Tagged statistics

LIES, DAMNED LIES, AND STATISTICS

Mark Twain by A.F.Bradley ©       Benjamin Disraeli  Photos.com/Jupiterimages ©

“There are three kinds of lies: Lies, Damned Lies, and Statistics.” The phrase is often claimed to be one of Mark Twain’s quotes, who, in fact, attributed it to Benjamin Disraeli, who served Queen Victoria as Prime Minister of the United Kingdom.  This quote is not found anywhere in Disraeli’s writings, although he may have said those words. Regardless of who said it, the sentence argues that statistics are often the worst kind of misinformation because they have the authority of numbers; the sources are hard to pinpoint and dispute and therefore are believed.

I fell into this trap recently and exposed my ignorance by quoting the Coronavirus mortality statistics, as given in large part, by the Chinese government.  The clue to me should have been “Chinese government.” The number of deaths I saw in the news was 3,132 deaths of a total of 92,303 (the newest numbers 3,886 deaths and 111,650 total cases). Nevertheless, mortality rates remain close to 3.4%.  The number of people that died is probably accurate.  It is hard to fake death, and counting the dead is likely to be reasonably precise. If they don’t move, they are probably dead.  The denominator has grown to 111,650 people that have the disease and are almost certain to be undercounted.  Mild cases are frequently not diagnosed or reported.  A runny nose is a runny nose, not COVID 19, in most people’s judgement, but then again, it could be.  The denominator is likely a much larger number judging from statistical analysis of previous epidemics. This would significantly revise mortality downward.

That, however, is not the essential source of misleading information. To accurately estimate disease mortality, raw numbers give the wrong impression.  In the seasonal flu mortality, we see every year, and the numbers are more accurately reported remaining under 1% because:

  • We are in the USA
  • Our health reporting is mandated by governmental agencies (not necessarily a strong point)
  • We try to be scientifically as honest as we can with checks and balances that report dishonesty or ulterior motives for inflating or deflating numbers
  • The people in charge of reporting don’t have quite the hidden motives or pressures as an autocracy generates, who initially punished reporting physicians
  • Our system has multiple sources of input that act as an automatic control mechanism
  • Our math adds the mild cases and estimates of asymptomatic people to the calculations (although this has to be retrospective)

Almost all comorbidities affect mortality rates: heart disease, chronic lung disease, renal failure, liver failure, obesity, neurologic disease, diabetes, cancer, immune deficiency, pregnancy, age, etc.

Without knowing the demographics of the population under discussion, raw data mortality rates are almost meaningless.  In the 2017-2018 flu season, which was one of the most severe epidemics in the last decade, if you were a healthy person, not pregnant, under 65 years of age, your risk of death was under 1%, but if you had a heart problem and were over 65 your mortality was over 10%.

How can we interpolate all this to the Coronavirus?  The 3.4% mortality rate is a raw number for many (but not all affected people); the mild ones didn’t even get counted.  I hate to use the term “fake news” because of its political overtones, but I think it is justified here.  The news media has “hyped up” this issue with inflated numbers for reasons that are suspicious to me.  To sell more newspapers? To be ahead in reporting doom and gloom? (which seems to be a habit of the news media culture). Or is it to affect elections? To create panic? To develop financial uncertainty? To…?  A recently published editorial in the most prestigious medical journal in the world, The New England Journal of Medicine written by Anthony S. Fauci (the nation’s leading expert on infectious diseases), H. Clifford Lane, and Robert R. Redfield, https://www.nejm.org/doi/full/10.1056/NEJMe2002387 has stated, and I quote, “the case fatality rate (of Covid 19) may be considerably less than 1% (the bold letters are mine). This suggests that the overall clinical consequences of Covid-19 may ultimately be more akin to those of a severe seasonal influenza (which has a case fatality rate of approximately 0.1%).”  So take everything with a grain of salt!