Characterizing Military and Veteran Health Data Platforms for Real-World Evidence Generation: A Tutorial for Investigators (Preprint)

BACKGROUND The United States federal government operates two of the largest integrated healthcare delivery systems in the world: the Military Health System (MHS) and the Veterans Health Administration (VHA). Both systems generate rich longitudinal real-world data (RWD) used extensively for clinical, epidemiologic, and regulatory-grade research. Despite their proven value, the two systems have not been comprehensively compared as research platforms in a single reference, limiting investigators' ability to identify the most fit-for-purpose data source for a given research question. OBJECTIVE To characterize the MHS Data Repository (MDR) and the VHA Corporate Data Warehouse (CDW)—accessed through the VA Informatics and Computing Infrastructure (VINCI)—as complementary and interoperable research data platforms; to compare their population characteristics, data architectures, longitudinal depth, and key strengths and limitations; and to introduce investigators to data access mechanisms, including the Department of War (DoW)-VA Infrastructure for Clinical Intelligence (DaVINCI). METHODS We conducted a descriptive review of published literature, federal agency documentation, and technical specifications describing the structure, governance, population demographics, and research use cases of the MDR and VHA CDW. We additionally describe their interoperability through DaVINCI. RESULTS The MDR captures data on more than 9.5 million beneficiaries at any time, encompassing direct care and TRICARE-financed civilian care, with exceptional pharmacy capture, familial linkage, and integration with the Serum Repository (DODSR). The VHA CDW serves approximately 9 million enrolled Veterans annually, holds records for more than 20 million unique individuals across nearly three decades of EHR use, and is integrated with the Million Veteran Program (MVP), the largest genetic data repository tied to a health care system in the United States. DaVINCI, which exchanges data between both repositories as source tables and in the OMOP Common Data Model, enables longitudinal follow-up across the military-to-Veteran transition, a capability not replicated elsewhere. The two platforms share a substantial overlapping population, since service members who separate from active duty and enroll in VHA care appear in both systems, making deduplication an important analytic consideration. Both platforms support retrospective and prospective research through established cooperative access mechanisms. CONCLUSIONS The MDR and VHA CDW are powerful, complementary, and increasingly interoperable federal health data platforms that together constitute one of the most capable real-world research ecosystems available globally. Understanding their distinct characteristics, overlapping populations, and access pathways is essential for investigators seeking to conduct high-impact, generalizable, regulatory-grade research using U.S. military and Veteran health data, with DaVINCI specifically designed to link these records across the transition. This overlap is a research asset enabling studies of how military-era exposures shape Veteran health outcomes.