Following the protocol versus following the patients
Intention-to-treat (ITT) analysis includes every enrolled participant in the arm they were randomly assigned to, regardless of whether they actually took the medication or completed the intervention. Per-protocol (PP) analysis counts only those who adhered to the assigned treatment.
ITT requires analyzing more participants in an arm than actually received that treatment, which dilutes the apparent effect but preserves the randomization. PP analyzes a more compliant subset, potentially inflating the effect size because people who adhere to treatment are often healthier or more motivated than those who don't.
When ITT and PP diverge
High dropout or non-adherence rates expose the gap between ITT and PP estimates. If a trial claims drug X is effective but ITT shows barely any benefit while PP shows substantial benefit, that divergence signals that the trial works only for adherent participants. The ITT result is closer to real-world effect because most patients will drop out or skip doses occasionally.