Abstract
BACKGROUND: Mild or pre-frailty is common and associated with increased risks of hospitalisation, functional decline, moves to long-term care, and death. Little is known about the effectiveness of health promotion in reducing these risks. This systematic review aimed to synthesise randomised controlled trials (RCTs) evaluating home and community-based health promotion interventions for older people with mild/pre-frailty. METHODS: We searched 20 bibliographic databases and 3 trials registers (January 1990 - May 2016) using mild/pre-frailty and associated terms. We included randomised controlled and crossover trials of health promotion interventions for community-dwelling older people (65+ years) with mild/pre-frailty and excluded studies focussing on populations in hospital, long term care facilities or with a specific condition. Risk of bias was assessed by two reviewers using the Cochrane Risk of Bias tool. We pooled study results using standardised mean differences (SMD) where possible and used narrative synthesis where insufficient outcome data were available. RESULTS: We included 10 articles reporting on seven trials (total n = 506 participants) and included five trials in a meta-analysis. Studies were predominantly small, of limited quality and six studies tested group exercise alone. One study additionally investigated a nutrition and exercise intervention and one evaluated telemonitoring. Interventions of exercise in groups showed mixed effects on functioning (no effects on self-reported functioning SMD 0.19 (95% CI -0.57 to 0.95) n = 3 studies; positive effects on performance-based functioning SMD 0.37 (95% CI 0.07 to 0.68) n = 3 studies). No studies assessed moves to long-term care or hospitalisations. CONCLUSIONS: Currently the evidence base is of insufficient size, quality and breadth to recommend specific health promotion interventions for older people with mild or pre- frailty. High quality studies of rigorously developed interventions are needed. PROSPERO REGISTRATION: CRD42014010370 (Review 2).