RESUMO
BACKGROUND: exercise therapy is highly recommended for falls prevention in older adults; however, poor exercise adherence may limit treatment effectiveness. OBJECTIVE: to assess the effectiveness of interventions to improve exercise adherence for community-dwelling adults (aged over 65 years), at risk of falling. METHODS: eight databases were searched to identify randomised/quasi-randomised trials. The Capability, Opportunity, Motivation model of behaviour (COM-B) was used to categorise the identified adherence interventions. Studies with similar interventions that provided adherence outcome data per group were analysed to establish pooled intervention effect. Protocol registration with Propsero: (CRD42016033677). RESULTS: of the 20 trials included (n = 4419), five provided data per group for adherence outcome. Meta-analysis of four studies (n = 482), containing interventions exploring the way exercise is delivered, demonstrated significantly better adherence in the intervention group (n = 166 experimental, n = 161 control Fixed effects model (FEM), SMD = 0.48 95% CI [0.26-0.70] P < 0.0001 I2 = 0%, very low GRADE evidence). Within this limited evidence base, interventions using telecommunication and the integration of exercise into activities of daily living appear most promising when delivering exercise at home. Meta-analysis to explore the effect that these interventions to improve adherence had on balance (n = 166 experimental, n = 161 control Random-effects model (REM), SMD = 0.82, 95% CI [-1.20-2.84] P = 0.43 I2 = 52%) and gait (n = 59 experimental, n = 56 control REM, SMD = 0.29, 95% CI [-1.62-2.20] P = 0.77 I2 = 48%), found no statistically significant effect. CONCLUSIONS: adherence to exercise can be positively influenced; however, insufficient data exists to support any single intervention that also achieves effective outcomes for balance and gait.