INTRODUCTIONOpen access to endoscopy facilities for general practitioners has resulted in increased endoscopic workload and an increase in hospital costs.1 Consequently, it is now under discussion whether the service should only be targeted at a selected group of patients who genuinely need endoscopy.2 Several studies have investigated whether empirical treatment instead of prompt endoscopy could increase appropriate use of endoscopy facilities and decrease costs. 3±7 Regrettably, most of the data are based on decision analysis 3±5 and, to our knowledge, only one randomized clinical trial 6 and one cohort study 7 have prospectively compared an empirical treatment strategy to prompt endoscopy in patients with dyspeptic symptoms. The randomized clinical trial demonstrated that empirical treatment was associated with higher costs and lower patient satisfaction, whereas in the cohort study, an empirical treatment strategy was cheaper and achieved comparable clinical results to the conventional strategy. A pitfall in both studies was that empirical treatment was given with either an H 2 -receptor antagonist (ranitidine) or prokinetics and not with proton pump inhibitors, the latter being the most effective acidsuppressing drugs available. 8,9 Thus in acid-related SUMMARY Background: Cost-effectiveness analysis, Helicobacter pylori research and the development of proton pump inhibitors are having an increasing impact on the management of dyspepsia. However, clinical trials have not always included both H. pylori diagnosis and proton pump inhibitors in their protocols. Methods: Patients who were referred for upper gastrointestinal endoscopy by their general practitioner were randomized to either prompt endoscopy followed by directed medical treatment (conventional group, n 38), or to empirical treatment with omeprazole and, in the case of symptom relapse, serological screening for H. pylori infection followed by eradication therapy in seropositive patients (empirical group, n 42). The study lasted for up to 1 year.