Incremental Cost-Effectiveness Analysis Comparing Rofecoxib with Nonselective NSAIDs in Osteoarthritis Ontario Ministry of Health Perspective John K. Marshall, 1 James M. Pellissier, 2 Cheryl L. Attard, 3 Sheldon X. Kong 4 and Michael A. Marentette 5 1 Division of Gastroenterology, Department of Medicine, McMaster University, Hamilton, Ontario, Canada 2 Merck Research Laboratories, Blue Bell, Pennsylvania, USA 3 Innovus Research Inc., Burlington, Ontario, Canada 4 Merck & Co., Inc., Whitehouse Station, New Jersey, USA 5 Merck Frosst Canada Ltd, Kirkland, Quebec, Canada Abstract Background: Clinical trials have shown rofecoxib, a selective inhibitor of cyclo- oxygenase-2, to be associated with fewer gastrointestinal complications than non- selective nonsteroidal anti-inflammatory drugs (NSAIDs). Objective: To evaluate the potential clinical and economic consequences of rofe- coxib prescription in Ontario, Canada, for patients with osteoarthritis (OA) aged >65 years who did not respond to paracetamol (acetaminophen) therapy. Design: Decision analytic modelling study. Methods: A model was constructed to compare rofecoxib and nonselective NSAIDs with respect to their gastrointestinal complications in patients with OA. The model had a 1-year horizon and considered direct medical costs from the perspective of the Ontario Ministry of Health. Event rates were estimated from a pooled analysis of 8 phase IIb/III clinical trials. The number of perforations, ulcers and bleeds (PUBs) with each strategy was used as the primary measure of effec- tiveness. Results: In the base-case scenario, the expected total cost per patient-day on nonselective NSAIDs was 1.60 Canadian dollars ($Can) versus $Can1.67 on rofecoxib (1999 values). Rofecoxib was associated with 0.0109 fewer PUBs per patient per year. The incremental cost to avoid 1 additional PUB by substituting rofecoxib for nonselective NSAIDs was $Can2247. The rofecoxib strategy be- came dominant if a gastroprotective agent was prescribed to more than 27.5% of the patients receiving nonselective NSAIDs. Conclusion: For patients with OA aged >65 years in whom paracetamol therapy has failed, rofecoxib may represent a cost-effective alternative to nonselective NSAIDs. Increased costs for drug acquisition are offset, in part, by avoidance of gastrointestinal complications and reduced use of gastroprotective agents. Rofe- coxib may offer increased benefit among patients at a higher risk of serious gastrointestinal events. ORIGINAL RESEARCH ARTICLE Pharmacoeconomics 2001; 19 (10): 1039-1049 1170-7690/01/0010-01039/$22.00/0 © Adis International Limited. All rights reserved.