Randomised controlled trials (RCTs) and quasi-RCTs of exercise interventions for older people that reported falls (number of falls or fallers, fall rate), measured participation and were conducted in a home, aged care facility or hospital setting were eligible for inclusion in the review. Participants were required to be aged over 60 years or be clearly described as elderly or senior; otherwise it was required that the mean age of the study population minus one standard deviation (SD) exceeded 60 years. The study intervention could comprise exercise alone or as part of a multi-faceted intervention. Controls could include placebo, alternative therapy or usual care. Studies were required to report participation at baseline and follow-up and to use self- or interview-administered generic instruments, designed for adults, that measured the International Classification of Functioning, Disability and Health (ICF) concept of participation using published methods (detailed in the review). Studies that referred participants to an exercise programme outside the study were excluded.
The mean age of participants in the included studies ranged from 75 to 85 years. Studies were set in the community or in aged care facilities and in several studies participants were recently discharged from hospital. More than half of the studies had fall history or risk factors for falls as an inclusion criterion. Interventions varied (such as endurance, strength, gait, balance and functional exercise, Tai Chi and/or physiotherapy) and were delivered in supervised groups, individually in classes or at home. Most studies delivered exercise as a single intervention; some included other interventions (such as goal-setting, cognitive behavioural therapy and home safety assessment). Intervention duration ranged from 1.5 to 12 months and frequency ranged from daily to three or more times over three months. Controls received alternative therapy, usual care or placebo. Participation was measured at 1.5 to 14 months after randomisation using one or more of nine different instruments.
One reviewer initially screened titles and abstracts retrieved by the search. Two reviewers performed the final selection of studies, with differences settled by discussion.