Evaluating Cardiovascular Health Disparities Using Estimated Race/Ethnicity: A Validation Study

Area of Science:

• Health Disparities Research
• Biostatistics
• Clinical Outcomes Research

Background:

• Administrative data methods for race/ethnicity estimation are increasingly used to identify health disparities.
• Validation of these indirect methods using clinically relevant outcomes is lacking.

Purpose of the Study:

• To evaluate the validity of indirect race/ethnicity identification (using residence and surname) against clinically relevant outcomes.
• To compare the effectiveness of drug coverage interventions based on self-reported versus indirectly assigned race/ethnicity.

Main Methods:

• Utilized data from 2387 participants in the Post-MI Free Rx Event and Economic Evaluation (MI FREEE) trial.
• Employed tests of interaction to assess differences in treatment effects on adherence and major vascular events/revascularization based on race/ethnicity assignment methods.

Main Results:

• The impact of full drug coverage on clinical events differed significantly based on self-reported race (interaction P=0.05) but not on indirectly estimated race (interaction P=0.83).
• Adherence outcomes showed similar results for both self-reported and Bayesian Improved Surname Geocoding (BISG)-estimated race/ethnicity for most medications.

Conclusions:

• Indirect race/ethnicity estimation yielded quantitatively and qualitatively different results compared to self-reported data.
• These indirect methods may not accurately capture race/ethnicity-related health behaviors, potentially impacting disparity research.