At nVoq, we are very fortunate to get a weekly pandemic update from our CEO, Charles Corfield. In this state of unrest, Charles provides his analysis and gives us a few points to ponder. We will be sharing Charles’ blogs weekly on Mondays until things return to normal or at least, the new normal.
Pandemic Update Week of April 20, 2020 – by Charles Corfield This week brought us the first data on community prevalence. At one hospital in New York they tested all the folks coming into labor and delivery, and they found that 1 in 8 tested positive; but of those who tested positive, 90% were asymptomatic. Given short comings with tests, it is quite possible that the population from which these new mothers came has an infection rate > 20%. Meanwhile, researchers at Stanford University tested people in Santa Clara county for antibodies and concluded that 2.5-4% of the population had been infected, which they noted meant that actual number of infections is 50-85 times the official case count.There are a couple of surprises which have preliminary support. One research team looked at the role of smoking in Chinese hospitalizations and found a much lower proportion of smokers than in the population at large. For what it is worth, approximately, 50% of men have a history of smoking (fairly evenly spread over age groups), compared with only 2% of women; 5-8% of hospitalizations had a history of smoking, but you would naively expect 25%. It will be interesting to see if researchers in other countries find a similar protective effect. Another (pleasant) surprise comes from doctors in New York and an article in the Lancet, which note that asthmatics are under-represented in hospitalizations.There is an oft quoted phrase about lies, damned lies, and statistics, which is really a comment about how humans form and then buttress their points of view. We are not as logical/sensible as we would like to believe. The first step is an intuitive reaction as to whether something seems right or wrong, the second step is where we recruit “evidence” in support of that intuition and it is not surprising that we gravitate to the “evidence” that is easiest to find or comes quickest to mind. This is called “availability bias”.Social media platforms (and echo chambers) ensure that a lot of supporting information is available, but availability is not the same as accuracy. In the science world the way the bias shows up is through the converse, where studies, which did not find an effect, have not been published, or were relegated to obscure journals. Thus, when I search for scientific papers, I bear in mind that important (negative) data may be missing. A partial remedy to this is when results have been replicated by independent research teams.We can see the availability bias playing out in the pandemic with the much hyped (hydroxy)chloroquine, you do not need to work hard to find lots of case reports of the form “My husband took (hydroxy)chloroquine for covid-19 and it cured him”. These attestations become what is known as “proof by anecdote” and the more they are repeated the more “available” they become. More people come to believe that this the go-to treatment that everyone should use. However, we know that most people will recover anyway, whether you give them (hydroxy)chloroquine or slug-juice, or nothing at all. Therefore, anecdotal reports are interesting in identifying interventions which are worth a closer look; but before you can claim that something works, you have to run a rigorous, randomized, controlled, double-blinded study with enough subjects to be meaningful.For those who have been singing the praises of (hydroxy)chloroquine ahead of any rigorous study results comes the inconvenient observation from rheumatologists who have patients already taking (hydroxy)chloroquine: their patients get covid-19 too… Oops. We too need to be mindful of availability bias as we go about our daily work -- it is sure to occur, but it won’t be obvious, it requires that we pause a moment and ask ourselves “Is <X> really true? Or is this a case of something being oft repeated and thus a case of availability bias?”