A Comparison of Two Methods for Estimating the Health Care Costs of Epilepsy


Address correspondence and reprint requests to Floyd J. Frost at Southwest Center for Managed Care Research, Lovelace Respiratory Research Institute, 2425 Ridgecrest Drive SE, Albuquerque, NM 87108, U.S.A. E-mail: ffrost@lrri.org


Summary: Purpose: Previous studies have estimated medical care costs of epilepsy by applying unit costs to estimated utilization or by summing costs for (a) ambulatory care and hospitalizations coded as epilepsy and (b) procedures and drugs specifically associated with the diagnosis or treatment of epilepsy. These methods may underestimate the cost of medical care for epilepsy. Two methods for estimating the medical care costs of epilepsy (“epilepsy-attributable cost method” and “case-control cost method”) were compared.

Methods: The study population was 655 individuals with an epilepsy diagnosis enrolled in a managed care plan in the south-western United States. The epilepsy-attributable costs were determined by summing costs for inpatient and outpatient encounters coded as epilepsy, procedures for the diagnosis or treatment of epilepsy, and drugs used to treat epilepsy. The case-control method determined costs by calculating the difference in total costs between cases and 1,965 age- and gender-matched controls.

Results: The case-control epilepsy costs were $2,923 per case compared with epilepsy-attributable costs of $1,335 per case. The case-control method found statistically significant differences in costs between cases and controls for inpatient care, prescription drugs, and 8 of 11 categories of outpatient care. The largest contributors to the discrepancy between estimates were inpatient care, emergency department care, laboratory tests, and “other specialist” care.

Conclusions: Epilepsy-attributable costs accounted for only 46% of the total difference in costs between epilepsy cases and controls. Persons with epilepsy use more medical services than controls, but a substantial portion of this care is not coded to epilepsy.