Thirty-four studies were included in the review. Participant numbers were reported for 25 studies (141,273 participants, range 11 to 79,471). The studies included eight RCTs (8,720 participants, range 54 to 3,999); two quasi-RCTs (135 and 825 participants); two prospective cohort studies (1,609 participants for one study); one retrospective cohort; one case-control study (70 participants); 15 before-after studies (129,914 participants, range 328 to 79,471 for eight studies); and five experimental studies (112 participants, range 11 to 45).
Eleven studies were adequately randomised. Two studies had appropriate allocation concealment. Participants were blinded to allocation in two studies. Attempts were made to blind outcome measurement in three studies. Adjustment for loss to follow-up was made in nine studies. Twenty-four studies had sufficient power.
Multifaceted fall prevention interventions (14 studies): The interventions in 12 of 14 studies gave sizable reductions in falls and/or fall-related injuries; results were significant in six studies.
Single environment-related interventions: Statistically significant reductions in falls were found with vinyl flooring versus carpeted flooring (one study) and bed rail reduction (one study).
Single care process and culture-related interventions: Individual studies found a significant reduction in falls with a medication review and modification, and patient education plus a multifaceted falls programme versus the programme alone. Another study found a significant fall reduction and significantly improved musculoskeletal function in elderly women with vitamin D and calcium supplementation versus no supplementation, which was not significant for supplementation with calcium alone. One study found a significant falls reduction with the addition of an exercise programme to a multifaceted fall prevention programme compared with the fall prevention programme alone. One study found no significant effect for additional exercise versus physiotherapy alone. Two studies found significant reductions in falls when volunteer companions were present. The use of blue identification bracelets to identify at risk patients versus no bracelets did not significantly reduce falls.
Single technology related interventions: No significant reductions in falls were found with the use of bed alarm systems (two studies).
Environment-related research
(nine studies):
There were significantly more falls in a 40-bed longitudinal ward versus either a 40-bed nuclear ward or a 28-bed nuclear ward (one study). There were significantly fewer falls in acuity-adaptable single-bed rooms with decentralised nurse stations versus two-bed rooms in coronary intensive care units (one study). More falls occurred with softer floors (three studies), but softer floors reduced the severity of injuries (four studies).
Multi-systemic fall prevention models: Two models were described which the authors reported had potential for reducing falls.