This covers the trials not being fully representative, but largely neglects why that is the case.
The paper defines a population "at high risk of drug-induced serious adverse events", which presumably means they're also the most likely people to be harmed or killed by the drug trial itself.
replies(2):