Propensity Score Matching: Role of Clinical Expertise in a Data-Driven Approach to Model Selection

ISPOR Europe 2022 -- Observational studies are not controlled and are subject to confounding factors that could influence outcomes. An open-label, non-randomized, prospective observational cohort study to assess post-procedural outcomes in two cohorts was conducted with the objective to provide further evidence of the value of clinical expertise alongside data-driven propensity score (PS) model selection in observational studies, as well as share learnings on process for model selection from a real-world study. This analysis provides further evidence to the added value of clinical expertise in PS matching model development and selection for controlling for bias in observational studies. Clinical expertise is important to instilling confidence in a largely data-driven approach, for the interpretation of the study's results in the medical community.