Adherence and persistence

This is the gap between what a trial showed and what a population gets, and it is usually larger than the effect being argued about. Roughly a third of US adults starting a GLP-1 drug for obesity were still persistent users after twelve months; among those also using continuous glucose monitoring the figure was two thirds. In hypertension, medication de-escalated by protocol and not promptly restored tracked a third more cardiovascular events.

Reminders, monitoring, cost, side effects and the number of daily doses all move the figure, and none of them moves it far. A treatment nobody takes has an effect of zero, whatever its trial reported.

Status

building

Direction

mixed

Would change our view

An intervention that raises long-term persistence by a margin large enough to show in a mortality endpoint. Evidence that adherence measured from claims — prescriptions filled — tracks what is actually taken. A trial reporting its own adherence as a primary outcome rather than a footnote.

Trackers

Gaps

Almost all adherence data is dispensing data, which records collection rather than consumption. Persistence is rarely followed beyond a year. Very little is known about who stops and why, as against how many.