Do Race, Ethnicity, and Geographic Location influence rate of Knee Arthroplasty?

Joint replacement surgery, particularly of the knee, is often helpful in relieving pain and improving joint function of moderate to severe osteoarthritis (OA). Evidence exists pointing to racial disparities among those undergoing knee arthroplasty. For example, Escalante et al (Arthritis Rheum 43:390, 2000) found, in an east Texas cohort, that Hispanics were less likely to undergo hip or knee arthroplasty than non-hispanic whites. These differences were independent of socio-economic standing. Additionally, the volume of joint replacement procedures has regional variability. Using a comprehensive, nationwide cohort, Skinner et al (NEJM 349(14):1350, 2003) examine the influence of race, ethnicity, and geographic location on rates of knee arthroplasty.

Methods: The U.S. Medicare database was scrutinized for all primary total knee arthroplasties for the years 1998 through 2000. Subjects were characterized according to sex, self-identified racial or ethnic background (black, Hispanic, or non-Hispanic white), and Hospital Referral Region.

Data were analyzed for regions in which the population of blacks and Hispanics was sufficient to overcome sampling error for small representative groups. Data were analyzed in terms of regional volume of procedures, degree of residential segregation, and income (obtained from year 2000 Census data).

Results: 430,726 knee arthroplasties in 403,251 persons were analyzed.

In 29 of the 30 Hospital Referral Regions with the largest black populations, rates of knee arthroplasties were significantly lower for black men than for white men. For 15 of the 30 regions, the rates of knee arthroplasties were significantly lower for black women than for white women.

In 5 of the 14 Hospital Referral Regions with the largest Hispanic populations, rates of knee arthroplasties were significantly lower for Hispanic men than for white men. No difference was observed between white and Hispanic women in the 14 regions with the highest Hispanic populations. In fact, the rate of knee arthroplasty in Manhattan N.Y. was significantly higher for Hispanic women than for white women (4.4 per 1000 vs. 2.9 per 1000).

Correcting the unadjusted national rates of knee arthroplasty based on geographic variation showed, in general, smaller differences between whites and non-whites.

Higher income and a lower regional level of segregation mitigated the effects of racial/ethnic differences to some extent, although not consistently. For example, for black men living in regions at or above the median income for blacks, a smaller difference in knee arthroplasty rates(blacks compared to whites) was observed than for black men living in regions with the incomes for blacks below the median (2.42 per 1000 vs. 2.79 per 1000).

In contrast, in regions with incomes at or above the median for Hispanics, the rate of knee arthroplasty for Hispanic women was higher than the rate for white women. The reverse was true for Hispanic women in regions below the median income for Hispanics.

Among black women, living in a region of high racial segregation was associated with a larger difference in knee arthroplasty rates between black women and white women.

Conclusions: Racial and ethnic disparities in rates of knee arthroplasty exist nationwide and are at least partially accounted for by geographic region.

Editorial Comments: This study is notable because it is one of the first attempts to consider the effect of geography on rates of health care utilization. Thus, while the investigators demonstrated significant racial/ethnic differences in the rates of knee arthroplasty among blacks, whites, and Hispanics, and between men and women, these differences were significantly attenuated when region of the country was factored in the analysis. Despite the adjustments, however, utilization of knee arthroplasty remained significantly lower for black men, in particular.

Many unanalyzed variables are potentially at play to explain the differences seen and include differences in access to health care, potential racial bias in recommending surgery, and potential racial aversion to surgical intervention. In addition, the study cannot account for regional or racial variation in the prevalence of OA. Also, since the accuracy of self-reporting of Hispanic ethnicity is variable, some of the lines between whites and Hispanics may be blurred in this analysis. Overall, however, the general trends are clear, most notably in terms of the under utilization of arthroplasty by black males. Further studies as to the particular barriers to knee arthroplasty in these subgroups may serve to diminish these differences in accessing a known effective therapy for a common disorder.