32,000 people in the study, of which 16,000 had previously had Covid.
They found a total of 238 re-infections.
Assuming 1% mortality from Covid, that group of 16,000 would have previously been 16,162, before being culled down to the study size by deaths.
You don't think an additional 162 would make a difference to the results?
(Obviously it doesn't make a difference if you read this study as "given someone survived Covid, what are the odds..." but lots of people are reading this study as "It's safer to catch Covid than get vaccinated." Given that, the deaths absolutely matter.)
> if you read this study as "given someone survived Covid, what are the odds..."
How else are we supposed to read it? I just don't understand what point is being made with "survivorship bias" here. It's not like individuals who have recovered from Covid have gone through some hyperselective winnowing process leaving only those with superheroic immune systems or something.
Obviously the study does not support the conclusion "it's safer to catch Covid than get vaccinated." However, for the millions of people who have had a confirmed infection and recovered, it is absolutely relevant to their risk calculus around whether to take the vaccine, and given that they are survivors, there is no "bias" to speak of.
You're assuming everyone read it the same as you, and now can't imagine differently.
There have been several mentions in this thread alone about how people have been using studies like this to recommend "Covid parties" (like chickenpox parties), or generally use it as further fuel to stay away from the vaccine, because they believe catching Covid is safer.
It's relevant because of the 1%. The numbers in the study are so low (8 people hospitalized in one, 1 person in the other group) that 1% makes a difference.
From the point of view that there is some risk involved from the first covid infection in order to get the immunity, so people will be less likely to correctly assess the risk of getting natural immunity vs getting vaccine in order to end up to the study's cohorts.
