When Race/Ethnicity Data Are Lacking: Using Advanced Indirect Estimation Methods to Measure Disparities

Abstract

A key aim of U.S. health care reforms is to ensure equitable care while improving quality for all Americans. Limited race/ethnicity data in health care records hamper efforts to meet this goal. Despite improvements in access and quality, gaps persist, particularly among persons belonging to racial/ethnic minority and low-income groups. This study describes the use of indirect estimation methods to produce probabilistic estimates of racial/ethnic populations to monitor health care utilization and improvement. One method described, called Bayesian Indirect Surname Geocoding, uses a person's Census surname and the racial/ethnic composition of their neighborhood to produce a set of probabilities that a given person belongs to one of a set of mutually exclusive racial/ethnic groups. Advances in methods for estimating race/ethnicity are enabling health plans and other health care organizations to overcome a long-standing barrier to routine monitoring and actions to reduce disparities in care. Though these new estimation methods are promising, practical knowledge and guidance on how to most effectively apply newly available race/ethnicity data to address disparities can be greatly extended.