Paper summary: risk of blindness from blood clots more than doubled amongst people who took the vaccines, with a particularly sharp effect for AstraZeneca. Most

Paper summary: risk of blindness from blood clots more than doubled amongst people who took the vaccines, with a particularly sharp effect for AstraZeneca. Most risk concentrated in the first two weeks, rest spread out over the first twelve. Paper assessment: not bad work overall. The usual problems such papers acquire in the process of getting past the gatekeepers are present, as expected, but can be ignored when you know what to look for. The good: Very large sample size, cohorts aligned according to many important variables. Goes well beyond just age/gender matching and includes comorbidities, medications. The bad: They do a lot of exclusions. Is it reasonable to exclude everyone with a history of RVO from the analysis? You'd imagine that people prone to it already might be affected worse and that this would matter for real world outcomes. Also, the paper shows signs of pre-publication tampering by someone other than the authors. The conclusions are dishonest and full with logical fallacies. But it's hard to blame the authors for this. If you don't include the usual bromides and deceptive comparisons then you can't get this sort of paper published. People who Follow The Science will draw the wrong conclusions as intended, the rest of us can incorporate it into our priors for future decision making purposes (when we're allowed).