Related Experiment Video
Updated: Jan 5, 2026
Point-of-Care Ultrasound for Peripheral Veno-Arterial Extracorporeal Membrane Oxygenation Without Left Ventricular Venting
Published on: January 17, 2025
Developing an Algorithm for Combining Race and Ethnicity Data Sources in the Veterans Health Administration
Susan E Hernandez1,2, Philip W Sylling3, Maria K Mor4,5,6
1Department of Health Services, School of Public Health, University of Washington, 1959 NE Pacific St, Magnuson Health Sciences Center, Room H-680, Box 357660, Seattle, WA 98195-7660.
Accurately identifying minority veterans in the Veterans Health Administration (VHA) is crucial for health equity. A new algorithm prioritizing the Survey of Healthcare Experiences of Patients (SHEP) and VA Corporate Data Warehouse (CDW) improves minority identification.
Area of Science:
- Health Services Research
- Health Equity
- Data Science
Background:
- Racial/ethnic disparities persist within the Veterans Health Administration (VHA).
- Accurate identification of minority veterans is essential for monitoring and addressing health equity.
- Existing VHA race and ethnicity data have variable accuracy across different databases.
Purpose of the Study:
- To develop an algorithm for constructing accurate race and ethnicity variables from VHA data sources.
- To compare demographic data accuracy across multiple VHA databases.
- To improve the identification of historically under-identified minority veteran populations.
Main Methods:
- Utilized VHA survey data (Survey of Healthcare Experiences of Patients - SHEP) and administrative databases (VA Corporate Data Warehouse - CDW, VA Defense Identity Repository - VADIR, Medicare) from 2003-2015.
- Employed measures of agreement including sensitivity, specificity, positive/negative predictive values, and Cohen kappa.
- Compared self-reported race/ethnicity from SHEP against other data sources to establish an algorithm.
Main Results:
- Agreement between SHEP and other data sources was high for White and Black veterans.
- Agreement was substantially lower for other minority groups.
- The VA Corporate Data Warehouse (CDW) showed better agreement than VADIR or Medicare.
Conclusions:
- An algorithm prioritizing data source precedence was developed for the VHA.
- The algorithm improves the accuracy of identifying historically under-identified minority veterans.
- Recommended data source order: SHEP, CDW, VADIR, Medicare.
