Is Doctor-patient Race Concordance Associated with Greater Satisfaction with Care?*

Is Doctor-patient Race Concordance Associated with Greater Satisfaction with Care?* THOMAS A. LAVEIST Johns Hopkins University AMANI NURU-JETER Johns Hopkins University Journal of Health and Social Behavior 2002, Vol 43 (September): 296306 We examined a national sample of African American, white, Hispanic, and Asian American respondents to test the hypothesis that doctor-patient race concordance is predictive ofpatient satisfaction. Our analysis examined racial/ethnic differences in patient satisfaction among patients in multiple combinations of doctor-patient race/ethnicity pairs. Additionally, we outline the determinants of doctor-patient race concordance. The analysis used the 1994 Commonwealth Fund Minority Health Survey to construct a series of multivariate models. We found that for respondents in each race/ethnic group, patients who had a choice in the selection of their physician were more likely to be race concordant. Whites were more likely to be race concordant with theirphysician compared to African American, Hispanic, and Asian American respondents. Among each race/ethnic group, respondents who were race concordant reported greater satisfaction with their physician compared with respondents who were not race concordant. These findings suggest support for the continuation of efforts to increase the number of minority physicians, while placing greater emphasis on improving the ability ofphysicians to interact with patients who are not of their own race. Efforts to increase the ranks of Ahcan of minority physicians (Carlisle et al. 1998; American and Hispanic healthcare providers Libby et al. 1997).Yet, although it is quite popare among the most commonly proposed solu- ular, the race concordance hypothesis has not tions to racial disparities in health. The under- been thoroughly tested. Is it true that patients lying hypothesis is that increasing the number express greater satisfaction if they have a of minority providers will increase the number physician of the same race group? In this paper of minority patients who have physicians of we examine this question. their own racialiethnic group, and that when the raceiethnicity of the patient and provider is concordant, the patient will receive better qual- BACKGROUND ity care and express greater satisfaction with Through most of the twentieth century, the that care. Medical schools have responded to treatment of racial and ethnic minorities withthis proposition by increasing the production in healthcare settings was dictated by a set of customs spawned by the American racial caste * This article was supported by a grant from the system and enforced by governmental policies. Commonwealth Fund to Dr. LaVeist. Address Healthcare facilities were racially segregated. reprint requests to Thomas A. LaVeist, Johns Hopkins University, Bloomberg School of Public Nonwhite patients received substandard care. Health, Department of Health Policy and Manage- Interpersonal communications during medical ment, 624 North Broadway, Room 441, Baltimore, encounters between patients and healthcare Maryland 21205, email: tlaveist@jhsph.edu. providers of different racial and ethnic groups

RACE CONCORDANCE AND PATIENT SATISFACTION were discourteous to minorities. These factors led to racial and ethnic disparities in access and utilization of quality health services, distrust of the medical care system among racial and ethnic minorities (LaVeist, Nickerson, and Bowie 2000b), and contributed to racial disparities in health status (LaVeist, Bowie, and Cooley-Quille 2000a; Smith 1999). To remedy this situation, calls for increases in the numbers of minority healthcare providers were met with the establishment of affirmative action programs which increased the numbers of minority physicians in the workforce (Carlisle et al. 1998; Libby et al. 1997). Evaluations of the impact of such policies have largely been limited to examinations of trends and patterns in minority physician production (Libby et al. 1997) and studies of the practice settings where minority physicians work (Cantor et al. 1996). For example, in a study of California-based physicians, Komaromy et al. (1996) demonstrated that black and Hispanic physicians were more likely than their white counterparts to care for black, Hispanic, and low income patients. Likewise, Moy and Bartman's (1995) national sample found that minority physicians were more likely to care for minority, medically indigent and sicker patients. Other studies have reported similar findings (Xu et al. 1997; Keith et al. 1985; Lloyd and Johnson 1982; and Rocheleau 1978). A related area of inquiry that has received some attention is the degree to which race plays a role in patients' selection of their healthcare providers. Gray and Stoddard's (1997) analysis of the National Medical Expenditure Survey showed that minority patients were significantly more likely to have a minority physician. However, while this study demonstrated a pattern, it was not able to determine if patients were selecting physicians as a matter of preference or if they had selected minority physicians because they were the only providers available in their communities. Saha et al. (2000) was able to overcome that shortcoming, demonstrating that black, white, and Hispanic patients sought care from physicians of their own race because of personal preferences and language, and not solely because of limited options. Bertakis (1981) found that minorities do not express a preference for the race of their physician. However, it is well documented that minority physicians are more likely to practice in minority communities (Moy and Bartman 1995), and, when available, people tend to select physicians of their same race (Saha et al. 2000). Whether or not patient outcomes are affected by doctor-patient race concordance is less clear. We could identify only a handful of studies that have attempted to address this question (Chen et al. 2001; Cooper-Patrick et al. 1999; Saha et al. 1999). Cooper-Patrick et al. (1999) conducted a telephone survey of 1,816 African American and white adults who were recent patients of a primary care practice to examine patient assessments of their physician's participatory decision-making style. Their analysis found that patients who were race concordant with their physician rated their visits as significantly more participatory than patients who were race discordant. Saha et al. (1999) found that African American race concordant patients were more likely to rate their physician as excellent and were more likely to report receiving preventive care and needed medical care. Chen et al. (2001) failed to find a race difference in utilization of cardiac catheterization by race concordance. While the empirical support for a direct effect of doctor-patient race concordance on patient outcomes is limited, the impact of patient's subjective ratings of satisfaction on health-related outcomes is better understood. Standardized surveys of patient satisfaction have gained wide acceptance as a key component of healthcare quality assessment and healthcare system performance (Scanlon et al. 2001; Harris-Kojetin et al. 2001; Cleary and McNeil 1988; Mukamel and Mushlin 2001; Simon and Monroe 2001). Satisfaction ratings are used by individual healthcare consumers and by employers as an aide in choosing health plans and providers (Crofton, Lubalin, and Darby 1999) and as such can indirectly impact the financial viability of a medical practice (Binderman 200 1). In addition to its intrinsic importance for individual healthcare consumers and thirdparty payers, patient satisfaction is also an important determinant of health-related outcomes, such as health services utilization (Zastowny, Roghmann, and Cafferata 1989; Roghmann, Hengst, and Zastowny 1979), decision to switch to another health plan (Murray et al. 2000; Allen and Rogers 1998; Hennelly and Boxerman 1983; Sorensen and Wersinger 1981), compliance with medical regimen

298 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR (Smith et al. 1987), and decision to initiate malpractice suits (Penchansky and Macnee 1994). Litt and Cuskey (1984) demonstrated that adolescents who reported greater satisfaction after an initial doctor visit were significantly more likely to keep their follow-up appointment compared with less satisfied patients. Subsequent studies of the effect of patient satisfaction on appointment keeping have confirmed this relationship (Freed et al. 1998; Carlson and Gabriel 2001; Ivanov and Flynn 1999). Newcomer, Preston, and Harrington (1 996) found that satisfaction with perceived physician quality and interpersonal satisfaction with the provider reduced the risk of health plan disenrollment (switching to another health plan during the open enrollment period). Schlesinger, Druss and Thomas (1999) conducted a large (n = 20,283) national study of employees from three corporations and found that 5843 percent of healthy respondents in fee-for-service plans who expressed dissatisfaction with their care elected to switch health plans during the open enrollment period. Other studies have also demonstrated a relationship between satisfaction and disenrollment from health plans (Murray et al. 2000; Allen and Rogers 1998; Hennelly and Boxerman 1983; Sorensen and Wersinger 198 1). Additionally, patients who report greater satisfaction with their care are more likely to comply with the prescribed medical regimen (Hall and Dornan 1990; Smith, Ley, and Seale 1987) and less likely to initiate a malpractice suite in the event of an adverse medical outcome (Moore, Adler, and Robertson 2000; Penchansky and Macnee 1994). Although the literature addressing these issues is small and not well developed, the available studies present compelling findings that further suggest an important role for patient satisfaction in research on health behavior and healthcare policy. The present study expands on previous findings in several respects. We explore predictors of doctor-patient race concordance, a phenomenon that has not been adequately studied in medical sociology. Additionally, our analyses go beyond examining race concordant versus discordant doctor-patient pairs to examine patient satisfaction within each available doctorpatient racial and ethnic pairing. Finally, we further explicate the consequences of efforts to increase minority health care providers and suggest consequent policy implications. METHODS Data for this study come from the 1994 Commonwealth Fund Minority Health Survey (Hogue and Hargraves 2000). The Minority Health Survey is a nationally representative sample of adults 18 years of age and over residing in households with telephones within the 48 contiguous United States. Interviews were conducted via telephone using randomdigit dialing. African Americans, Hispanics, and Asian Americans were over-sampled in the Minority Health Survey. The present analysis examines the sub-sample of 2,720 respondents who reported having a usual source of care (910 whites, 745 blacks, 676 Hispanics, and 389 Asian Americans). Because of over-sampling of minority respondents, all analyses of the full sample will incorporate sampling weights. The weights will maintain the representative nature of the sample. However, weights will not be used for race-specific analyses. Outcome Variables The first set of analyses examine factors associated with race concordance, and the second set of analyses examine patient satisfaction. Race concordance was specified as a binary variable indicating that the respondent's race is concordant with the race of their physician. Patient satisfaction is a five-item index. The individual items that make up this index ask how good a job your doctor does at: (1) proving good health care, (2) treating you with dignity, (3) making sure you understand what you've been told, (4) listening to your health problems, and (5) being accessible by phone or in-person. The item response codes were a 4- point scale ranging from poor to excellent. These items were combined to form an index by averaging across the items, resulting in a patient satisfaction summary score ranging from 1 to4. Independent Lfzriables In analysis examining race concordance, patient's racelethnicity and physician choice were the primary independent variables. Race was specified as a set of binary variables indicating white non-hispanic, black non-

RACE CONCORDANCE AND PATIENT SATISFACTION Hispanic, Hispanic, and Asian American. Physician choice was measured by the question, "how much choice do you have in where you go for medical care?" Respondents indicating that they have "a great deal" or "some" were coded as having choice, and responses of "very little" or "no choice" were coded as no choice. In analysis of patient satisfaction, physician-patient race concordance was the primary independent variable. The covariates were: not speaking English as the primary language, sex, age, income, education, and health insurance. "English as the primary language" is a binary variable which indicates that English is not the respondent's primary language. Sex was specified as a binary variable indicating male. Age is a continuous variable. Annual income was specified as a continuous variable. Education is a continuous variable coded as less than high school graduate, high school graduate, some college, college graduate, and more than college graduate. Health insurance is a set of binary variables indicating private insurance, Medicare, Medicaid, and uninsured. RESULTS 299 In Table 1, we display analyses examining the distribution of the covariates among the sub-samples. Each raciavethnic group is generally evenly divided by gender. The white and Afncan American respondents tended to be older than the Hispanic and Asian samples. While nearly 43 percent of white respondents and over 33 percent of African American respondents were over age 5 1, only 19.6 percent of Hispanics and 18.6 percent of Asian American respondents were. White and Asian American respondents tended to have higher incomes than the Hispanic or African American respondents. Over 18 percent of whites and nearly 13 percent of Asian American respondents earned over $75,000 per year, compared with 8.6 percent of African Americans and 7.6 percent of Hispanics. In addition, Asian Americans and whites tended to have higher levels of education compared with African Americans and Hispanics. Thirtyone percent of whites, 23.8 percent of African Americans, 20.7 percent of Hispanics, and 44.3 percent of Asians were college graduates. Table 1 also shows that while most respon- TABLE 1. Sample Descriptive Statistics by Respondent's Racdethnicity Variable White Black Hispanic Asian American Male 52.9 50.0 50.6 50.0 Age*** 18-30 16.7 20.4 30.0 30.2 3140 22.2 27.0 30.3 29.5 41-50 18.3 18.9 20.1 21.7 5145 22.5 20.9 14.7 14.4 6694 20.4 12.7 4.9 4.2 Income*** < $7,501 7.7 13.2 10.4 9.1 $7,501-$15,000 9.8 12.6 11.4 10.3 $15,001-$25,000 15.3 19.3 19.0 9.6 $25,001-$35,000 18.8 18.6 19.8 22.3 $35,001-$50,000 20.2 17.4 19.3 22.6 $50,001-$75,000 14.5 10.3 12.7 13.5 $75,001-$100,000 7.3 5.8 5.4 6.9 > $100,000 6.5 2.8 2.2 5.8 Education*** < High School Graduate 10.3 17.7 19.4 19.0 High School Graduate 32.9 31.6 29.7 21.5 Some College 25.9 27.0 30.2 15.1 College Graduate 20.0 13.8 15.3 31.4 Post College 11.0 10.0 5.4 12.9 Insurance*** Private 86.6 75.6 71.1 69.5 Medicare 24.1 19.9 8.7 8.3 Medicaid 7.7 15.4 10.7 9.7 Uninsured 6.9 10.8 20.9 22.2 Doctor Choice*** 76.3 65.3 59.5 49.7 English not primary language*** *p<.05, **p<.ol,***p<.ooi 1.1 1.7 30.2 80.7 Note: Insurance status categories are not mutually exclusive. For example, a respondent can have Medicare and a private "Medigap" policy. Also, one could be dually eligible for Medicare and Medicaid.

JOURNAL OF HEALTH AND SOCIAL BEHAVIOR dents had a private source of insurance, more than one fifth of Hispanic and Asian American respondents were uninsured. Consistent with their relatively older age, white (24.1 percent) and African American (19.9 percent) respondents were more likely to have Medicare than Asian Americans (8.3 percent) or Hispanics (8.7 percent). Also consistent with their lower incomes, African Americans (15.4 percent) were more likely to have Medicaid compared with whites (7.7 percent), Hispanics (10.7 percent) and Asian Americans (9.7 percent). Majorities of white (76.3 percent), Afncan American (65.3 percent), and Hispanic (59.5 percent) respondents reported that they had a choice in selecting their physician. A minority of Asian American respondents (49.7 percent) reported not having a choice. Finally, the majority of Asian American respondents (80.7 percent) indicated that English was not their primary language, as did nearly one third of Hispanics (30.2 percent). Table 2 displays the relationship between race of respondent and race of physician. The table shows that with the exception of Asian Americans, respondents were most likely to have a white physician. Other than white physicians, the second most common configuration was for Afncan American, Hispanic, and Asian American respondents to have physicians of their own racelethnic group. Twenty-one percent of African American respondents, 18.9 percent of Hispanic respondents, and 52.2 percent of Asian American respondents had physicians who were of the same race or ethnicity. In all, 46.8 percent of respondents were race concordant. By far, the race group most likely to be race concordant with their physician was whites. However, Asian Americans were more likely to be race concordant than have a white physician. White respondents were 3.94 times more likely than Afncan American respondents and 4.53 times more likely than Hispanic respondents to have a physician of the same race. However, whites were only 1.64 times more likely to be race concordant with their physician than Asian Americans. Afncan American and Hispanic respondents were also more likely than whites to have physicians from minority groups other than their own. Nine percent of Afncan American and 9.5 percent of Hispanic respondents reported having an Asian American physician, while 7.1 percent of whites reported having an Asian American physician. 8.5 percent of African Americans and 9.3 percent of Hispanics reported having a physician of another race, compared with only 3.7 percent of whites. Of all groups, Asian American respondents were least likely to have a physician who was a non-asian minority. In Table 3, we test whether patient's race is associated with doctor-patient race concordance within multivariate models. We used logistic regression since the response variable is binary. Model 1 is a weighted analysis based on the full sample including all race groups. The model shows that respondents belonging to the nonwhite raciallethnic groups were less likely to be race concordant with their physicians compared with whites, adjusting for sex, age, income, education, health insurance, and having the ability to select one's own doctor. Additionally, persons who had a choice in the selection of their physician and those for whom English is not the primary language were significantly more likely to be race concordant. In Model 2 we examine correlates of race concordance among whites. The model contains the same set of variables as those in Model 1, with the exception of patient race. Age is associated with race concordance such that younger whites wsre more likely to be race TABLE 2. Physician-patient Race Concordance in the 1994 Commonwealth Minority Health Survey Patient's Race Asian White Black Hispanic American Physician's Race (n = 910) (n = 745) (n = 676) (n = 389) White 779 (85.6%) 436 (58.5%) 406 (60.1%) 175 (45.0%) Black 14 (1.5%) 162 (21.7%) 15 (2.2%) 5 (1.3%) Hispanic 19 (2.1%) 17 (2.3%) 128 (18.9%) 2 (.5%) AsianlPacific Islander 68 (7.5%) 75 (10.1%) 71 (10.5%) 203 (52.2%) Other 30 (3.3%) 55 (7.4%) 56 (8.3%) 4 (1.0%)

RACE CONCORDANCE AND PATIENT SATISFACTION 301 TABLE 3. Logistic regression analysis of correlates of physician-patient race concordance, coefficient (standard error). Asian Total Sample White Black Hispanic American Variable Model 1 Model 2 Model 3 Model 4 Model 5 Female (reference) - (.09)** - - - - - - - - Male -.219 -.I31 (.18) -.I64 (.19) -.342 (.21) -.I57 (.21) Age Income -.205 (.04)*** -.I85 (.08)*,061 (.03)*,073 (.05) -.211 (.08)**,086 (.06)*.002 (.09) -.074 (.07),041 (.09) -.065 (.06) Education -.046 (.05),032 (.08) -.086 (.09) -.082 (.lo) -.075 (.09) Insurance Private (reference) -- - - - - - - - - Medicare,494.524 (.28).445 (.29) -.498 (.41) -.336 (.44) Medicaid -.411 (.la)* -.471 (.32) -.303 (.31),250 (.37) -.I24 (.40) Uninsured,051 (.15),062 (.35),218 (.35),138 (.30) -.018 (.29) Doctor Choice 3.485 (.12)*** 3.467 (.21)*** 2.065 (.31)*** 2.760 (.37)*** 2.575 (.25)*** English not primary 283 (.24)***,135 (.68) -.208 (.66) 1.021 (.22)***.718 (.29)** language Patient Race White (reference) - - - - - - - - - Black -2.916 (.16)*** Hispanic -3.483 (.22)*** Asian American -2.22 1 (.20)*** Constant -1.113 (.21)*** -1.460 (.38)*** -2.920 (.459)*** 1.021 (.21)*** -2.386 (.52)*** Model Statistics x2 2089.17 443.87 86.63 119.06 159.15 df 12 9 9 9 9 D.ooo.000,000,000,000 concordant. Also, respondents who reported patient satisfaction rating for Asian/Pacific having a choice in the selection of their physi- Islander doctors compared with white doctors. cian were more likely to be race concordant. Asian American patients also reported less sat- We then test the same set of independent isfaction compared with white patients. variables among African American respon- Respondents with higher incomes reported dents in Model 3. In this model we find that greater satisfaction. Additionally, physician among African Americans who reported hav- choice was associated with greater satisfacing at least some choice in selecting their tion, but English language was not. physician, there was a greater tendency to In Model 2, we examine the correlates of select an African American doctor. Also, patient satisfaction among the white sub-samyounger African Americans were more likely ple. The model shows that among white to have an African American physician. respondents patient satisfaction ratings were Among Hispanics and Asian Americans, race significantly lower only for Hispanic and concordance is also predicted by physician AsianIPacific Islander physicians. White choice. Hispanics and Asian Americans for respondents did not report lower ratings of satwhom English was not the primary language isfaction for African American doctors were more likely to be race concordant. (although the number of cases were small, n = Now we turn to an assessment of the effect 14). Doctor choice remains an important deterof race concordance on patient satisfaction. In minant in the white sub-sample, as does Table 4 we display a series of ordinary least income. Model 3 examines the independent squares multiple regression models for the full variables among African Americans. The sample, followed by models within each model shows that Afncan Americans who are raciallethnic sub-group. Model 1, specified for race concordant with their physicians reported the full sample, shows a significantly higher a higher level of satisfaction compared with patient satisfaction rating for African African Americans who had white or Asian American doctors and a significantly lower doctors. There is also a significant effect of

Total Sample White Variable Model 1 Model 2 Male Age Income Education Insurance Private (reference) Medicare Medicaid Uninsured Doctor Choice English not primary language Patient Race White (reference) Black Hispanic Asian American Physician Race White Black Hispanic AsianPacific Islander Other Constant Model Statistics Adj R2 F 4 *p<.05, **p<.01, ***p<,001 Note: Data in parentheses are standard errors. JOURNAL OF HEALTH AND SOCIAL BEHAVIOR TABLE 4. Regression of Correlates of Patient Satisfaction, Racelethnic Specific Analysis, (Standardized Coefficient) Asian Black Hispanic American Model 3 Model 4 Model 5 physician choice on patient satisfaction. Additionally, older age was associated with greater satisfaction with care. Among Hispanics (Model 4), race concordance is associated with patient satisfaction. Hispanic respondents were significantly more satisfied if they had a Hispanic physician compared with doctors of all other race groups. Not spealung English as a primary language was not a significant correlate of patient satisfaction. Analysis testing for an interaction between race concordance and primary language among Hispanic respondents (analysis not shown) did not yield a significant effect. However, among Asian American respondents (Model 5), there was a significant effect of not being a primary English speaker on patient satisfaction. Also, Asian American respondents reported greater satisfaction with care if they were race concordant with their physician compared to having a white or Hispanic doctor. DISCUSSION We examined a national sample of white, African American, Hispanic, and Asian American respondents to test the hypothesis that doctor-patient race concordance is associated with patient satisfaction. We first examined correlates of doctor-patient race concordance, and found that respondents who had the ability to choose their own doctor were significantly more likely to be race concordant with their physician compared with respondents that lacked choice in their physician. This was consistent for all raciallethnic groups. Additionally, with the exception of Asian Americans, everyone was most likely to have a white physician, and white respondents were significantly more likely to be race concordant than Afncan American, Hispanic, or Asian American respondents. Moreover, for Hispanic and Asian American respondents, if English was not their primary language, they were

RACE CONCORDANCE AND PATIENT SATISFACTION more likely to be race concordant with their physician. Our analysis found support for the concordance-satisfaction hypothesis. Respondents of each racial and ethnic group reported the highest level of satisfaction if they were race concordant. Moreover, all respondents reported greater satisfaction with physicians from their own race. One caveat is that several doctorpatient combinations had fifteen or fewer respondents. (i.e., white, Hispanic, and Asian American respondents with African American physicians; Asian American patients with Hispanic doctors, and doctors of some "other" race group) Therefore, estimates for these categories may not be reliable (see Table 2). The available data do not allow for a fiiller exploration of why patients choose to be race concordant with their physician. However, it is possible to speculate.-to this end we present three hypotheses: (1) the intra-group connectedness hypothesis, (2) the internalized bigotry hypothesis, and (3) the experiential hypothesis. That respondents of each raciavethnic group were more likely to select a physician of their on race or ethnicity may suggest that patients are more likely to trust and feel greater comfort with physicians of the same race group. The doctor-patient relationship is highly dependent upon trust (Thom 2001). Patients may perceive that a physician of their same raciavethnic group will exercise a greater sense of agency with regard to their care. By agency we mean that healthcare providers will operate in patients' best interests (Mechanic and Schlesinger 1996). If this is the case, the source of that trust among patients may be due to an intrinsic sense of connection to members of their own raciallethnic group. Patients may be associating their own sense of raciallethnic connectedness with greater trust in physicians of their own group (whether it is deserved or undeserved). Further research is needed to specifically address this possibility. A second possibility is that race concordant patient preferences may be a manifestation of the internalization of broader societal racism. Thus, the preference for race concordance may be less a matter of in-group connectedness and more a matter of negative attitudes about outgroup members. Patients' preferences may be shaped by lingering distrust resulting from a historical context where doctor-patient encounters between race groups (when they occurred) were characterized by discrimination and inferior care. Additionally, physicians who were raciallethnic minorities were allowed to practice only on minority patients. Within such a context, racial and ethnic minorities may have come to prefer healthcare providers of their own raciallethnic group, and white patients may regard the care provided by nonwhite physicians to be of lower quality. Alternatively, if greater patient trust among race concordant doctor-patient pairs does exist, the source of that trust may be experiential. Overt racial segregation and discrimination in health care has abated; however, numerous studies have found that less overt forms of discrimination persist (Schulman et al. 1999; van Ryn and Burke 2000; Todd, Samaroo, and Hoffman 1993; Todd et al. 2000). Thus, patients may be reacting to their past experiences or the experiences of others, where physicians not of their same raciallethnic group have been less courteous or less engaging. Previous studies have found that physicians tended to rate minority patients more negatively than racial non-minorities. van Ryn and Burke (2000) found that physicians rated minority patients as less intelligent, less compliant, and more likely to engage in risky health behaviors. Such attitudes may be perceived by patients via nonverbal cues or verbal tone or inflections. In addition to negative experiences driving patients to preferences for race concordance, patient's may prefer race concordance because of positive experiences. Previous research suggests that patients tend to choose physicians based on their perception of the physician being knowledgeable, interested in patient's concerns, the amount of time spent with patients, and having good interpersonal communications skills (Hill and Gamer 1991). Saha et al. (2000) found that patients were more likely to attribute these qualities to physicians of their same raciallethnic group, and Cooper-Patrick et al. (1999) demonstrated that both African American and white patients rated their doctor visits as more participatory if they were race concordant with their physician. Whether patient preferences for race concordance with their physician is caused by internalized bigotry, feelings of intra-group connectedness, or previous experiences will lead to different paths for policy. If race concordance preferences result from greater trust

rtal OF HEALTH AND SOCIAL BEHAVIOR and comfort among persons of similar cultural groups, then this argues for the continuation of efforts to increase racial and ethnic diversity among the ranks of healthcare providers. If race concordance preferences are motivated by bigotry then the solution lies in the furtherance of broader social reforms regarding societal norms for racial tolerance. However, if the etiology of the race concordance preference is a reaction to past experiences within the healthcare system, then the reforms should be more narrowly focused on the educational process for healthcare providers or the process of selecting persons for training as healthcare providers. While it is clear that patients prefer physicians of their same race or ethnicity if given the opportunity, our analyses did not examine the effect of race concordance on quality of care, patient compliance, or utilization of medical services. Previous studies have found that patient satisfaction is associated with each of these important health-related outcomes (Litt and Cuskey 1984; Carlson and Gabriel 2001; Ivanov and Flynn 1999). However, whether race concordance is predictive of health services utilization and whether patient satisfaction mediates or moderates this relationship has not been assessed. Analyses of these questions, in combination with the present study, will form a stronger basis to form policy implications. It has been projected that racial and ethnic minority populations in the United States will be an increasingly larger and, eventually, by mid-century, a majority of the United States population. As patients tend to prefer physicians of their own racial or ethnic group, the demand for minority physicians will likely increase in the coming decades. However, as Libby et al. (1997) has demonstrated, at the current rates of production, the supply of minority physicians will fall short of future demand. These trends foreshadow a probable decreased likelihood of race concordance for African Americans and Hispanics in the fiture, a reduction of satisfaction among minority patients, and perhaps a decline in quality of care among minorities. The avoidance of these outcomes can be achieved by increasing minority physician production, while placing greater emphasis on improving the ability of physicians to interact with patients who are racially, ethnically, or otherwise culturally different than they are. REFERENCES Allen, H.M. Jr, and W.H. Rogers. 1998. "The Consumer Health Plan Value Survey: Round Two." Health Affairs 17:26548. Bertakis, Klea. 1981. "Does Race Have an Influence on Patients' Feelings Toward Physicians?" Journal of Family Practice 13:383-87. Binderman, J. 2001. "Variables Affecting the Financial Viability of Your Practice." Journal of Medical Practice Management 17:7-10. Cantor, J.C., Miles E.L., Baker L.C., and Barker DC. 1996. "Physician Service to the Underserved: Implications for Affirmative Action in Medical Education." Inquiry 33:167-80. Carlisle, David M., Jill E. Gardner, and Honghu Liu. 1998. "The Entry of Underrepresented Minority Students into Us Medical Schools: an Evaluation of Recent Trends." American Journal of Public Health 88: 1314-18. Carlson, Matthew J., and Roy M. Gabrel. 2001. "Patient Satisfaction, Use of Services, and Oneyear Outcomes in Publicly Funded Substance Abuse Treatment." Psychiatric Services 52:123@36. Chen, Jersey, Saif S. Rathore, Martha J. Radford, Yun Wang, and Harlan M. Krumholz. 2001. "Racial Differences in the Use of Cardiac Catheterization after Acute Myocardial Infarction." New England Journal of Medicine 344:144349. Cleary, Paul D, and McNeil BJ. 1988. "Patient Satisfaction as an Indicator of Quality Care." Inqui~25:25-36. Cooper-Patrick, Lisa, Joseph J. Gallo, Junius J. Gonzales JJ, Hong T. Vu HT, Neil R. Powe, Christine Nelson C, and Daniel E. Ford. 1999. "Race, Gender, and Partnership in the Patient- Physician Relationship." Journal of the American Medical Association 282:583-89. Crofton, Christine, James S. Lubalin, and Charles Darby. 1999. "Consumer Assessment of Health Plans Study (CAHPS). Foreword." Medical Care 37:MSl-9. Freed, Lorraine H., Jonathan M. Ellen, Charles E. Irwin Jr, and Susan G. Millstein. 1998. "Determinants of Adolescents' Satisfaction with Health Care Providers and Intentions to Keep Follow-up Appointments." Adolescent Health 22:475-79. Glass, K.P., L.E. Pieper, and M.F. Berlin. 1999. "Incentive-based Physician Compensation Models." Journal of Ambulatory Care Management 22:3&46. Gray, Bradford and Jeffery J. Stoddard. 1997. "Patient-Physician Pairing: Does Racial and Ethnic Congruity Influence Selection of a

RACE CONCORDANCE AND PATIENT SATISFACTION Regular Physician?" Journal of Community Health 22:247-59. Hall, JA, and Dornan MC. 1990. "Patient Sociodemographic Characteristics as Predictors of Satisfaction with Medical Care: a Meta-analysis." Social Sciences in Medicine 302311-18. Harris-Kojetin, Lauren, Lauren McCormack, Elizabeth Jael, Judith Sangl, and Steven Garfinkel. 2001. "Creating More Effective Health Plan Quality Reports for Consumers: Lessons from a Synthesis of Quality Testing." Health Services Research 36:447-76. Hennelly, Virginia D., and Stuart B. Boxerman. 1983. "Disenrollment from a Prepaid Group Plan. A Multivariate Analysis." Medical Care 21:115467. Hill, CJ, and S.J. Gamer. "Factors Influencing Physician Choice." Hospital Health Services Administration. 1991. Winter; 36(4):491-503. Hogue, Carol J.R. and Martha Hargraves. 2000. "The Commonwealth Fund Minority Health Survey of 1994: An Overview." Pp. 3-18 in Minority Health in America: Findings and policy implications from the Commonwealth Fund Minority Health Survey, edited by Carol J.R. Hogue, Martha A. Hargraves and Karen Scott Collins. Baltimore MD and London: Johns Hopkins University Press. Ivanov, L.L. and B.C. Flynn. 1999. "Utilization and Satisfaction with Prenatal Care Services." Western Journal of Nursing Research, 21:372-86. Keith, S.N., Bell R.M., Swanson A.G., and Williams A.P. 1985. "Effects of Affirmative Action in Medical Schools. A Study of the Class of 1975." New England Journal of Medicine 313: 1519-25 Komaromay, Miriam, Kevin Grumbach, Michael Drake, Karen Vranizan, Nicole Lurie, Dennis Keane, and Andrew B. Bindman. 1996. "The Role of Black and Hispanic Physicians in Providing Health Care for Underserved Populations." New England Journal of Medicine 334:1305-10. LaVeist, Thomas A., Janice Bowie, and Michelle Cooley-Quille. 2000a. "The Middle Years: Minority in Young Adulthood." HCFA Review 21:9-22. LaVeist, Thomas A,, Kim Nickerson and Janice Bowie. 2000b. "Attitudes about Racism, Medical Mistrust and Satisfaction with Care among African-American and White Cardiac Patients." Medical Care Research and Review 57(Supplement 1): 146-61. Libby, Donald L., Zijun Zhou, and David A. Kindig. 1997. "Will Minority Physician Supply Meet US Needs?'Health Afairs 16:205-14. Litt, I.F. and W.R. Cuskey 1984. "Satisfaction with Health Care: A Predictor of Adolescents' Appointment Keeping." Journal of Adolescent Health Care 5: 196-200. Lloyd, S.M. Jr, and Johnson D.G. 1982. "Practice Patterns of Black Physicians: Results of a Survey of Howard University College of Medicine Alumni." Journal of the National Medical Association 74: 129-41. Mechanic, David, and Mark Schlesinger. 1996. "The Impact of Managed Care on Patients' Trust in Medical Care and Their Physicians." Journal of the American Medical Association 275: 1693-97. Moore, Philip J, Nancy E. Adler, and Patricia A. Robertson. 2000. "Medical Malpractice: the Effect of Doctor-patient Relations on Medical Patient Perceptions and Malpractice Intentions." Journal of Western Medicine 173:244-50. Moy, Ernest and Barbara A. Bartman. 1995. "Physician Race and Care of Minority and Medically Indigent Patients." Journal of the American Medical Association 273: 1515-20. Mukamel, Dana B. and Alvin I. Mushlin. 2001. "The Impact of Quality Report Cards on Choice of Physicians, Hospitals and HMOs: A Midcourse Evaluation." Joint Commission Journal of Quality Improvement 27:20-27. Murray, B.P., R.B. Dwore, G. Gustafson, R.J. Parsons, and L.H. Vorderer. 2000. "Enrollee Satisfaction With HMOs and its Relationship with Disenrollment." Managed Care Inteface, 13:5541. Newcomer, R, Preston S, and Harrington C. 1996. "Health Plan Satisfaction and Risk of Disenrollment among Social/HMO and Fee-forservice Recipients." Inquiry 33:144-54. Penchansky, Roy, and Macnee Carol. 1994. "Initiation of Medical Malpractice Suits: a Conceptualization and Test." Medical Care 34:280-82. Rocheleau, Bernard. 1978. "Black Physicians and Ambulatory Care." Public Health Reports 93 :278-82. Roghmann, KJ, A. Hengst, and T.R. Zastowny. 1979. "Satisfaction with Medical Care: Its Measurement and Relation to Utilization." Medical Care 17:461-79. Safran, Dana G., Deborah A. Taira, William H. Rogers, Mary Kosinski, John E. Ware, and Alvin R. Tarlov. 1998. "Linking Primary Care Performance to Outcomes of Care." Journal of Family Practice 47:213-20. Saha, Somnath, Miriam M. Komaromy, R. D. Koepsell, and Andrew B. Bindman. 1999. "Patient-physician Racial Concordance and the Perceived Quality and Use of Health Care." Archives of Internal Medicine. 159:997-1004. Saha, Somnath, Sara H. Taggart, Miriam Komaromy, and Andrew B. Bindman. 2000. "Do Patients Choose Physicians of Their Own Race?" Health Aflairs 19:76-83. Scanlon, Dennis, Charles Darby, Elizabeth Rolph and Hilary Doty. 2001. "The Role of Performance Measures for Improving Quality in

306 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR Managed Care Organization." Health Services Research 36361 9-42. Schlesinger, Mark, Benjamin Druss, and Tracey Thomas. 1999. "No Exit? The Effect of Health Status on Dissatisfaction and Disenrollment from Health Plans." Health Services Research 341547-76. Schulman, Kevin A., Jesse A. Berlin, William Harless, Jon F. Kerner, Shyrl Sistrunk, Bernard J. Gersh, Ross Duke, Christopher K. Taleghani, Jennifer E. Burke, Sankey Williams, John M. Eisenberg, Jos& J. Escarce, and William Ayers. 1999. "The Effect of Race and Sex on Physicians' Recommendations for Cardiac Catheterization." New England Journal of Medicine, 340:6 18-26. Simon, L.P., and A.F. Monroe. 2001. "California Provider Group Report Cards: What Do They Tell Us?'and American Journal of Medical Quality 16:61-70. Smith, David B. 1999. Health Care Divided Race and Healing a Nation. Ann Arbor: University of Michigan Press. Smith, N.A., F? Ley, J.P Seale, and J. Shaw. 1987. "Health Benefits, Satisfaction and Compliance." Patient Education Counsel 10:279-86. Sorensen, Andrew A. and Richard F? Wersinger. 1981. "Factors Influencing Disenrollment from an HMO." Medical Care 19:766-73. Thom, David H. 2001. "Physician Behaviors That Predict Patient Trust." Journal of Family Practice 50:323-28. Todd, K.H., C. Deaton, A. P. D'Adamo, and L. Goe. 2000. "Ethnicity and Analgesic Practice." Annals of Emergency Medicine. 35: 11-16. Todd, K.H., N. Samaroo, and J.R. Hoffman. 1993. "Ethnicity as a Risk Factor for Inadequate Emergency Department Analgesia." Journal of the American Medical Association. 269:1537-39. Van, Ryn Michelle and Jane Burke. 2000. "The Effect of Patient Race and Socio-economic Status on Physicians' Perceptions of Patients." Social Science and Medicine, 509 13-28. Xu, Gang, Sylvia K. Fields, Christine Laine, Jon J. Veloski, Barbara Barzansky, and Carlos J. Martini. 1997. "The Relationship Between the Racelethicity of Generalist Physicians and Their Care for Underserved Populations." American Journal of Public Health 87:8 17-22. Zastowny, Thomas R., Klaus J. Roghrnann, and Gail L. Cafferata. 1989. "Patient Satisfaction and the Use of Health Services: Explorations in Causality." Medical Care 27:705-23. ThomasA. LaVeist is director of the Center for Health Disparities Solutions and associate professor of sociology, and health policy and management at the Johns Hopkins Bloomberg School of Public Health. His research focuses on the impact of race, ethnicity, social class and health and social policy on health status and health services utilization. Amani Nuru-Jeter is a doctoral candidate in the Department of Health Policy and Management, Johns Hopkins University Bloomberg School of Public Health. Her interests include research on racial disparities and income inequality and health.