Abstract
A simple model of a Markov chain was used to study the long term outcome of different strategies for the treatment of duodenal ulcer. Maintenance treatment with H2 receptor antagonists surpassed intermittent drug treatment and proximal gastric vagotomy with respect to the relapse free interval and severe postoperative morbidity. With maintenance treatment the rate of complications and the number of deaths related to ulcer were slightly higher than after proximal gastric vagotomy. Nevertheless, because the few deaths from proximal gastric vagotomy occur at the initiation of treatment the loss of life years during maintenance treatment exceeded that of proximal gastric vagotomy only after 20 years. Despite its rarity, severe postoperative morbidity after proximal gastric vagotomy far exceeded that after the few emergency operations which would become necessary in the course of maintenance treatment. The superiority of maintenance treatment over proximal gastric vagotomy remained insensitive to changes in the assumptions underlying the recurrence rate with both treatments and the postoperative morbidity of proximal gastric vagotomy.