A statistical approach that includes only participants who completed a clinical trial according to the study protocol—received all planned doses, attended all visits, and did not violate any protocol rules. Participants who dropped out, missed doses, or switched treatments are excluded.

Per-protocol analysis answers “what happens when everything goes perfectly?” while intention-to-treat analysis answers “what happens in the real world?” Per-protocol results almost always look better than ITT results because they exclude the patients who tolerated the treatment poorly or found it ineffective. Both analyses are typically reported in well-designed trials; if a study reports only per-protocol results, the actual treatment effect is likely smaller than presented.