous, and doesn’t require massive government action, quarantines or wrecking our economy.  It’s very simple: good hygiene.

3 thoughts on “Dissecting COVID19 Statistics: what they really mean.

  1. Great reply. Retired Infection Control Nurse. Wash your hands, cover your mouth when coughing or sneezing ( preferably with a tissue, then properly dispose of it ) avoid crowds when possible and stay home when sick.

  2. I’m not a medical professional, but I do deal with data a lot. (And I am over 65.) I appreciate the effort to stick with data in trying to understand the effects of this virus. There appear to be at least 3 ways to define the denominator, based on the question we are trying to answer:
    1. Mortality rate of the virus among an entire population, whether at risk or not.
    2. Mortality rate of the virus among the at-risk population.
    3. Mortality rate of the virus among those who actually contract the virus.

    The foregoing article appears to address #2 above. In my view, the most relevant number is #3, because the denominators of #s 1 & 2 are not directly connected to the impact of the virus. Whereas, with #3, where we are most concerned about how many of those who get the virus succumb to it, both the numerator and denominator are relevant to the virus and its impact. Beyond that, you can further break that down into various demographic categories.

    In any case, when we look at numbers from various sources, consistent with this article, it is essential to understand which of the questions is being answered and what the denominator is defined as. And, above all, I applaud the effort to avoid panic and stick with the basics of personal hygiene, avoiding crowds, and staying home when sick.

    • Thanks for the comment. I think that in terms of overall societal impact, (1) is the most important. (2) and (3) are of supreme interest to the health care community; in one sense, they don’t care how many people catch something. They care a lot about how many show up at their door, and how sick they are. The other factor is the rate at which sick people show up. This is where COVID19 is such a dilemma. The fraction of (1) is smaller than the seasonal flu; the fraction of (3) is the about the same. But the rate will cause a spike higher than normal.

      This all points out a key problem: there is no one number that captures everything everybody needs to know. Using the “worst” number for all purposes can cause serious problems, just as using the “best” number can. Right number for the right job! Again, thanks!

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.