To compare the efficacy of conventional versus on-demand (symptomatic) treatment of duodenal ulcer, 81 patients were randomized into two groups. Group A (n = 40) patients were treated with ranitidine 150 mg twice daily until complete ulcer healing was achieved. Group B (n = 41) received a similar dose of ranitidine until complete relief of pain was achieved, irrespective of ulcer healing. Recurrence of ulcer in group A was treated with a full course of treatment until complete healing of the ulcer was achieved again, whereas, in group B, treatment was given only until pain recurrence was symptomatically controlled. Endoscopic examination was performed each month. Analysis of the results at 8 wk and 28 wk showed that 1) ulcer healing in group A was significantly superior to that in group B up to 24 wk, but at 28 wk the difference was no longer statistically significant (95% vs 70%), 2) the number of painful days were similar in the two groups, 3) group A patients took treatment for a significantly longer period than those in group B, 4) the cost of treatment per patient in group A was significantly greater than that in group B, 5) the recurrence rate assessed in patients followed for 28 wk after complete ulcer healing was similar in the two groups, and 6) the ulcer-related complications were not significantly different in the two groups. These findings indicate that, although on-demand treatment results in slower ulcer healing, it is not associated with an increase in the duration of pain and incidence of complications. A major advantage of this approach was a significant reduction in the cost of treatment. It is concluded that on-demand treatment is an attractive alternative therapeutic approach in the management of duodenal ulcer disease.