OBJECTIVE: To estimate the cost of falls in LTCFs. METHODS: The study employed a non-randomized, before and after comparison with control group design. A multi-facility long-term care company provided data from residents institutionalized between January 1, 2002 and October 30, 2004. Data included Minimum Data Set (MDS) observations, Resource Utilization Group (RUG) classifications, and demographics. An index date was assigned to each resident to identify pre-and post-periods. The index date was defined as the date of the first fall for fallers and as the date of the fifth MDS measurement for non-fallers. Direct medical cost estimates were based on MDS measures of hospital, emergency room, and physician utilization and on average Medicare reimbursement rates. Costs related to changes in resident functioning were estimated from RUG payment rates. Total reimbursement per resident per day (PRPD) was calculated as the sum of RUG and medical service reimbursements. Fall-related costs were estimated by comparing between-group differences in pre- to post-index period changes in reimbursement. Regression analysis was used to control for between-group differences. The dependent variable was the natural log of post-period total reimbursement. Independent variables included group, pre-period reimbursement, post-period length of stay, age, gender, race, and severity of illness as measured by a modified Charlson Comorbidity Index. RESULTS: The sample included 1298 fallers and 1509 non-fallers. Fallers had substantially more fractures and higher medical services utilization in the post-period than non-fallers. Total reimbursement for fallers decreased from $107 to $37 PRPD compared to a decrease from $98 to $24 for non-fallers. Regression analysis indicated that reimbursement in the post-period was 40% higher for fallers than non-fallers after controlling for demographic and disease differences and pre-period reimbursement. CONCLUSION: Falls in LTCFs result in substantial costs, primarily due to higher hospitalization rates.