Physician Racial Bias and Medical Care

Background and Significance: Well-documented disparities exist in quality of health care for racial/ethnic populations in the US. The following question guided this research: Why would health care providers, who generally believe in quality of care for all, offer poorer quality or inappropriate care to racial/ethnic minority patients? A recent report from the Institute of Medicine, Unequal Treatment, (2003) speculates that health care provider (physician) implicit racial bias may contribute to racial/ethnic health care disparities. Empirical evidence to support this theory is missing.

Objective: This study aimed to measure physician implicit and explicit racial attitudes and stereotypes and to assess the relationship between physician characteristics, implicit and explicit measures and treatment recommendations.

Methods: Pediatricians were recruited for a single session web survey in Autumn 2005 from one academic department at a research university. Response rate was 58% (N=95). Three Implicit Association Tests (IAT) were used to measure physician attitudes about race, and stereotypes about patient compliance and quality of care. Explicit attitudes were measured by self-report. Case vignettes were used to assess quality of medical care.

Results: A positive IAT mean score that differs from zero is considered to show some degree of implicit White preference. A significant implicit preference for Whites vs. African Americans was found on the Race IAT (N= 43, M= 0.18, SD= 0.44, P= .01). This finding was weaker when compared to findings from others who take the Race IAT. A significant association found between Whites rather than African Americans and the concept of “compliant patient” (N= 88, M= 0.25, SD= 0.42, P= .001) and a moderate association between African Americans rather than Whites and the concept of “preferred medical care” (N= 45, M= - 0.21, SD= 0.33, P= .001). When asked about their own practice, pediatricians' explicit attitudes favored African Americans. No significant relationship was found between implicit and explicit measures. A significant positive relationship was found between physician age and implicit attitudes about race and compliance (r= .22, P= .05). There was no significant relationship between implicit bias and treatment recommendations on the case vignettes.

Conclusion: Research is needed to determine whether pediatricians generally hold less pro-white bias than others in society. We empirically identified an implicit “perceived compliance” and race bias. Implications of this research for clinical practice and expanding notions of cultural competence will be discussed. Further research is needed to explore the relationship between physician implicit attitudes and quality of care.