OBJECTIVES: For patients with reflux esophagitis, long-term therapeutic options include proton pump inhibitor (PPI) therapy and/or antireflux surgery. An earlier cost-effectiveness analysis concluded that at 5 yr, medical therapy was less expensive but similarly effective to fundoplication, but the results were sensitive to estimates on quality of life and long-term medication usage, which were derived from "expert opinion." Recently, data from randomized controlled trials addressing these variables have become available. We have incorporated these new data into a revised Markov model to examine the cost-effectiveness of surgical versus medical therapy in patients with severe reflux esophagitis. METHODS: A Markov simulation model was constructed using specialized software (DATA PRO 4.0, Williamstown, MA). Total expected costs and quality-adjusted life-years were calculated for long-term medical therapy and for laparoscopic Nissen fundoplication. Probabilities were obtained from the medical literature using Medline®. Procedural and hospitalization costs used were the average Medicare reimbursements at our institution. Medication costs were the average wholesale price. The analysis was extended over a 10-yr time horizon at a discount rate of 3%. RESULTS: The discounted analysis shows that medical therapy is associated with total costs of $8,798 and 4.59 quality-adjusted life-years, whereas the surgical strategy is more expensive ($10,475) and less effective (4.55 quality-adjusted life-years). The results were robust to most one-way sensitivity analyses. CONCLUSIONS: Long-term medical therapy with PPI is the preferred strategy for patients with gastroesophageal reflux disease and severe esophagitis. Our study highlights the importance of using primary, patient-derived data rather than expert opinion.