Impact of endoscopy on mortality from occult cancer in radiographically benign gastric ulcers. A probability analysis model.

Endoscopy is commonly used in the management of patients with radiographically benign gastric ulcers to detect occult malignancy. Clinical studies examining the cost-effectiveness of using endoscopy in such patients, however, have not been done. To address this issue using probability analysis, a probability tree was designed incorporating the possible clinical courses of patients with radiographically benign gastric ulcers managed with and without endoscopy, and probability estimates for each course were derived by compiling data from the literature. Probability and sensitivity analysis was used to compare the impact on overall mortality rate and cost-effectiveness of six commonly practiced methods of using endoscopy to manage patients with radiographically benign gastric ulcers: (1) all follow-up by upper gastrointestinal x-ray only; (2) endoscopy for nonhealing ulcers only; (3) endoscopy for all ulcers before medical therapy with all follow-up by upper gastrointestinal x-ray; (4) endoscopy for all ulcers after an initial trial of medical therapy; (5) endoscopy for all ulcers before therapy and for nonhealers; (6) endoscopy before therapy, and all follow-up by endoscopy. This analysis predicts that the greatest decrease in mortality rate occurs when endoscopy is used before medical therapy and for all follow-up, reducing the estimated number of deaths per 1000 patients with radiographically benign gastric ulcers from 36.7 with follow-up by upper gastrointestinal x-ray only to 27.2. However, initial endoscopy with all subsequent follow-up by upper gastrointestinal x-ray increased the overall death rate by only a small amount, to 28.0, and was consistently the most cost-effective method, requiring 116 endoscopies and approximately 60,000 diagnostic dollars per additional 5-yr survivor.