Objective The aim of this study was to determine the accuracy of the diagnosis of gastro-oesophageal reflux disease (GORD) by the Reflux Disease Questionnaire (RDQ), family practitioners, gastroenterologists and a test of esomeprazole therapy. Methods This was a single-blind, single-arm study over 3–4 weeks from September 2005 to November 2006. Each symptom-based diagnostic assessment was made blinded to prior diagnoses. Patients were those presenting to their family practitioner with troublesome upper gastrointestinal symptoms (n=308). The RDQ was completed and a symptom-based diagnosis was made by the family practitioner. Placebo esomeprazole was started. Gastroenterologists made a symptom-based diagnosis and then performed endoscopy with 48 h oesophageal pH and symptom association monitoring to determine the presence/absence of GORD. Symptoms were recorded during treatment with 40 mg of esomeprazole for 2 weeks. The main outcome measure was RDQ scoring for the presence of GORD compared with symptom-based diagnosis by family physicians and gastroenterologists. Results GORD was present in 203/308 (66%) patients. Only 49% of the patients with GORD selected either heartburn or regurgitation as the most troublesome symptom. Sensitivity and specificity, respectively, of the symptom-based diagnosis of GORD, were 62% and 67% for the RDQ, 63% and 63% for family practitioners, and 67% and 70% for gastroenterologists. Symptom response to esomeprazole was neither sensitive nor specific for the diagnosis of GORD. Conclusions The RDQ, family practitioners and gastroenterologists have moderate and similar accuracy for diagnosis of GORD. Symptom response to a 2 week course of 40 mg of esomeprazole does not add diagnostic precision. Clinical trial number NCT00291746.