Due to the significant expense of obtaining frequent endoscopy and pH monitoring measures as outcome variables available for use in a multihospital clinical trial of gastroesophageal reflux disease, and the lack of a suitable inexpensive index of disease activity, evaluated for both reliability and validity, the study planning committee decided to develop an index of gastroesophageal reflux disease activity in a pilot study--to precede the clinical trial. In particular, the purpose of the pilot study was to find a reliable, valid, and inexpensive index of gastroesophageal reflux disease which could be obtained independently of the treating physician and used as an outcome variable in the clinical trial. This paper describes the pilot study and the statistical methodology used to derive and evaluate a gastroesophageal reflux disease activity index model. In addition, the results of the activity index's use in the subsequent clinical trial's longitudinal analyses are presented. Comparisons with the more expensive, and thus less frequently obtained, endoscopy and pH monitoring outcome variables are described.