Drowning in Coronavirus Research, Part I

Coronavirus research is an epidemic, growing exponentially, and in particular there has been a huge upsurge in modeling papers.  Everyone has a model, and I suspect everyone of them is worthless.  Anyway, I am trying to keep up and provide summaries of a few interesting ones.  We will start with a couple of stories that aren’t technical research papers.  First I ran across the following interview on March 30 with the Centers for Disease Control director.   (CDC Interview)   Here is his perspective at the time. The virus is going to be with us for a long time; it isn’t going away.  The virus is three times as transmissible or infectious as flu.  As many as 25% of people are asymptomatic.  (uh, guess again, try 50%, 75%, 90%, because that is what the evidence is suggesting)  Among symptomatic people, virus shedding has probably been occurring for at least two days before the symptoms appear.  He acknowledges that due to the lockdowns we have just delayed the spread, not stopped the virus.  He thinks seasonality will help, but for now that is a hypothesis that has fallen out of favor.  He thinks we use the summer months to prepare for a new wave in the fall.  He says the virus can’t go from person to person easily, it needs to be within 6 feet.  (uh, at this point, and even on March 30, this was just nonsense, it was pretty clear the virus could survive on surfaces and could remain airborne for much longer distances and times)  He says 99% of people will recover, but it is the elderly who are particularly vulnerable.  I think overall he is realistic about the virus not being eradicated, but unduly pessimistic about the actual rate of serious illness and death.

Next up, a story from the New York Times on March 13, on the atrocious (only of course the NYT doesn’t call it that) early modeling that drove panicked decision-making on shutdowns.   (NYT Story)    According to the story, the “modelers”, and if that isn’t a dirty word by now, it should be, presented four scenarios, which varied in assumptions about transmissibility of the virus and how serious the illness was.  50 expert teams were then tasked with modeling the outbreak.  These geniuses (idiots?) projected that 160 to 240 million Americans might be infected and 200,000 to 1.7 million could die.  And apparently most alarmingly to these “experts”, 2.4 million to 21 million people might need to be hospitalized.  So, let’s stop there and notice how useful those very small, limited ranges of projected hospitalizations and deaths are for decision-makers, and how the media, and the politicians just naturally gravitate to the highest number.  And of course, the story relates, the modelers relied heavily on Chinese data in developing their assumptions.  Seriously?  Why would anyone trust those numbers.  You read this crap and you understand how dangerous relying on experts can be in these situations, because they clearly can’t figure out to develop projections that are even close to reality, absurdly underestimating the true number of infections, and the percent of those infections that are asymptomatic, or lead to mild or moderate illness, and grossly overestimating the likely number of hospitalizations and deaths.

And right on cue, yet another story indicating that in fact many more people have been infected than is shown by infection testing and that these people are all asymptomatic or had mild illness.  (Boston Tests)   This test occurred in Boston and involved 200 people who did a finger prick test for antibodies in a neighborhood.  64 of these people or 32%, had antibodies to the coronavirus.  About half said they had a symptom of coronavirus in the past four weeks.  I have to say, it is getting to the point where it is undeniable that we have very, very many more people who have been infected than the official case count, and every one of those is a non-serious illness.

Finally, for this segment, under the too much sense category, is a column written by two health researchers at the Incidental Economist site.  (IE Column)   The column, called the economic cost of flattening the curve, is yet another call for recognizing that the efforts to mitigate the spread of the epidemic are not cost-free.  The authors say that the costs of the mitigation measures should also have been considered.  They used the “quality adjusted life year” to estimate this.  They first estimated the number of life-years being lost to coronavirus; not lives but life years, in other words how much life expectancy remained at the time of death.  They projected that to be 7.8 years on average. They used the usual wide range of total deaths from coronavirus in the US.  They then figured out what the estimated economic stimulus cost would be and came up with $4 trillion and a cost per QALY of $300,000 to $2.5 million.  By comparison, most health economists think a health intervention is worth it only at a QALY of $100,000.  I think a better figure is how much economic output is lost.  Right now that looks like at least 20% of the U.S. economy, but over not just a year but for a much longer time, although the percent of loss will gradually decline.  So let’s assume the first year it is 20%, then drops in half, then drops in half again.  That is around $8 trillion in lost economic output.  And you have to add to that all the loss of quality adjusted life years because of the lockdowns, the suicides and the decreased life expectancy from stress, etc.

What we are doing is simply insane.