Background and aim An early and accurate evaluation by a general practitioner is needed to screen out non-gastroesophageal reflux disease (GERD) patients. A recent questionnaire (QUEST) highlighted problems with specificity and complexity, so the aim of the present study was to design a simplified questionnaire. Methods When admitted to hospital to undergo an upper gastrointestinal endoscopy for suspected GERD, 333 patients completed a 50-item questionnaire requiring 'yes/no' answers to different combinations of questions relating to symptoms of upper gastrointestinal tract conditions (e.g. GERD, ulcers and functional dyspepsia) and psychosomatic symptoms. The endoscopic diagnosis was then correlated with the rate of positive answers to each question. Results Based on the analysis of the 50 items, the 8-10 questions most often answered affirmatively by each of the GERD and non-GERD groups were chosen for the simplified questionnaire. Three draft questionnaires were compiled. After calculating the sensitivity, specificity and accuracy in relation to the diagnosis of GERD and other conditions, it was found that questionnaire B (selection of persons answering 'yes' to at least one of questions 1-5 and exclusion of persons answering 'yes' to at least three of questions 7-10) had a high sensitivity, high specificity and low false positive rate. Conclusion A novel questionnaire was developed. It was designed to detect the symptoms of GERD while simultaneously excluding non-GERD patients. This simplified nine-item simplified questionnaire had a sensitivity of 79.8%, a specificity of 53.6% and an accuracy of 63.4%.