3 ms·

But this is exactly where we should double our efforts, because the problem is technical: we do have lots of data to analyze already. The influence of anti-his

But this is exactly where we should double our efforts, because the problem is technical: we do have lots of data to analyze already. The influence of anti-histamine drugs was relatively quickly discovered because everybody expected asthma patients would get worse and the opposite was happening, so this was an obvious thing to examine. But many more drugs may play a role and these are hiding in plain sight - I hope researchers try to organize around it and make more meaningful discoveries.

SketchySeaBeast 5y ago
I'm not saying it wouldn't be a good idea if it was possible, but do we have enough sanitized, high quality, and complete medical data to pull that off? Given the number of potential interactions with all the different medications and conditions, could we actually pull out those links without a ton of confounding variables? Would it produce signal, or just noise?
- hdjjhhvvhga 5y ago
  Given enormous resources spent on everything COVOD-related, from tests and vaccines to the costs of lockdowns and restrictions, I think we should at least try to analyze the data looking for the low-hanging fruit, then try to account for confounding factors, and if still interesting look at it closer. Not easy, but definitely possible for large hospitals - they do have enough data, they just need competent people to analyze them and make preliminary findings. On the other hand, we are probably most interested in patients who got infected but did not end up in a hospital. Were they taking any meds routinely? I don't think it's easy to gather this kind of data unless people share them voluntarily with an organization they trust.
  - s1artibartfast 5y ago
    I think you are underestimating the the challenge, but I like your optimism. If states like California can hire 20,000 contact tracers, it seems plausible to hire 20,000 people to transcribe medical records into a database for analysis and remove personal identification.

SketchySeaBeast 5y ago

I'm not saying it wouldn't be a good idea if it was possible, but do we have enough sanitized, high quality, and complete medical data to pull that off? Given the number of potential interactions with all the different medications and conditions, could we actually pull out those links without a ton of confounding variables? Would it produce signal, or just noise?

hdjjhhvvhga 5y ago
Given enormous resources spent on everything COVOD-related, from tests and vaccines to the costs of lockdowns and restrictions, I think we should at least try to analyze the data looking for the low-hanging fruit, then try to account for confounding factors, and if still interesting look at it closer. Not easy, but definitely possible for large hospitals - they do have enough data, they just need competent people to analyze them and make preliminary findings. On the other hand, we are probably most interested in patients who got infected but did not end up in a hospital. Were they taking any meds routinely? I don't think it's easy to gather this kind of data unless people share them voluntarily with an organization they trust.
- s1artibartfast 5y ago
  I think you are underestimating the the challenge, but I like your optimism. If states like California can hire 20,000 contact tracers, it seems plausible to hire 20,000 people to transcribe medical records into a database for analysis and remove personal identification.