The effect of insurance status on use of recombinant erythropoietin therapy among end-stage renal disease patients in three states.

Recombinant human erythropoietin (rHuEPO) has been demonstrated to be effective in ameliorating anemia among persons with chronic renal failure, and is associated with improved functional status and quality of life. Access to rHuEPO has been examined by a variety of clinical, demographic, geographic, and facility characteristics. However, rHuEPO utilization based on insurance status has not been previously examined. All Medicare and Medicaid prevalent end-stage renal disease (ESRD) patients receiving dialysis services in California, Georgia, and Michigan in December 1991 were identified using state and federal administrative program data. The population in each state was stratified by insurance status as follows: Medicare-entitled, Medicare/Medicaid dually entitled, and Medicaid-only entitled. Insurance coverage of the ESRD population by Medicaid, as either a primary or secondary payer, differed greatly by state. In December 1991, the proportion of Medicaid-only and Medicaid/Medicare dually eligible dialysis patients ranged, respectively, from 8% and 43% in California, to 3% and 26% in Michigan, and to 3% and 18% in Georgia. Compared with the Medicare-entitled population, the Medicaid/Medicare dually eligible and Medicaid-only populations disproportionately comprised women, black patients, and individuals younger than 20 years. Using Lee's two-stage binary logit model, dual-eligibility was found to be associated with an increased access to rHuEPO. Compared with their state-specific, dually eligible counterparts, the odds of receiving rHuEPO was lower for Medicare-entitled patients in California (odds ratio [OR], 0.84; 95% confidence interval [CI], 0.76,0.93) and Georgia (OR, 0.65; 95% CI, 0.53,0.80), and lower for Medicaid-only patients in Georgia (OR, 0.02; 95% CI, 0.01,0.05) and Michigan (OR, 0.34; 95% CI, 0.23,0.52). We hypothesize that the absence of substantial copayments associated with rHuEPO, approximately $1,000 per year for a portion of Medicare-entitled patients, resulted in increased access among the dually eligible ESRD population. Dosing of rHuEPO was associated primarily with patient hematocrit level (P < 0.0001) and was unrelated to insurance status. Regardless of insurance status, an unexpectedly large number of Medicare prevalent dialysis patients receiving rHuEPO in each state (31%, 42%, and 41% in California, Georgia, and Michigan, respectively) had hematocrit values lower than 0.28, indicating inadequate treatment of anemia. Eleven percent of all patients receiving rHuEPO in California and nearly 20% in Georgia and Michigan were deemed to be severely anemic (hematocrit < 0.25). The wide variability in access to rHuEPO among the Medicaid-only populations may be indicative of state-specific differences in Medicaid prior approval, copayments, and other drug restrictions. We conclude that the Medicaid-only ESRD population excluded from Medicare coverage is particularly vulnerable to cost-containment measures that focus on expensive technologies such as rHuEPO.