All gastroscopies and colonoscopies performed in two U.K. teaching hospitals over a period of one year were audited to investigate whether endoscopic reporting of gastroscopies and colonoscopies by different endoscopists is consistent. Endoscopic diagnoses were retrieved from the hospitals' endoscopy databases. The results of 1814 colonoscopies and 2127 gastroscopies were analysed using chi2 (Chi squared). The frequency of reporting common diagnoses was variable and the differences between specialist endoscopists were highly significant, including for important conditions such as peptic ulceration (range 2-10%, p = 0.001) and colonic polyps (16-45%, p < 0.001). There is a large variation in the frequency of the diagnoses reported by different endoscopists. This is unlikely to be explained by casemix or chance. This may have major implications for the health of patients. More emphasis must be placed during training on the correct interpretation of endoscopies.