Interventions for preventing falls in the elderly.

BACKGROUND Fractures in the elderly often result from a simple fall. OBJECTIVES To assess the effects of programmes designed to reduce the incidence of falls in community dwelling, institutionalised, or hospitalised elderly people. SEARCH STRATEGY We searched MEDLINE, EMBASE, CINAHL, PsycLIT, Social Science Citation Index, Dissertation Abstracts, Index to UK Theses, the Cochrane Register of Controlled Trials, and bibliographies of identified studies. We contacted known workers in the field. Trials were also obtained from the Cochrane Musculoskeletal Injuries Group trials register. Date of the most recent search: May 1997. SELECTION CRITERIA Randomised trials of interventions designed to minimise the effect of, or prevent exposure to, any putative risk factor for falling in elderly individuals living in the community, in institutional care, or in hospital. The main outcomes of interest were number of fallers or falls, or the number sustaining a fall resulting in injury. Trials that focused on intermediate outcomes such as improved balance or did not report fall outcomes, were excluded. DATA COLLECTION AND ANALYSIS Two reviewers selected trials for inclusion. For each included trial, quality assessment and data extraction was carried out independently by two reviewers. Results of trials of similar design were pooled. MAIN RESULTS Eighteen trials and one pre-planned meta-analysis were included. The analysis of four trials which studied the effect of exercise alone did not establish protection against falling (Peto odds ratio 1.05; 95% confidence interval 0.74 to 1.48). Based on one trial, there was no evidence to support exercise in conjunction with health education classes (Peto odds ratio 1.72; 95% confidence interval 0.78 to 3.75), or of health education classes alone (Peto odds ratio 1.25; 95% confidence interval 0.51 to 3.03) for the prevention of falls. However, significant protection against falling was apparent from interventions which targeted multiple, identified, risk factors in individual patients (Peto odds ratio 0.77; 95% confidence interval 0. 64 to 0.91), and from interventions which focused on behavioural interventions targeting environmental hazards plus other risk factors (Peto odds ratio 0.81; 95% confidence interval 0.71 to 0.93). REVIEWER'S CONCLUSIONS Health care purchasers and providers contemplating fall prevention programmes should consider health screening of at risk elderly people, followed by interventions which are targeted at both intrinsic and environmental risk factors of individual patients. There is inadequate evidence for the effectiveness of single interventions such as exercise alone or health education classes for the prevention of falls.