Psychometric evaluation of the Hendrich Fall Risk Model

The aim of this paper is to report a psychometric evaluation of the Hendrich Fall Risk Model. Thoroughly developed and tested instruments for assessment of fall risk are needed to identify patients at risk of falling, to enable the implementation of preventative measures. Data from 1977 patients/residents in 45 nursing homes and 7197 patients from 47 hospitals were evaluated in a cross-sectional survey. The internal consistency of the Hendrich Fall Risk Model was examined using the Kuder-Richardson Test. The dimensions of the model were revealed by exploratory factor analysis and the Care Dependency Scale was used to investigate construct validity. Using Spearman Rho the sum of weighted items was correlated with the sum of unweighted items to obtain information about the practicability of a weighted total score. The study was carried out in 2003. The internal consistency of the model was not high (alpha = 0.54). Additionally, factor analysis showed that the model had more than one dimension. The correlation between the fall risk model and Care Dependency Scale was quite high for hospital patients and the total group (Spearman Rho = -0.71 or -0.76 respectively, P < 0.01) and medium for nursing home residents (Spearman Rho = -0.51, P < 0.01). The total scores of the weighted and unweighted items correlated highly (Spearman Rho = 0.96, P < 0.01). The use of this risk model is not recommended for nursing homes. For hospitals, we advise the use of unweighted items.