OBJECTIVE. Preventing hospital falls and injuries requires knowledge of fall and injury circumstances. Our objectives were to determine whether reported fall circumstances differ among hospitals and to identify predictors of fall-related injury. DESIGN. Retrospective cohort study. Adverse event data on falls were compared according to hospital characteristics. Logistic regression was used to determine adjusted odds ratios (aORs) with 95% confidence intervals (CIs) for risk factors for fall-related injury. SETTING. Nine hospitals in a Midwestern healthcare system. PATIENTS. Inpatients who fell during 2001-2003. RESULTS. The 9 hospitals reported 8,974 falls that occurred in patient care areas, involving 7,082 patients; 7,082 falls were included in our analysis. Assisted falls (which accounted for 13.3% of falls in the academic hospital and 9.8% of falls in the nonacademic hospitals; P < .001) and serious fall-related injuries (which accounted for 3.7% of fall-related injuries in the academic hospital and 2.2% of fall-related injuries in the nonacademic hospitals; P < .001) differed by hospital type. In multivariate analysis for the academic hospital, increased age (aOR, 1.006 [95% CI, 1.000-1.012]), falls in locations other than patient rooms (aOR, 1.53 [95% CI, 1.03-2.27]), and unassisted falls (aOR, 1.70 [95% CI, 1.23-2.36]) were associated with increased injury risk. Altered mental status was associated with a decreased injury risk (aOR, 0.72 [95% CI, 0.58-0.89]). In multivariate analysis for the nonacademic hospitals, increased age (aOR, 1.007 [95% CI, 1.002-1.013]), falls in the bathroom (aOR, 1.46 [95% CI, 1.06-2.01]), and unassisted falls (aOR, 1.83 [95% CI, 1.37-2.43]) were associated with injury. Female sex (aOR, 0.83 [95% CI, 0.71-0.97]) was associated with a decreased risk of injury. CONCLUSION. Some fall characteristics differed by hospital type. Further research is necessary to determine whether differences reflect true differences or merely differences in reporting practices. Fall prevention programs should target falls involving older patients, unassisted falls, and falls that occur in the patient's bathroom and in patient care areas outside of the patient's room to reduce injuries.