Drug trials: incidence of side effect lower than background

[StringJunky]

How is it to be interpreted in drug trials when the incidence of a condition in a trial is less than in the population? For example, myocarditis in covid-19 vaccines. Someone brought this up elsewhere, and I wondered.

[Prometheus]

Depends on the phase of trial and the trial design. In phase 1 and 2 you're recruiting predominantly young healthy volunteers so you'd expect less adverse outcomes due to chance. By phase 3 you should be recruiting participants that reflect the drugs target population, but people likely to have adverse reactions are screened out.  

[StringJunky]

23 minutes ago, Prometheus said:

Depends on the phase of trial and the trial design. In phase 1 and 2 you're recruiting predominantly young healthy volunteers so you'd expect less adverse outcomes due to chance. By phase 3 you should be recruiting participants that reflect the drugs target population, but people likely to have adverse reactions are screened out.  

So, if the myocarditis cases are screened out, does the drug get the blame for those found in the trial?

[Prometheus]

Again depends. Statisticians would compare incidence in the active and control groups, accounting for type 1 error inflation. Doctors would follow up and assess the severity and likelihood of causal link (which is often subjective - it helps if there is some known or  plausible causal pathway).

But are the myocarditis cases detected during post marketing surveillance rather than clinical trials? So the adverse event wasn't detected in the clinical trials but only when it was released and used by millions. This happens with particularly rare adverse events.