Falls risk stratification is complex and difficult to achieve due to the many variables that contribute to patient fall risk. More than 35 factors are believed to contribute directly to inpatient falls. Falls risk assessment tools must be sufficiently sensitive to identify all at-risk patients and sufficiently specific to avoid diverting hospital resources needlessly. At present, more than 15 falls risk assessment tools have been described, and 3 tools have been validated in multiple studies of different patient populations: the Saint Thomas’s Risk Assessment Tool in Falling Elderly Inpatients (STRATIFY), the Morse Falls Scale (MFS), and the Hendrich Fall Risk Model II (H2Model).1 A review of 31 articles, including 22 previously un-reviewed studies, recommended 7 assessment tools, including the Timed Up-and-Go test, STRATIFY, and H2Model.2
Falls risk assessment tools are typically applicable in specific populations, and, in most cases, the ability of specific tools to identify at-risk patients depends on the patient population and other variables that are specific to the locale.1,2 To ensure validity, each tool must be assessed in the location and patient population and in the context of the balance between specificity and sensitivity.
Both reviews present data from numerous studies of falls risk assessment tools and provide information regarding tool performance for specific conditions and patient populations. Highlights include a summary of research that directly compared falls risk tools in one environment; recommendation of tools that are useful in medical-surgical and surgical units; and discussion of a Cochrane review that analyzed 41 intervention studies. For instance, comparison of the H2Model with the MFS and STRATIFY in an Italian study in an acute geriatric unit resulted in sensitivity and specificity of 70% and 61% for the H2Model at a cutoff score of ≥ 5, while the MFS had a higher sensitivity (88%) and lower specificity (48.3%) in the same population. The lower specificity and greater ease of use of the H2Model was deemed more suitable in this patient population.1 Studies that evaluated the STRATIFY, H2M, and MFS in medical and surgical units showed that STRATIFY performed best in medical patients younger than 65 years old, with a sensitivity of 35% to 93% and specificity of 46.7% to 93%.2 However, STRATIFY failed to predict inpatient falls in elderly patients. Under similar conditions, the MFS had a sensitivity of 57% and 88% and a specificity of 73% and 48%. TIn medical and surgical units, no single tool had sensitivity and specificity over 70%, and the H2Model and STRATIFY were better at identifying patients at low risk. Similarly, in the inpatient rehabilitation setting, few tools demonstrated high sensitivity, and STRATIFY had a specificity below 70% and poor sensitivity under these conditions. Variable efficacy of the different tools under diverse conditions highlights the need for validation.
1. Simpson JR, Tosenthal LD, Cumbler EU, Likosky DJ. Inpatient falls: defining the problem and identifying possible solutions. Part I: An evidence-based review. Neurohospitalist. 2013;3(3):135-143.
2. Lee J, Geller AI, Strasser DC. Analytical review: focus on fall screening assessments. Physical Med Rehab. 2013;5:609-621.