OBJECTIVES: Examine patterns of published economic “value messages” for anti-epileptic drugs (AEDs).
METHODS: Using literature review best practices, identified, reviewed, and abstracted data from comparative economic analyses published in English and referenced in PubMed or presented at ISPOR. For each study, documented comparators, “winners” and “losers”, explanation of economic advantage (if any) study sponsor (if any), year published, country of interest, and study design.
RESULTS: We identified 26 studies containing at least one comparative economic “value message” for an AED. A total of 57% (15) were published as manuscripts; 53% (14 of 26) were sponsored by a drug manufacturer (4 manuscripts and 10 conference abstracts); and 38% (10 of 26) were US-oriented. Of the 14 sponsored studies, Ortho-McNeil (topiramate) sponsored 6 (only 1 published; only 1 US-oriented); UCB (levetiracetam) 4; Novartis (carbamazepine, oxcarbazepine) 3; and GSK (lamotrigine) 1. With only one exception (Ortho-McNeil), sponsored studies generated positive messages for sponsors’ products. The 26 studies generated 39 comparative messages. There was at least one “winning” message for 11 of the 13 AEDs studied. Topiramate was the most frequent “winner” (35% of all messages expressed economic superiority of topiramate over comparators). Lamotrigine was the most frequent “loser” (45% of all economic messages). There was at least one message showing economic superiority over lamotrigine for 7 of the 13 AEDs. For generically available AEDs, the explanation for cost savings stemmed from lower drug price, with no evidence of clinical inferiority. For levetiracetam, the explanation for cost-effectiveness stemmed from reduced seizure frequency, a better side effect profile, and improved adherence. The rationale for topiramate’s economic advantages was unclear from conference abstracts.
CONCLUSIONS: Several manufacturers of branded AEDs (Ortho-McNeil, UCB, Novartis) have produced studies describing their drug’s economic value, while others have done very little work in this area. Patterns emerge in methods and comparators.
