covid, diabetes
Amusingly, what makes me believe this result more is the population statistics (of a larger than the preexisting trend would imply increase in t1 diagnosis rate in children): otherwise I'd be very suspicious of the lack of attempt to consider that onset of diabetes is not instantaneous (so having more exposure to the family doctor will pull the time of diagnosis earlier; I couldn't easily find how quick it is) and the somewhat confusedly described attempt to exclude cases where the diagnosis of diabetes was first (reading between the lines, they seem to have 3mo resolution and there's at least a bit of a reason to expect exposure to COVID to be higher around the time one is diagnosed with diabetes, because that involves lots more contact with people).
re: covid, diabetes
Yeah, this is the part that convinces me more. I found that amusing, because usually an attempt to actually study a correlation on an individual level instead of something related to "pitacy and ocean temperature are correlated" is more convincing, but here is the other way round.
re: covid, diabetes
@timorl or rather, the trend convinces me more than the correlation
re: covid, diabetes
@robryk Kinda importantly this trend continues, I learned about this not from a study, but from reports from a hospital, where pre-pandemic they had ~1-2 new diagnoses of diabetes/week and now it’s more around 1/day. If it was just exposure to doctors (which I’m not convinced is that much higher due to COVID anyway) the trend would’ve subsided by now.