Purpose: There are approximately 600,000 individuals each year discharged from US hospitals with a diagnosis of peptic ulcer disease. In 1994, the NIH disseminated guidelines recommending that all patients with documented ulcers be tested for H. pylori infection and treated if present. In addition, NSAID use has been identified as the cause of up to 50% of ulcers in the Medicare population. The purpose of this study is to examine state-wide and interstate variation in the process of care, and to evaluate the quality of care for hospitalized Medicare beneficiaries with peptic ulcer disease. Methods: Five Peer Review Organization (PRO) Quality Improvement collaborators (Colorado, Connecticut, Georgia, Oklahoma and Virginia) used 1995 Medicare claims files to select state-wide samples of inpatients with a principal diagnosis of peptic ulcer during the period from January, 1, 1995-June, 30, 1995. Each collaborator selected 100% of the claims until 550 cases were selected in each state. Quality of care indicators were developed by a multi-state group with content experts and include: the percent of ulcer patients screened for H. pylori, the percent of patients who had a biopsy who received a tissue test, the percent of H. pylori positive patients who received appropriate therapy, the percent screened for pre-admission NSAID use, and the percent counseled about the risks of NSAID use. Statistical analysis was performed using Pearson's Chi-squared test. Results: 2,621 patients were eligible for medical record review. There was a median age of 78, with 43% of the sample male. Inter-rater reliability testing for chart abstraction resulted in a mean Kappa value of .81. Only 57% of hospitalized Medicare beneficiaries with PUD were screened for H. pylori, and only 73% of those testing positive were treated with appropriate anti-H, pylori therapy. 84% of patients who had endoscopic biopsies taken received a tissue test for H. pylori. 74% of patients were screened for pre-admission NSAID use, but only 24% and 2% were counseled about NSAID use and the specific ulcer risks of NSAID use respectively. There was statistically significant regional variation for four of the six quality indicators: H. pylori screening, screening for non-steroidal anti-inflammatory drug use, counseling about NSAID use, and counseling about ulcer risks associated with non-steroidal anti-inflammatory drug use. Conclusions: Well accepted guidelines for the management of NSAID and H. pylori related PUD exist, but are not always observed. There is an opportunity to improve the quality of care by improving screening for H. pylori in hospitalized Medicare beneficiaries with PUD. Substantial opportunities also exist to improve counseling about NSAIDS and their ulcer risks. Improving H. pylori screening and treatment in those infected may prevent hospital admissions, ulcer recurrences and complications of ulcer disease.