The rise (of exercise) and fall (of injuries)

1 March 2014

Do community exercise programmes for the elderly reduce injuries as well as falls?


Organised exercise prevents injury from falling by older adults living in the community. The rate of all injury falls may be reduced by about 40 per cent and fracture falls by up to 60 per cent. Hospitals should consider investing in organised exercise for older people as a means of harm reduction and demand management.


Patient falls in hospital are currently a national focus, but amount to a small part of the burden of injury from falls. Residential care has also been focusing on client falls in your region, but attention has not been given to the opportunity to prevent admission of independent-living older people for fall-related injuries. You know exercise prevents falls in older people living in the community, but you also want to know what effect it has on injury levels, which may be more persuasive for your district health board.


Among community-dwelling older people, do exercise programmes reduce injuries from falls?


PubMed – clinical queries (therapy, broad): falls prevention AND exercise AND community.


El-Khoury F, Cassou B, Charles MA, Dargent-Molina P. The effect of fall prevention exercise programmes on fall-induced injury in community- dwelling older adults: a systematic review and meta-analysis of randomised controlled trials. BMJ 2013;347:f6234.


A systematic review of randomised controlled trials where exercise for falls prevention was compared to either ‘no intervention’ or a ‘placebo control’, such as social visits. Inclusion criteria were:

Type of study was randomised controlled trials published in English or French that provided data on injurious falls, serious falls, fall-related injuries or fall-induced fractures;

Patients were people aged 60 years or older, living in the community, but excluding participants that had specific neurodegenerative disease or other condition not corrected by exercise;

Intervention was exercise for falls prevention, but excluding trials that had exercise as part of a multifactorial intervention;

Outcomes were rate ratio of injurious falls (falls and falls injuries were classified from trial reports based on the ProFANE definitions).


The search strategy was specified and the electronic databases searched included the Cochrane Library, PubMed, Embase and CINAHL to July 2013. Trials were restricted to those published in English or French. Two reviewers screened study titles, abstracts and full texts of papers for inclusion and independently-assessed full texts to determine eligibility for inclusion, with additional information sought from authors where necessary. The two reviewers independently assessed the risks of bias and extracted the data. Risk of bias assessment followed that specified by the Cochrane Collaboration and included random sequence generation (selection bias) randomisation, allocation concealment (selection bias), blinding (detection bias), and incomplete outcome data (attrition bias). Publication bias was assessed using funnel plots. Overall – this was a high quality review, perhaps only marred by the language restriction.


Seventeen trials involving 4305 participants, with a mean overall age of 76.7 years. Seven trials selected participants on the basis of being high risk (history of falls, age over 80, or functional limitations). Fourteen trials delivered the intervention in a group setting, and six of these supplemented group sessions with home-based exercise. Study size varied with sample sizes of 53 to 486. Most trials focused on balance, gait and strength training with two being Tai Chai only. Twelve trials were judged as at low risk of bias on random sequence generation, eight on allocation concealment, nine on blinding of falls and injurious falls assessment, eleven on incomplete outcome data, and six on ascertaining serious harm injuries. The six trials confirmed serious injuries using medical records. Intervention duration ranged from 5.5 weeks to one year and from six to 30 months. Four trials had a ‘no intervention’ control. Ten trials provided data that could be included in a meta-analysis. The funnel plot constructed from the ten trials included in the meta-analysis was symmetrical, suggesting publication bias was probably not present. Only minor adverse effects were reported in two trials and were restricted to musculoskeletal discomfort after exercise. Six trials stated there were no adverse effects, while others did not report on ill effects.

See table below.


This review strengthens the case for investment in exercise programmes for community-dwelling older adults. The programmes not only prevent falls but prevent harm from falling.

Sensitivity analyses removing lower quality trials had little effect on rate ratios.

Reviewer: Dr Andrew Jull, RN PhD, Associate Professor, University of Auckland & Nurse Advisor – Quality & Safety, Auckland District Health Board.