The Predictor wrote:
Thanks for the clarification. The idea I was trying to get across is that the patients should be separated in more specific groups in the studies and not just lumped together as a group of people who are all in the hospital for COVID. Example categories should be age, pre-existing conditions, estimated time 'til death, etc.
Maybe if enough studies like that come out we can rule out which groups triumph on x medication and which ones don't.
Generally in a large enough trial it's feasible to do this.
You have to be careful analyzing too many subgroups looking for an effect. If every subcohort you analyze has a change of turning up a false positive, the more you look at the more likely you are to get one. This is often abused by people trying to show a positive effect for their dead-in-the-water drug.
Lots of statistical and methodological ways around this.
Generally it's very expensive to run an RCT, especially for 1000s of patients.