Participants who drop out of a study before it ends, before outcomes can be measured.
Loss to follow-up refers to participants who are enrolled in a study but do not complete it, so their outcome data is missing at later time points. High or differential loss to follow-up (where dropout rates or reasons differ between groups) can bias results, as those lost may differ systematically from those retained. As a rough guide, loss above 20% is often flagged as a threat to validity, though no universally agreed threshold exists.
This guide was auto-drafted and is pending editorial review.