RACIAL AND ETHNIC DISPARITIES IN MEDICAL HISTORY TAKING: DETECTING SUBSTANCE USE AMONG LOW-INCOME PREGNANT WOMEN Objective: To determine whether providers’

medical history-taking of substance use varies with their patients ’ race or ethnicity. Design/Setting/Participants: The medical re-cords of 1083 low-income women who de-livered at an urban teaching hospital over a 12-month period were reviewed and data were abstracted. The frequency with which the presence or absence of substance use was documented was compared among Black, Hispanic, and White women. Multivariate models predicting documentation of data on substance use were also built. Results: Information on substance use was more often documented in the medical records of Black and Hispanic women than in those of White women. For example, 74 % of Black women, 78 % of Hispanic women, and 60 % of White women had data on cocaine use documented in their medical records (P5.0001). Multivariate analyses found that, after controlling for other factors, Black (odds ratio [OR] 4.1) and Hispanic (OR 5.3) non-clinic patients were more likely thanWhite non-clinic patients to have documented data on substance use history in their medical records. No racial/ ethnic disparities were found among clinic patients. Among White women, clinic patients were more likely than non-clinic patients to have documented data on substance use. Among other races and ethnicities, no dispar-ities were found between places of care. Conclusions: The differential collection of information on women’s medical histories may be influenced by organizational factors and/or provider factors. The standard imple-mentation of history-taking protocols would reduce the influence of institutional and personal biases on the care provided to pregnant women and enable all patients to receive appropriate referrals to treatment