Prospective testing of a scoring system designed to improve case selection for upper gastrointestinal investigation

Abstract

A recently described scoring system designed to assess the individual risk of finding serious pathology in patients referred for upper gastrointestinal investigation has been prospectively tested in 1279 patients undergoing first-time endoscopy and 321 patients undergoing radiologic examination. The scoring system has been confirmed to give a reasonable prediction of the likelihood of finding serious pathology in two hospitals with differing endoscopic practice, and also to be applicable to patients attending for radiology. The system works best at defining a low-risk group (score less than 412, 26% of total) in which the incidence of serious pathology was 3%. All cases of malignancy (n = 55) occurred in patients scoring greater than 464 (50% of total). A simple table is described that allows for the easy calculation of score at a glance without the use of a computer. We believe that this scoring system, which can be implemented in seconds, is the simplest yet described and that it could prove to be a useful educational aid.