Medicaid is the single largest source of health care coverage in the US. However, health policy research on Medicaid has historically lagged due to the limited availability of high-quality administrative claims data across states.
In 2019, the US Centers for Medicare & Medicaid Services released the T-MSIS Analytic Files (TAF), a federal Medicaid claims dataset that has catalyzed policy-relevant research on the Medicaid program. TAF data quality represents a significant improvement relative to the quality of previous Medicaid claims datasets. Nonetheless, variation in data quality across states can introduce nontrivial bias and measurement errors.
To ensure that research is high quality, credible, and interpreted and applied appropriately, Dr. William Schpero, assistant professor of population health sciences, and colleagues have developed the TAF Analysis Reporting Checklist, described in an article in JAMA Health Forum. The Checklist was developed through the Medicaid Data Learning Network (MDLN) at AcademyHealth, a consortium of more than 70 academic and nonprofit institutions, which Dr. Schpero co-founded.
The Checklist is designed to promote transparency and reproducibility in research using TAF data and safeguard the quality of research used to inform Medicaid policy. It includes four categories of items recommended for reporting in studies using TAF data: details on the specific data used, how the analytic sample was defined, what states and/or territories were excluded from the analysis based on data quality concerns, and additional information on measuring spending and conducting longitudinal analyses.
“Our goal was to create a resource for researchers using TAF, as well as the editors, peer reviewers, and readers who assess and consume their work,” Dr. Schpero said.
Checklist items are recommendations that can be adapted to different study designs and are meant to complement existing reporting guidelines, like the STROBE checklist for observational research.
“The Checklist represents a multi-year effort by the MDLN to develop best practices for working with the TAF data,” Dr. Schpero said. “We benefited from a ton of input across a broad cross-section of the Medicaid research community.”
- Highlights