Analysis populations define various subgroups of study participants used for specific statistical analyses. The primary populations include Intention-to-Treat (ITT), Per-Protocol (PP), modified Intention-to-Treat (mITT), and the Safety Population. These definitions are critical for interpreting study results and must be prospectively defined in the statistical analysis plan prior to database lock.
The Intention-to-Treat (ITT) population includes all randomized participants and follows the “analyze as randomized” principle, regardless of protocol adherence or treatment completion. This provides a conservative estimate of the treatment effect and reflects effectiveness under real-world conditions. The Per-Protocol (PP) population excludes participants with major protocol deviations and provides an estimate of the treatment effect under optimal conditions (efficacy). Regulatory authorities typically expect both analyses for confirmatory studies.
The Safety Population includes all participants who received at least one dose of the study medication and is used for all safety analyses. A modified ITT (mITT) may include various meaningful restrictions, such as minimum exposure requirements or the availability of post-baseline measurements. CROs must implement robust processes for the definition, documentation, and quality control of these populations. Sensitivity analyses using different population definitions strengthen the robustness of study conclusions and address potential sources of bias.