Interoperability & HIE

Abbie Zhang, Daniel Lipsey +6

Does a higher volume of health information exchange (HIE) improve outcomes, and do effects differ between community and VA direct care? Using VHA electronic health record data from January 2022 to December 2023 (3144 medical center-months; ~2.4 million patients monthly), the authors instrumented HIE volume with each center's count of organizational exchange partners, with center and month fixed effects. In community care, a 1-SD increase in HIE volume was associated with 4.07 fewer 30-day readmissions, 1.45 fewer avoidable hospitalizations, and 0.25 fewer inpatient deaths per center-month; in VHA direct care, 7.07 additional readmissions and 0.19 additional inpatient deaths, with no significant change in avoidable hospitalizations.

Hannah A Burkhardt, Sarah J Blach +8

Can a health system-governed data platform overcome the fragmentation, latency, and quality limits of real-world data? This descriptive platform paper reports on Truveta's partnership model, architecture, and applications, covering de-identified electronic health record data on 130 million US patients — roughly 1 in 3 Americans — aggregated from participating health systems and linked to closed claims, mortality, and social determinants data. Records are ingested daily, normalized to standard ontologies, and processed with NLP to extract concepts from notes, imaging, and pathology text. The data have supported over 100 publications spanning treatment effectiveness, device surveillance, COVID-19 vaccine safety, and health equity. No comparative effect estimates are reported.

Josh Lemieux, Emily R Pfaff +8

Can national health information networks be repurposed to acquire EHR data for research? In a pilot, the All of Us Center for Linkage and Acquisition of Data worked with eHealth Exchange, the largest US health information network, to route participant-authorized queries to one health information exchange and one hospital system; returned FHIR and C-CDA records were mapped to the OMOP common data model and compared with existing All of Us EHR data. Retrieved records added complementary information and improved completeness, though the abstract reports no effect sizes. Barriers included inconsistent capacity to transact authorizations and variable data quality.