Exploring physician assistant data sources

Objectives: To assess four physician assistant (PA) proprietary datasets and inform researchers about data quality for addressing healthcare policy and workforce questions. Methods: The quality of datasets was assessed by experienced researchers. Descriptive analysis included overview, collection methodology, variables, and availability. Assessment included each dataset's strengths and limitations. Results: Datasets from the American Academy of Physician Assistants, National Commission on Certifi cation of Physician Assistants, Physician Assistant Education Association, and Optum Provider360 Database include overlap in variables refl ecting organizational mission and/or design. Attributes include variables for validation; limitations were lack of public use fi les, requirements for specifi c data requests or data purchase. The datasets do not have unique identifi ers and cannot easily be linked. Conclusions: The PA datasets contain variables of interest but are limited in scope. Better data collection and shared platforms could further the understanding of PA workforce characteristics and contributions to American healthcare. Researchers await more comprehensive, longitudinal, linked, and publicly available datasets.