A questionnaire was constructed and validated to improve the accuracy of symptom assessment in diagnosing gastro-oesophageal reflux disease (GERD). The GERD questionnaire consisted of four questions describing an upward moving, uncomfortable feeling in the chest frequently accompanied by retrosternal burning that is improved with antacids. It was found that if a patient answered yes to all four questions the likelihood was 85% that erosive oesophagitis would be detected on endoscopy or that pathological gastro-oesophageal reflux on 24-hour pH-monitoring would be documented, or both. The GERD questionnaire was used to identify 269 patients with probable GERD who after one week on placebo entered a 2-week double-blind placebo-controlled study which was completed by 251 patients. Cimetidine (400 mg) b.d. was given to 124 patients and placebo to 127 patients. On diary cards the patients noted the number, the mean duration and the mean severity of GERD symptoms episodes. Cimetidine was significantly superior to placebo in increasing the percentage of symptom-free days, and in reducing the median number of daily symptom episodes. This trial demonstrates that 400 mg cimetidine b.d. is effective in improving GERD symptoms in patients identified by a descriptive, validated GERD questionnaire.