What Works for Whom in Pediatric OCD: Description of Causally Interpretable Meta-analysis Methods and Report on Trial Data Harmonization

Improving patient outcomes will be enhanced by understanding “what works, for whom?” enabling better matching of patients to available treatments. However, answering this “what works, for whom?” question requires sample sizes that exceed that of most individual trials. Conventional methods for combining data across trials, including aggregate-data meta-analysis, suffer from key limitations including difficulty accounting for differences across trials (e.g., comparing “apples to oranges”). Casually interpretable meta-analysis (CI-MA) addresses these limitations by pairing individual-participant-data (IPD) across trials with advancements in transportability methods to extend causal inferences to clinical “target” populations of interest. Combining IPD across trials also requires careful acquisition and harmonization of data, a challenging process for which practical guidance is not well-described in the literature. To fill these gaps, we describe methods and work to date for (1) harmonization of the largest IPD dataset available from 28 randomized controlled trials for pediatric obsessive compulsive disorder (OCD) treatments, along with target data from a clinical sample of treatment-seeking youth ages 4-20 with OCD and (2) the application of CI-MA to facilitate an understanding “what works, for whom?”. We also detail dissemination strategies and partner involvement planned throughout the project to enhance future clinical utility of CI-MA findings.