Using Bayesian Imputation to Assess Racial and Ethnic Disparities in Pediatric Performance Measures

Objective: To analyze health disparities in pediatric HEDIS quality of care measures and to determine how imputation has an impact on these disparities. Data Sources: Five HEDIS measures are calculated based on 2012 administrative data for 145,652 children in two public insurance programs in Florida. Methods: The Bayesian Improved Surname and Geocoding (BISG) imputation method is used to impute missing race and ethnicity data for 42% of the sample (61,954 children). Models are estimated using the imputed data as well as dropping observations with missing race and ethnicity data. Principal Findings: Using BISG, we find that age, female, enrollment in Children’s Medical Services Network, and more severe illness are associated with a higher likelihood of HEDIS compliance. The effect of race and ethnicity on HEDIS compliance varies across the measures. Using the BISG imputed race and ethnicity analysis as the benchmark, dropping those individuals who do not self-report their race and ethnicity substantially alters the coefficient estimates. In particular, the race and ethnicity coefficient estimates are systematically dampened if they are positive and magnified if they are negative. Conclusions: These results provide further support for the importance of appropriately accounting for missing race and ethnicity data through advanced imputation methods.