Can this be done 8 pm 4/22

RACE OF PHYSICIAN AND SATISFACTION WITH CARE AMONG AFRICAN-

AMERICAN PATIENTS Thomas A. LaVeist, PhD, and Tamyra Carroll

Baltimore, Maryland

The purpose of this study is to examine predictors of physician-patient race concordance and the effect of race concordance on patients' satisfaction with their primary physicians among African American patients. The specific research question is, do African American patients express greater satisfaction with their care when they have an African American physician? Using the Commonwealth Fund, Minority Health Survey, we conduct multivariate analysis of African American respondents who have a usual source of care (n = 745). More than 21% of African American patients reported having an African American physician. Patient income and having a choice in the selection of the physician were significant predictors of race concordance. And, patients who were race concordant reported higher levels of satisfaction with care com- pared with African American patients thatwere not race concordant. (J NatlMedAssoc. 2002; 94:937-943.)

INTRODUCTION Race relations have been among the most

vexing problems facing American culture, and healthcare has not been immune to this. Through most of the 20th century, healthcare facilities were racially segregated, with African Americans generally receiving suboptimal care. In addition, African American physicians were typically barred from practicing medicine on white patients.1 2 While many of these condi- tions have abated, vestiges of this history re- main. There is substantial contemporary evi-

© 2002. From the Center for Health Disparities Solutions, Bloomberg School of Public Health, Johns Hopkins University. Address reprint requests to Dr. Thomas LaVeist, Johns Hopkins University, Bloomberg School of Public Health, 624 North Broadway, Baltimore, MD 21205; phone (410) 955-3774; fax (410) 614-8964; or direct e-mail to [email protected]

dence of race disparities in access, utilization and quality of care.3"14'15 Among the most commonly proposed solu-

tions to race disparities in quality of care, is the recommendation to increase the number of African American healthcare providers. This suggestion supposes that increasing the num- ber of providers will increase the likelihood that African American patients will have an African American provider. In turn, physician- patient race concordance is expected to lead to better quality care and greater patient satisfac- tion.28 Medical schools have responded to this proposition by increasing the production of minority physicians.4'5 Yet, although quite pop- ular, the race concordance hypothesis has re- ceived only limited scrutiny. In this paper, we address the question, do African American pa- tients express greater satisfaction with their

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 94, NO. 11, NOVEMBER 2002 937

DOCTOR RACE AND PATIENT SATISFACTION

care when they have an African American phy- sician?

METHODS Data for this analysis came from the 1994

Commonwealth Fund Minority Health Survey (MHS). The MHS is a sample within the con- tiguous United States of adults 18 years of age and older residing in households with tele- phones.6 Interviews were conducted via tele- phone using random digit dialing. African Americans were over-sampled in the MHS, re- sulting in a total sample of 1048 African Amer- ican respondents. The present analysis exam- ines the subsample of 745 African American respondents who indicated that they had a usual source of care.

Dependent Variables The first set of analyses examined predictors

of race concordance, and the second set exam- ined patient satisfaction. Race concordance was specified as a binary variable indicating that the respondent's race is concordant with the race of the physician. Patient satisfaction is assessed by a five-item scale. Respondents were asked how good their doctor is at: (1) providing good healthcare overall; (2) treating them with dig- nity; (3) making sure the patient understands what he/she has been told; (4) listening to health problems; (5) being assessable by phone or in person. The responses were on a scale of 1 to 4, with 1 being excellent and 4 being poor. In the regression analysis, responses to the five items were summarized to form an index rang- ing from 5 to 20. This index had a Chronbach's Alpha of 0.81.

Independent Variables In analysis predicting race concordance,

physician choice was the primary independent variable. Physician choice was measured by the question, "How much choice you do have in where you go for medical care?" Respondents indicating that they have a "great deal" or "some" were coded as having a choice, and responses of "very little"~or "no choice" were

coded as no choice. In analysis of patient satis- faction, physician-patient race concordance was the primary independent variable.

Covariates The covariates were sex, age, income, educa-

tion, and health insurance. Sex was specified as a binary variable indicating male. Age was spec- ified as a set of binary variables indicating age: 18-30, 31-40, 41-50, 51-65, and 66-94. An- nual income was specified as a continuous vari- able. Education was specified as a set of binary variables indicating less than high school grad- uate, some college, college graduate, and more than college graduate. Health insurance was specified as a set of binary variables indicating private insurance, Medicare, Medicaid, and un- insured.

RESULTS Table 1 presents distributions of the vari-

ables included in the analysis. The table shows that the respondents were evenly divided by gender, with 50% of the respondents being male. Age of the respondents was spread fairly evenly within a range of 18 to 94 years of age. The median age fell within the range of be- tween 41 and 50 years old. And respondents over age 66 represented the smallest category (12.7%). The median income for the sample was between $25,001 and $35,000 per annum, with 91.4% of respondents reporting a salary under $75,000.

Eighty-two percent of respondents had ob- tained at least a high school education and 50.8% received some higher education. More than 75% of respondents had private health insurance and slightly less than one-fifth had Medicare. Just over 15% of the respondents had Medicaid and 10% were uninsured. Fi- nally, nearly two-thirds of respondents re- ported having a choice for their physician.

Table 2 examines the distribution of physi- cian's race among African American respon- dents. The table shows that nearly 22% of re- spondents had an African American doctor. The largest percentage of respondents (58.5%)

938 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 94, NO. 11, NOVEMBER 2002

DOCTOR RACE AND PATIENT SATISFACTION

Table 1. Sample Descriptive Statistics

Variable Percent Male 50.0 Age

18-30 20.4 31-40 27.0 41-50 18.9 51-65 20.9 66-94 12.7

Income <$7,500 13.2 $75,001-$15,000 12.6 $15,001-$25,000 19.3 $25,001-$35,000 1 8.6 $35,001 -$50,000 17.4 $50,001-$75,000 10.3 $75,001-$100,000 5.8 >$ 100,000 2.8

Education <HS grad 17.7 HS grad 31.6 Some college 27.0 College grad 13.8 Post college 10.0

Insurance* Private 75.6 Medicare 19.9 Medicaid 15.4 Uninsured 10.8

Doctor choice 65.3

*Insurance status categories are not mutually exclusive. For example, a respondent can have Medicare and a private healthcare policy. Also, one could be dually eligible for Medicare and Medicaid.

reported having a white physician. About 10% of respondents had Asian or Pacific Islander physicians and 9.7% had a doctor who was of Hispanic descent or another ethnic group (His- panic physicians were not analyzed as a sepa-

Table 2. Race of Physician Among African American Respondents who Reported Having a Usual Source of

Care

Physician's race (n=745) White 436 (58.5%) Black 162 (21.7%) Hispanic/other 46 (6.5%) Asian/pacific islander 75 (10.1 %)

Table 3. Logistic Regression Analysis of Predictors of Physician-patient Race Concordance, Odds Ratio (95%

Confidence Interval)

Female B Male .863 (.593,1.25)

Age 18-30 B 31-40 .41 (.24, .71) 41-50 .47 (.27, .83) 51-65 .53 (.31, .92) 66-94 .25 (.11, .1)

Income 1.13 (1.0, 1.26) Education

K-12 B High School .62 (.34, 1.13) Some College .73 (.39, 1 .36) College Graduate .64 (.31, 1.30) Post College Degree .64 (.29, 1.38)

Insurance Private B Medicare 1.43 (.75, 2.74) Medicaid .69 (.37, 1.32) Uninsured 1.19 (.59, 2.39)

Doctor Choice 1.91 (.1.00, 3.64) Model Statistics Hosmer and Lemeshow X2 = 4.71

df = 8 p = .00

rate category because their total numbers were too small, only 17 respondents reported having a Hispanic physician).

Table 3 presents the results of logistic regres- sion models examining predictors of race con- cordance. The table shows that patient's age, income, and having a choice of physician are significant predictors of physician-patient race concordance. Younger respondents were more likely to be race concordant. Higher income was associated with a greater likelihood of race concordance. And, respondents who report they have the ability to choose their own phy- sician had nearly double the odds of being race concordant with their physician compared with patients that did not have choice.

In Table 4, the analysis turns to an assess- ment of patient satisfaction. The table displays bivariate analysis of each item comprising the five-item patient satisfaction scale arrayed by race of physician. The general pattern among

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 94, NO. 11, NOVEMBER 2002 939

DOCTOR RACE AND PATIENT SATISFACTION

Table 4. Patient Satisfaction, Percent responding excellent

Race of Physician

Asian/ How would you rate your doctor on the job he/she is African Hispanic/ Pacific

doing.... American White Other Islander Total In providing healthcare overall?* 67.3% 49.8% 68.1% 56.8% 56.0% In treating you with respect and dignity?* 78.3% 59.5% 65.7% 65.3% 64.8% Making sure you understand what you are told about your 69.6% 60.0% 71.4% 60.0% 63.1% medical problems?

Listening to your health concerns and taking them 70.2% 52.8% 77.1% 54.1% 59.0% seriously?*

In being accessible either by phone or in person?* 57.1% 41.0% 54.3% 44.6% 46.1%

* = p-value for x2 test < .001

the variables is that African American respon- dents reported the greatest satisfaction with Af- rican American physicians and the least satis- faction with white physicians. Respondents were more then 35% more likely to rate African American physicians as excellent at "providing healthcare overall," compared with white phy- sicians. For all physicians, the lowest rating was for "being assessable by telephone or in per- son."

Table 5 displays ordinary least squares (OLS) multiple regression analyses examining race concordance as a predictor of patient sat- isfaction. Model 1 tests the bivariate effect of race concordance on patient satisfaction, and Model 2 tests the relationship with adjustments for a set of covariates. The table shows that race concordance is an important predictor of pa- tient satisfaction. Patients who have white or Asian/Pacific Islander physicians were signifi- cantly less satisfied with their care, compared with African Americans who are race concor- dant. Model 2 adjusts for covariates. The model

shows that the significant effect of physician race on patient satisfaction persists after model adjustment. The model also shows statistically significant controls for patients age 51 to 65, high school graduates, those with post-gradu- ate educations, and those having the ability to choose their doctor.

DISCUSSION We conducted analysis of the 1994 Common-

wealth Fund Minority Health Survey to explore the correlates of physician-patient race concor- dance and to determine whether race concor- dance was predictive of patient satisfaction. More than one-half of the African American respondents reported having a white physician. However, this is likely reflective of the fact that whites comprise the overwhelming majority of all physicians in the US. Our analysis found that the percentage of African American phy- sicians caring for African American respon- dents (21.7%) was much higher than the per- centage of African American physicians in the general population (3.6%). This confirms pre- vious analyses showing that African American physicians are more likely to practice in African American communities, compared with doc- tors of other race/ethnic groups.5,7-'2

Race concordance among African Ameri- can patients appeared to be a matter of choice rather than merely a byproduct of constrained options caused by geographic limitations. Respondents who reported hav- ing the ability to choose their own physician were significantly more likely to have an Af- rican American physician. Moreover, more affluent African American respondents were more likely to have an African American phy-

940 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 94, NO. 11, NOVEMBER 2002

DOCTOR RACE AND PATIENT SATISFACTION

Table 5. Ordinary Least Squares (OLS) Regression of Predictors of Patient Satisfaction-race/Ethnic Specific Analysis, Standardized Coefficient (P-value)

Model 1 Model 2 Physician race

Black White -.166 (p = .0001) -.55 (p = .0002) Asian/pacific islander -.093 (p=.02) -.098 (p = .02) Hispanic/other .026 (p = .53) .031 (p = .44)

Female B Male - .024 (p = .53) Age

18-30 31-40 .016 (p = .75) 4 1-50 .037 (p = .45) 51-65 .128(p= .011) 66-94 .046 (p = .44)

Income .023 (p= .61) Education

K-12 High school graduate .115 (p = .04) Some college .088 (p = .13) College graduate .039 (p = .46) Post graduate degree .131 (p = .012)

Insurance Private B Medicare .045 (p = .42) Medicaid .001 (p = .98) Uninsured .059 (p = .13)

Patient does not have ability to choose doctor B Patient has ability to choose doctor .169 (p = .0001) Model statistics Adj R2 = .024 Adj R2 = .064

F = 6.69 F = 3.83 p =.000 p =.000

sician, compared with less affluent persons. These respondents, presumably, have a greater choice of providers.

While physician-patient race concordance leads to greater patient satisfaction, there is room for improvement. As Table 4 shows, Af- rican American respondents were more satis- fied with Hispanic and other race physicians at making sure patients understood what they were being told and listening to patient's health concerns and taking them seriously. Ad- ditionally, only 57.1% of respondents reported that African American physicians were doing an excellent job at being accessible by tele- phone or in person. We point out that the Hispanic/other physician category is somewhat

of a miscellaneous category, so findings related to this group must be interpreted with caution. However, it is meaningful that African Ameri- can physicians did not obtain the highest rat- ings for all categories. Moreover, Chen et al.,13 demonstrated that African American physi- cians were as likely as white physicians to fail to refer African American cardiac patients for cor- onary Angiography when the procedure was indicated. Thus, while African American pa- tients are more satisfied with their care, it is not clear that all dimensions of quality are en- hanced by race concordance. For example, are African Americans more likely to receive pre- ventive health services? Are their health ser- vices utilization patterns different? Is their com-

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 94, NO. 11, NOVEMBER 2002 941

DOCTOR RACE AND PATIENT SATISFACTION

pliance better? These questions remain to be explored in future research.

Nevertheless, in spite of some shortcomings in satisfaction among race concordant African Americans, our finding of greater patient satis- faction supports the need for the continuation of efforts to increase African American physi- cian production. One might interpret these results as support-

ing a return to racial segregation in medicine; that is, that patients are better off with physi- cians who are of their own racial or ethnic group. However, we would strenuously resist this interpretation. Instead, we favor providing patients with an array of options that maxi- mizes choice, and we view the race of physician as merely one factor to be considered among many others. As Table 5 shows, patient choice is also highly associated with patient satisfac- tion. As to the matter of whether increasing the

number of African American physicians will lead to better health outcomes for African Americans, an interpretation of our findings within the context of other related literature leads us to conclude that it most likely will. Increasing the number of African American physicians will expand the opportunities for African American patients who choose to be race concordant. Our findings suggest that there will be greater satisfaction among African Americans who are inclined to choose doctors in concordance with race.

Previous findings demonstrate that African American physicians are more likely to practice in minority and under-served communities.7-"1 Studies of patient satisfaction and other related concepts (such as patient-centeredness) find improved outcomes in race concordant physi- cian-patient pairs.'6 There is a very limited lit- erature on race concordance and medical out- comes,'3 but the one available study indicates that African American physicians are as likely as white physicians to under-refer African Ameri- can patients for heart surgery. Chen's'3 find- ings are inconsistent with the findings in the present analysis, as well as other studies,'6"7

and may be anomalous. Clearly, there is a need for further study to confirm or refute Chen's findings. However, whether or not future re- search confirms Chen, patient satisfaction is an important outcome in its own right. In addition to its intrinsic importance for individual health- care consumers and third-party payers, patient satisfaction also is an important determinant of health-related outcomes, such as health ser- vices utilization,'8'19 decision to switch to an- other health plan,2s23 compliance with medical regimen,24 and the decision to initiate malprac- tice suits.25 Moreover, the Institute of Medicine recommends including patient satisfaction as an indicator of quality of care.26'27

As the African American population contin- ues to grow and becomes increasingly affluent, the demand for African American physicians will grow, as well. However, as previous analyses have shown, the current rate of African Amer- ican physician production, will not keep pace with future demand.5 As such, alternatives such as effective "cultural competency" training, are needed if we are to achieve the objective of eliminating race disparities in health, improve quality of care among African American pa- tients, and make the healthcare system respon- sive to the changing racial demographics of our nation.

AKNOWLEDGEMENT This research was supported by a grant from the Com-

monwealth Fund to Dr. LaVeist.

REFERENCES 1. Gamble VN. A legacy of distrust: African Americans and

medical research. AmJPrev Med. 1993;9(6 Suppl):35-38. 2. Smith DB. Health Care Divided: Race and Healing a Nation.

Ann Arbor, MI: University of Michigan Press, 1999. 3. Mayberry RM, Mili F, Ofili E. Racial and ethnic differ-

ences in access to medical care. Med Care Res Rev. 2000;57(Suppl 1):108-145.

4. Carlisle DM, Gardner JE, Honghu Liu. The entry of underrepresented minority students into US medical schools: an evaluation of recent trends. AmJPub Hlth. 1998;88:1314-1318.

5. Libby DL, Zhou Z, Kindig DA. Will minority physician supply meet US needs? Health Affairs. 1997;16:205-214.

6. Hogue, Carol JR, Hargraves MA. The Commonwealth Fund Minority Health Survey of 1994: an overview, in Carol JR, Hogue, Hargraves MA and Collins KS (eds.), Minority health in America: findings and policy implicationsfrom the Commonwealth Fund

942 JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 94, NO. 11, NOVEMBER 2002

DOCTOR RACE AND PATIENT SATISFACTION

Minority Health Survey, Baltimore and London: Johns Hopkins University Press, 2000.

7. Cantor JC, Miles EL, Baker LC, Barker DC. Physician service to the underserved: implications for affirmative action in medical education. Inquiry. 1996;33:167-180.

8. Komaromay M, Grumbach K, Drake M, et al. The role of Black and Hispanic physicians in providing health care for un- derserved populations. New EngJ Med. 1996;334:1305-1310.

9. Moy E, Bartman BA. Physician race and care of minority and medically indigent patients. JAMA. 1995;273:1515-1520.

10. Keith SN, Bell RM, Swanson AG, Williams AP. Effects of affirmative action in medical schools: a study of the class of 1975. New EngJMed. 1985;313:1519-1525.

11. Lloyd SM Jr., Johnson DG. Practice patterns of Black physicians: results of a survey of Howard University College of Medicine alumni. J Nat Med Assoc. 1982;74:129-141.

12. Rocheleau B. Black physicians and ambulatory care. Pub Hlth Repts. 1978;93:278-282.

13. Chen J, Rathore SS, Radford MJ, Wang Y, Krumholz HM. Racial differences in the use of cardiac catheterization after acute myocardial infarction. New EnglJMed. 2001;344:1443-1449.

14. Geiger HJ. Racial and ethnic disparities in diagnosis and treatment: a review of the evidence and a consideration of causes. In Unequal Treatment; Confronting Racial and Ethnic Disparities in Health Care. 2002 IOM. Washington, DC: National Academy Press.

15. Kressin NR, Peterson IA. Racial differences in the use of invasive cardiovascular procedures: review of the literature and prescription for future research. Ann Intern Med. 2001;135:352- 366.

16. Cooper-Patrick L, Gallo Jj, Gonzales Jj, et al. Race, gen- der, and partnership in the patient-physician relationship.JAMA. 1999;282:583-589.

17. Saha S, Komaromy MM, Koepsell RD, Bindman AB. Patient-physician racial concordance and the perceived quality and use of health care. Arch Int Med. 1999;159:997-1004.

18. Zastoway TR, Roghmann KJ, Cafferata GL. Patient satis- faction and the use of health services: explorations in causality. Med Care. 1989;27:705-723.

19. Roghmann KJ, Hengst A, Zastoway TR. Satisfaction with medical care: its measurement and relation to utilization. Med Care. 1979;17:461-479.

20. Murray BP, Dwore RB, Gustafson G, Parsons RJ, Vor-

derer LH. Enrollee satisfaction with HMOs and its relationship with disenrollment. Managed Care Interface. 2000;13:55-61.

21. Allen HM Jr., Rogers WH. The consumer health plan value survey: round two. Hlth Aff 1998;17:265-268.

22. Hennelly VD, Boxerman SB. Disenrollment from a pre- paid group plan: a multivariate analysis. Med Care. 1983;21:1154- 1167.

23. Sorensen AA, Wersinger RP. Factors influencing disen- rollment from an HMO. Med Care. 1981;19:766-773.

24. Smith NA, Ley P, Seale JP, Shaw J. Health benefits, satisfaction and compliance. Pat Educ Counsel. 1987;10:279-286.

25. Penchansky R, Macnee C. Initiation of medical malprac- tice suits: a conceptualization and test. Med Care. 1996;34:280- 282.

26. Hurtado MP, Swift EK, CorriganJM (Eds.). Envisioning the National Health Care Quality Report. 2001 IOM. Washington DC: National Academy Press.

27. Institute of Medicine, Committee on Quality of Health Care in America. Crossing the Quality Chasm: A New Health System for the 21st Century 2001. IOM. Washington DC: National Acad- emy Press.

28. LaVeist TA, Nuru-Jeter A. Is doctor-patient race concor- dance associated with greater satisfaction with care? J Hlth Soc Behv. (forthcoming)

We Welcome Your Comments

Journal of the National Medical Association welcomes your Letters to the Editor about articles that appear in the JNMA or issues relevant to minority health care.

Address correspondence to Editor-in-Chief, ]NMA, 1012 Tenth St, NW, Washington, DC 20001; fax (202) 371-1162; or ktaylor @nmanet.org.

JOURNAL OF THE NATIONAL MEDICAL ASSOCIATION VOL. 94, NO. 1 1, NOVEMBER 2002 943