Objective: Under the Pharmaceutical Benefits Scheme, the use of H2‐receptor antagonists (H2A) in the treatment of dyspepsia and heartburn is only subsidised when there is a proven diagnosis of ulcer. This study compared the costs of this Australian practice with a simulation of British practice, which allows unrestricted prescribing of subsidised H2A.Design: Patients with heartburn and/or dyspepsia were prospectively randomised to either a “British” group treated freely at the discretion of their general practitioner without necessarily being investigated or an “Australian” group where use of H2A was allowed only after gastroscopy or a barium meal had demonstrated a peptic ulcer or ulcerative oesophagitis. The patients were followed up for six months and all direct and indirect costs were recorded.Setting: Forty‐nine Sydney general practitioners recruited primary care patients for the study.Patients: Any patient with heartburn or dyspepsia was considered for recruitment; 139 patients entered the study and 137 completed it.Main outcome measures: The outcome measures were the costs of general practitioner consultations, specialist consultations, radiology and gastroscopy, other tests, H2A, other medications, personal costs, and total cost per patient.Results: The cumulative total cost per patient at the end of the study was equivalent in the “Australian” ($392) and “British” ($406) groups. A higher initial cost per patient of H2A in the “British” group was offset by a rapid decrease in the proportion that continued to use H2A and by the cost of specialist consultations and investigations in the “Australian” group.Conclusion: Over a six‐month period the cost of early investigation of heartburn and dyspepsia was equivalent to the cost of a therapeutic trial of H2A.