ABSTRACT
OBJECTIVE: To design a decision analytic model to help determine the costs associated with various treatment regimens for gastroesophageal reflux disease (GERD). STUDY DESIGN: A decision analytic model incorporating Markov processes was developed to calculate clinical and direct economic outcomes for patients with GERD after 2 years of treatment. PATIENTS AND METHODS: We used retrospective data in the Markov model to generate clinical and economic outcomes. The primary data sources were the 1993 MarketScan claims database, the 1992 National Hospital Discharge Survey, and the clinical literature. RESULTS: Patients with mild GERD (17.6% of patients) contributed 37.8% of costs, while those with moderate to severe disease (14.4% of patients) contributed 49.9% of costs. The remaining 12.3% of costs was spent on the 68% of patients with non-GERD diagnoses. The class of drugs with the highest acquisition cost--proton pump inhibitors--had the lowest total cost per case. The high level of efficacy of these drugs may explain this result. Sensitivity testing showed no evidence that our model's results depended heavily on any one probability or cost factor. CONCLUSIONS: This model showed that patients with moderate to severe GERD were the most expensive cases to treat and that proton pump inhibitors resulted in the lowest total cost per case. Further testing and manipulation of the model are required to gain a better understanding of the trade-offs involved in different options for GERD management.
