[rsfern]

VAERS is for surfacing hypotheses about rare side effects

For serious risk quantification and causal analysis, we have things like the vaccine safety datalink, which links all electronic health records across a bunch of hospital systems, covering IIRC 3% of the US population. The UK has something similar I think. I think transparency in that system (VSD) could be better, but it has the same problem this thread is discussing, that anonymizing the records may be at odds with making the analysis reproducible.

https://pubmed.ncbi.nlm.nih.gov/34477809/