For reporting of side effects, is there any threshold for statistical significance?
That's a good question. I don't know. (But I know someone I could ask.)
From what I understand, anything that happens to any trial participants during the trial has to be reported. So even if there's no stated statistical cutoff, there would be a de-facto cutoff just because trials only have so many people in them. If the trial was thousands or 10's of thousand of people, but not millions, everything could have been done correctly and this could have still been missed.
It's also interesting to me that all of the people who experience blood clots are women.
Historically sex as a biological variable (SABV) has been ignored in research studies. Interestingly, most of the drug recalls in the last couple of decades have been because of women-specific side effects that weren't discovered during trials. Human trials are pretty even between men and women today, but animal trials (which is the first step for almost all research) typically uses male animals, simply because boy mice are cheaper to purchase than girl mice. Also most of the cell lines that scientist use for research are male cells. Awareness has been growing over this as a major oversight in science.
In the past few years the US National Institute of Health, the European Commission, and the Canadian Institute of Health Research have all instituted SABV policies. Some scientific journals have stopped publishing research where SABV isn't considered. So things are improving, but I'm curious if this is related to this specific J&J blood clotting issue.
Anyway, I went off on a tangent, but if you want to learn more.
https://orwh.od.nih.gov/sex-gender/nih-policy-sex-biological-variable
https://en.wikipedia.org/wiki/Sex_as_a_biological_variable