Run-In Period

A run-in period is a defined stretch of time between screening and randomisation during which everyone gets the same thing. In a placebo run-in that is a dummy treatment, in an active run-in it is the study drug itself, and in a stabilisation run-in it is the background therapy alone. The point is to standardise the starting state, measure adherence, obtain stable baseline values, or identify who should not be randomised, all of it done before allocation so that neither arm is affected differentially.

Placebo run-ins are used to drop people who cannot take the treatment reliably or who improve on their own; the Physicians' Health Study famously randomised aspirin only after an 18-week run-in. Active run-ins are used where tolerability is the gate, including titration schedules that walk participants past the early gastrointestinal effects of an incretin agonist before the randomised comparison starts. Stabilisation run-ins are standard in trials with a background regimen that must be steady before anything is added.

Run-ins raise statistical power in two ways: fewer dropouts and less noise in the baseline, both of which shrink the required sample size. That gain is bought with enrichment. A cohort filtered for adherence and tolerance will show a larger effect and fewer discontinuations than the same drug used in an unselected population, so the design trades external validity for internal validity quite deliberately.

That trade is where the misreading happens. Adverse event and discontinuation rates from a trial with an active run-in describe a population from which the intolerant were already removed, so a clinic will see worse. Anything that happens during the run-in is also pre-randomisation and cannot be compared between arms, which does not stop run-in improvement being reported as though it were a treatment result.