4 ms·

What is an n=1 study?

• dekhn 5y ago
  This paper is based on a single patient. Normally, in any sort of health study, you want to work with a large population for statistical reasons. However, for many practical reasons (such a very rare event like this) you can't do that. While there are statistical techniques to deal with small n studies ("If nobody dies, did anything go wrong?" from Glantz's stats book), mainstream science never makes a strong assertion (as this paper does) of causality. My read of the paper does not find any evidence truly supporting their assertion, but this isn't really uncommon, doctors publish case reports like this all the time with the hope that they're the first person to actually report on a real problem (and hope all the other times they reported are conveniently ignored). And, of course, as a truly honest scientist, I will say: it's entirely possible they did find a situation where it occurred. It may be that all my complaining is actually against a legitimate problem. But, to be honest, when all the data is evaluated holistically, it is unlikely that this is a major concern. And the real problem is that somebody is going to go share this with some website that publishes "COVID vaccine causes hepatitus" and then a whole bunch of people won't get vaccinated.