OBJECTIVE. To evaluate an intervention to prevent falls at a hospital. DESIGN. A quasi-experimental intervention with historical and contemporaneous control groups. SETTING AND PARTICIPANTS. Nursing staff and patients in the medicine service (comprising 2 intervention floors and 2 control floors) at an academic hospital. INTERVENTION. Nursing staff were educated regarding fall prevention during the period from April through December 2005. Data on implemented prevention strategies were collected on control and intervention floors. Mean monthly fall rates were compared over time and between intervention and control floors, using repeated-measures analysis of variance. RESULTS. Postintervention fall knowledge test scores for the nursing staff were greater than preintervention test scores (mean postintervention test score, 91%; mean preintervention test score, 72%; P < .001). Use of prevention strategies was greater on intervention floors than it was on control floors, including patient education via pamphlets (46% vs 15%; P < .001), use of toileting schedules (36% vs 25%; P = .016), and discussion of high-risk medications (51% vs 30%; P < .001). The mean fall rate for the first 5 months of the intervention was 43% less than that for the 9-month preintervention period for intervention floors (3.81 falls per 1,000 patient-days vs 6.64 falls per 1,000 patient-days; P = .043). Comparisons of mean rates for the overall 9-month intervention period versus the 9-month preintervention period showed a 23% difference in the fall rate for intervention floors, but this did not reach statistical significance (5.09 falls per 1,000 patient-days vs 6.64 falls per 1,000 patient-days; P = .182). CONCLUSION. The nursing staff's knowledge and use of prevention strategies increased. Fall rates decreased for 5 months after the educational intervention, but the reduction was not sustained.