Meta-comment about medical studies. Suppose that an intervention's effect is contingent on some variable. For high-profile treatments, wouldn't it make sense to

Meta-comment about medical studies. Suppose that an intervention's effect is contingent on some variable. For high-profile treatments, wouldn't it make sense to isolate those variables (for the cases where it did help, but the overall result over the entire population was neutral) and run a second trial?