Epoetin alpha (Epo) for use in dialysis patients represents a substantial expenditure for academic medical centers. OBJECTIVE: The purpose of this study was to evaluate the efficiency and accuracy of EPO billing and reimbursement systems administered and dispensed in 8 geographically distributed academic health centers. METHOD: A retrospective chart review was performed for 604 patients across the 8 centers, during the third or fourth quarter of 1989 or the third quarter of years 1990 through 1993. Data collected included HCFA provider number, Medicare number, quantity of Epo prescribed, quantity of Epo recorded as administered for home dialysis, number of doses of Epo not administered as prescribed, reason for missing dose, available hemoglobin and hematocrit values, and relevant information to explain skipped dialysis treatments. The number of units administered or dispensed for home use was matched to an abstract of the HCFA reimbursement files for comparison. RESULTS: Approximately 18% of the total Epo actually administered and/or dispensed during the course of the study was not represented in the HCFA reimbursement data. This ranged from 2% to 45% in the different centers. The total volume of Epo administered or dispensed per center was not related to the percentage of under-representation in the HCFA reimbursement files. CONCLUSION: Epo administered and/or dispensed to Medicare primary dialysis outpatients was under-represented in HCFA reimbursement data. As a result, institutions may be experiencing a significant loss of reimbursement. The exact reasons for this discrepancy are unclear. Institutions should evaluate their Epo reporting/billing policies and procedures to potentially increase revenue recovery they are due.