OBJECTIVES: To compare the pharmaceutical costs of a new proposal of reimbursement of proton pump inhibitors (PPI) in the treatment of GERD with the present regulation in which PPI's are reimbursed only when endoscopy demonstrates esophagitis.The new reimbursement proposal includes empiric therapy (without endoscopy), symptomatic treatment of non-erosive GERD, and chronic 'on-demand' therapy. METHODS: A decision tree model was developed for treatment of patients with GERD resistant to H2 receptor antagonists. Calculations were performed using MS Excel. Response rates of different therapies and probabilities of findings at endoscopy were derived from literature. Costs from the payer's perspective were calculated for the first 48 weeks of treatment using the mean price of the PPI on the Belgian market on Jan 1st 2003. RESULTS: The present reimbursement system and the new proposal represented a mean 48 weeks cost per case of respectively €351 and €204. Sensitivity analysis revealed that the cost increased to € 255 when the efficiency of empiric treatment rose from 67 % to 80% using a double dose during 8 weeks.The cost related to treatment varied from €112 to €236 using respectively the cheapest and the most expensive PPI ; as compared to the actual treatment this resulted in a cost decrease of respectively €138 and €233. Number of days treated while on 'on-demand' therapy also significantly influenced costs: €171 for 1/7 and €270 for 7/7. CONCLUSIONS: The cost reduction with the new reimbursement proposal is robust to changes in efficiency rate, in PPI-price and in duration of 'on demand' therapy.