Per-Protocol Analysis

In Detail

Per-protocol (PP) analysis is a statistical methodology that includes only those participants who completed the clinical trial in compliance with the protocol — meaning they received the full course of treatment, attended all required visits, met all eligibility criteria, and did not have major protocol deviations. Participants who dropped out, missed doses, switched treatments, or violated key protocol requirements are excluded from the per-protocol analysis. This approach provides an estimate of the treatment's effect under ideal conditions — what happens when patients actually take the drug as directed. Per-protocol analysis typically shows larger treatment effects than intention-to-treat analysis because it excludes "noise" from non-compliant participants. However, this exclusion can introduce bias because the reasons participants drop out or deviate from the protocol may be related to the treatment itself. For example, if participants drop out because of side effects, excluding them would make the treatment appear safer and more effective than it actually is in the broader population. For this reason, regulatory agencies consider ITT analysis the primary analysis and treat per-protocol analysis as supportive or sensitivity analysis. The two analyses are expected to be concordant — if the ITT analysis shows the treatment works and the per-protocol analysis confirms it, the evidence is considered robust. If the results diverge, it raises questions about the reliability of the findings. Per-protocol analysis is particularly useful for non-inferiority trials, where it provides a more conservative estimate (the opposite of its role in superiority trials). Understanding both ITT and PP analysis helps patients and physicians evaluate the strength of clinical trial evidence.