OBJECTIVES: This study upgrades a previous CMA of PHT versus LEV for early seizure pharmacoprophylaxis where PHT was seen to be the cheapest alternative from an institutional perspective and thereby the optimal one. That CMA assumed equal efficacy, based on no statistically significant difference in efficacy (patients seizure free for PHT/LEV = 83.3/85.3 percent). CMA, while an early standard form of economic evaluation, has been subjected to a number of critiques, suggesting its inappropriateness. We explore the implications of using the previous data exactly as reported, but developing a more appropriate CEA model. As the LEV alternative was thereby more costly but also more effective, may be cost-effective.

METHODS: A randomized comparative trial provided parameters in the original CMA and here. We reproduced the previous CMA to ensure we used the parameters identically, assuming equal efficacy of the two drugs. A second analysis corrected several errors in the CMA and performed a probabilistic sensitivity analysis (PSA).

RESULTS: The expected values of per patient costs were $151 for PHT and $412 for LEV, exactly as reported in the CMA. The deterministic Incremental cost-effectiveness ratio (ICER) for LEV vs PHT was $13,031 per additional seizure prevented while the PSA produced an ICER of $16,866 per additional seizure prevented.

CONCLUSIONS: This work has shown the value of using CEA over CMA in economic evaluation. The previous CMA indicated an optimal choice of PHT. Our results are less clear. If one is willing to pay more than $16,866 for seizure avoidance, then LEV is the optimal treatment based on these data. We suspect our ICER estimate is biased upward due to LEV cost and utility advantages not incorporated here or in the CMA: seizure treatment cost avoidance, 1 day length of hospital stay advantage and reported long-term outcomes advantages. Further analysis is warranted.