OBJECTIVE To analyze the effects of endoscopy and care in a gastroenterology ward on 30-day mortality among Italian patients hospitalized for acute non-variceal upper gastrointestinal hemorrhage (UGIH). METHODS We conducted a population-based study based on administrative data contained in the Regional Hospital Information System (RHIS) for the Lazio Region (Italy). We identified all hospitalizations with a main diagnosis of UGIH during period 2000-2005. Discharge data were analyzed for procedures performed, ward where the patient was cared for, comorbidities, vital status at discharge. Vital status 30 days after admission was cross-checked with the Regional Registry of Causes of Death. Logistic regression models were performed taking into account patients' risk factors (OR and C.I. 95%). RESULTS A total of 13,427 hospitalizations for UGIH (mean patient age, 68 years; 60% males) were identified. The 30-day mortality was 6.9%. Significantly lower rates were observed among hospitalizations that included endoscopy (OR 0.30, 95% C.I. 0.26-0.34), specialist care (OR 0.55, 95% C.I. 0.37-0.82), or both (OR 0.12, 95% C.I. 0.07-0.22). The protective effects of endoscopy and specialist care remained strong after adjustment for potential risk factors. CONCLUSIONS Endoscopy, per se, reduces mortality among patients hospitalized for UGIH, and care in a gastroenterology ward may offer additional protective effects.