Duke 2008 | ASHE | Equity and Efficiency in Health and Healthcare

Discrimination is thought to be a significant contributor to racial/ethnic disparities in health care use and health outcomes. Racial/ethnic concordance between patients and physicians is hypothesized to reduce discrimination by fostering favorable prejudice, modification of negative stereotypes, increased clinical certainty, trust and compliance. Designing effective policies to address disparities requires understanding their source, particularly the extent to which within-provider variation is important. In this analysis, I investigate importance of discrimination in the outpatient, primary care setting by examining the role of concordance in rates of preventive screening and the length of time spent with the physician.

Discrimination and racial/ethnic concordance are, by definition, within-provider variation. While concordance itself is not random, this analysis is restricted to physicians who see both white and minority patients and specifically tests whether a patient who is the same race/ethnicity as their provider receives better quality care relative to a patient who sees the same provider and is not concordant. Geographic region is controlled for, as are factors related to patient need and access to the health care system. While previous research on the impact of concordance has focused on patients' subjective views of their satisfaction with, and quality of, care, this analysis estimates the role of concordance in more objective quality of care measures. I also address a methodological weakness in the existing literature by separately measuring the effects of patient race, physician race, and racial concordance.

I use a pooled sample of the National Ambulatory Medical Care Survey (2001-2003), a nationally-representative sample of outpatient visits. Confidential data files that include physician demographics and geographic identifiers were accessed through the National Center for Health Statistics' Research Data Center. The analysis of screening rates (blood pressure, tobacco use, and cholesterol) is restricted to patient subgroups to match clinical guidelines.

After restricting my sample to physicians who see both white and minority patients and controlling for patient characteristics, physician demographics, and geographic location, I find little evidence that concordance plays an important role in these primary care outcomes. Two exceptions are that racial/ethnic concordance increases cholesterol screening rates by 2-3 times among black and Hispanic men and appears to reduce rates of tobacco cessation counseling among black and Hispanics. Generally speaking, physician race is a much more important predictor of preventive screening and duration of visit than patient race or concordance, but the direction of the effect varies by outcome. Given the relatively large magnitudes of the effects of physician race relative to the small contributions of concordance, for some outcomes minority patients may actually be worse off in concordant pairs. This highlights the methodological importance of measuring the role of concordance separately from patient and physician race. It also suggests that policies aimed at increasing the number of minority physicians need to be combined with physician education and training to improve the quality of primary care.