An algorithm to screen long-term care residents at risk for accidental falls

Background and aims: The process applied to identify fall risks in frail elderly persons remains a matter of debate. We intended to develop a fall screening instrument for clinically defined subgroups of long-term care residents, to be administered by nursing staff. Methods: Fall risk indicators were selected by multiple logistic regression in three pre-defined subgroups. The first consisted of residents who were not able to transfer, defined as a change from sit-to-stand position, without physical assistance (NAT). The second subgroup comprised residents who were able to transfer, but who had had a recent fall during the last 6 months (AT-F). Residents who were able to transfer but had had no recent fall (AT-NF) were in the third subgroup. The prospective observational study included 472 long-stay residents (mean age 84 years, 79% female) from three community nursing homes, with a follow-up period of 12 months. Results: Fall incidence was highest in the AT-F subgroup: 6066 per 1000 resident years. The risk indicators identified included a positive fall history and restraint use in the NAT group, transfer assistance in the AT-F group, and urinary incontinence and visual impairment in the AT-NF group. Conclusions: The identification of different risk indicators in the subgroups indicates that specific strategies may be more appropriate to improve the effectiveness of fall prevention in long-term care, than the application of one strategy to all residents. The identification of incontinence, visual impairment, and restraints as risk indicators stresses the need for intervention studies which specifically address these items.