Individual Participant Data Meta-Analysis

The IPDMA approach enables measure harmonisation, participant-level predictor modelling, consistent missing-data handling, and better-powered moderation estimates. IPD-MA can generate evidence unavailable from published papers or aggregate-data meta-analysis when patient-level data are dispersed across studies. Pooling individual participant data is presented as a way to improve patient-level specificity and statistical power compared with aggregate-data meta-analyses. Traditional aggregate-data meta-analyses are limited because they cannot directly test individual-level associations and are vulnerable to ecological bias and confounding. IPD-MA is often more robust than aggregate-data meta-analysis for subgroup, dosing, and heterogeneous paediatric questions. IPD-MA is generally faster and less costly than new clinical trials. A one-stage approach analyses all participant-level data within a single hierarchical model. An IPDMA pools de-identified participant-level data from multiple trials. IPD meta-analysis pools deidentified patient-level datasets and maps variables to common definitions for combined analysis. The study will use a one-stage IPDMA with Bayesian mixed-effects mode…