Improving Occupational Health Disparity Research: Testing a method to estimate race and ethnicity in a working population. Issue 8 (2nd April 2018)

• Record Type: Journal Article
• Authors: Smith, Caroline K.; Bonauto, David K.
• Is Part Of: American journal of industrial medicine. Volume 61:Issue 8(2018)
• Page Start: 640
• Page End: 648
• Publication Date: 2018-04-02
• Subjects: gender stratified -- methods -- ROC curves -- surveillance -- workers' compensation; Medicine, Industrial -- Periodicals; Médecine du travail -- Périodiques; 616.9803
• DOI: 10.1002/ajim.22850
• Languages: English
• ISSNs: 0271-3586

Abstract

Abstract : Background: Race and ethnicity data are often absent from administrative and health insurance databases. Indirect estimation methods to assign probability scores for race and ethnicity to insurance records may help identify occupational health inequities. Methods: We compared race and ethnicity estimates from the Bayesian Improved Surname Geocoding (BISG) formula to self‐reported race and ethnicity from 1132 workers. Results: The accuracy of the BISG using gender stratified regression models adjusted for worker age and industry were excellent for White and Latino males and Latino females, good for Black and Asian Pacific Islander males and White and Asian Pacific Islander females. American Indian/Alaskan Native and those who indicated they were "Other" or "More than one race" were poorly identified. Conclusion: The BISG estimation method was accurate for White, Black, Latino, and Asian Pacific Islanders in a sample of workers. Using the BISG in administrative datasets will expand research into occupational health disparities.