The goal of this study was to assess recombinant human erythropoietin (EPO) therapy among Medicare-entitled black and white in-center hemodialysis patients in order to replicate an earlier report of racial disparity in access to EPO.

Data were derived from patient medical records, from April to December 1992, which were reviewed as part of the USRDS Case Mix/Adequacy Study. Patients were selected by following a detailed protocol in order to ensure that a national random sample of the target population was identified. Patients on hemodialysis during the period ftom 1990 to 1991 were included in this study. (Data from the previous study were derived from Health Care Financing Administration claims files.)

The study found no racial differences in access to treatment with EPO. Using the HCFA data to identify EPO use by patients in the current data set, the relative odds (RO) that black patients received EPO compared with white patients was 0.93 (95% confidence interval=0.80, 1.10); using medical records to identify EPO use, the RO was 0.94 (95% confidence interval=0.80, 1.11); using the HCFA data to identify EPO, but controlling for comorbid factors and health status from medical records, the RO was 1.01 (95% confidence interval=0.86, 1.18); and using the medical records to identify EPO and control for medical factors, the RO was 1.03 (95% confidence interval=0.83-1.23).

Income had no influence on access to EPO, the RO was one (1.00) for $5000 differences in median family income as per census data by ZIP code area. This finding is in marked contrast to other studies reporting an income-race effect on access to higher technology medicine.

A higher percentage of black patients (55.4%) than white patients (47.8%) had a hematocrit level below 29 (a measure of anemia). The authors suggested that low hematocrits may indicate reduced access to EPO therapy, a lower dose of dialysis regardless of EPO status, inadequate dose of EPO, or a lower response to EPO. Preliminary analysis on a subset of patients showed black patients did receive a 5% lower dose of dialysis than white patients.

The authors suggest several potential reasons for the different results found in this study versus the previous one. First, the earlier report used HCFA EPO reimbursement records rather than medical records, and the sensitivity of the HCFA data was only 75% according to an analysis included with the present study. Second, the present study was conducted several years after the initial study; so, it is possible that the factors influencing the racial discrepancy previously reported have since been resolved. Third, the previous study did not control for comorbid factor; thus, it is possible that such factors and overall health status explained the racial patterns previously reported.

The authors conclude that “these results indicated that an equal fraction of prevalent black and white hemodialysis patients were undergoing EPO therapy in 1990 and 1991, but there is a suggestion of a somewhat greater need for EPO therapy among blacks according to the findings of a lower hematocrit in blacks receiving and not receiving EPO.”