Primary care interventions to address physical frailty among community-dwelling adults aged 60 years or older: A meta-analysis

Abstract

Introduction: The best interventions to address frailty among older adults have not yet been fully defined, and the diversity of interventions and outcome measures makes this process challenging. Consequently, there is a lack of guidance for clinicians and researchers regarding which interventions are most likely to help older persons remain robust and independent. This paper uses meta-analysis to assess effectiveness of primary care interventions for physical frailty among community-dwelling adults aged 60+ and provides an up-to-date synthesis of literature in this area.

Methods: PubMed, CINAHL, Cochrane Register of Controlled Trials, and PEDro databases were searched, and RCTs, controlled pilot studies, or trials with similar study designs addressing frailty in the primary care setting among persons aged 60+ were chosen. Study data was abstracted following PRISMA guidelines, then meta-analysis was performed using the random effects model.

Results: 31 studies with a total of 4794 participants were analysed. Interventions using predominantly resistance-based exercise and nutrition supplementation seemed to improve frailty status versus control (RR = 0.62 (CI 0.48-0.79), I2 = 0%). Exercise plus nutrition education also reduced frailty (RR = 0.69 (CI 0.58-0.82), I2 = 0%). Exercise alone seemed effective in reducing frailty (RR = 0.63 (CI 0.47-0.84), I2 = 0%) and improving physical performance (RR = 0.43 (CI 0.18-0.67), I2 = 0%). Exercise alone also appeared superior to control in improving gait speed (SMD = 0.36 (CI 0.10-0.61, I2 = 74%), leg strength (SMD = 0.61 (CI 0.09-1.13), I2 = 87%), and grip strength (Mean Difference = 1.08 (CI 0.02-2.15), I2 = 71%) though a high degree of heterogeneity was observed. Comprehensive geriatric assessment (RR = 0.77 (CI 0.64-0.93), I2 = 0%) also seemed superior to control in reducing frailty.

Conclusion: Exercise alone or with nutrition supplementation or education, and comprehensive geriatric assessment, may reduce physical frailty. Individual-level factors and health systems resource availability will likely determine configuration of future interventions.

Fig 1. Prisma flow diagram showing identification, screening, eligibility assessment, and inclusion of studies.

Reasons for exclusion of articles were: Ineligible outcome (13 studies); ineligible comparator (21), other (not yet published, irrelevant, not an intervention etc.) (34), unable to contact authors (4), authors contacted but primary data no longer accessible (1).

Fig 2. Risk of bias assessment of included studies.

Showing risk of bias as either low (green), unclear (yellow), or high (red) for included studies, for random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, or other bias.

Fig 3. Effect of interventions on frailty status.

A) Shows Risk ratio (RR, square data markers) with 95% confidence intervals (CI, horizontal lines) for meta-analysis of change in frailty status (dichotomised to not frail vs. frail; events indicate individuals whose status did not change) for interventions using exercise plus nutrition supplements or nutrition supplements alone. B) Shows RR with 95% CI for meta-analysis of change in prevalence of frailty (dichotomised to not frail vs. frail; events indicate frail individuals) for interventions using exercise plus nutrition education, exercise only, and comprehensive geriatric assessment. Size of data markers indicates study weighting in random effects meta-analysis. Diamond-shaped data markers indicate overall RR and CI for each grouping of interventions. I² statistic is reported for each group. Left side favours intervention.

Fig 4. Effect of interventions on physical performance and gait speed.

A) Shows SMD (square data markers) with 95% CI (horizontal lines) for meta-analysis of mean scores ± SD for physical performance tests for interventions using exercise. B) Shows SMD with 95% CI for meta-analysis of gait speed as mean time taken to walk a set distance (measured in seconds) or mean speed ± SD (measured in metres per second) for interventions using exercise plus nutrition supplements, exercise only, or nutrition supplements only. C) Shows Mean Difference with 95% CI for meta-analysis of gait speed (measured metres per second), reported either as change scores or as final measurements ± SD, separated into sub-groups for clarity, for interventions using exercise plus nutrition education. Size of data markers indicates study weighting in random effects meta-analysis. Diamond-shaped data markers indicate overall SMD or Mean Difference with CI for each grouping of interventions. I² statistic is reported for each group. Right side favours intervention.

Fig 5. Effect of interventions on leg strength.

A) Shows SMD (square data markers) with 95% CI (horizontal lines) for meta-analysis of mean scores ± SD for leg strength (measured in kilogrammes, kilogramme metres, feet per pound, Newtons, or Newton metres) for interventions using exercise plus nutrition supplements, exercise only, and nutrition supplements only. B) Shows SMD with 95% CI for meta-analysis of leg strength reported as mean change from baseline ± SD (measured in Kilogrammes or Newton Metres), for interventions using hormone supplementation. Size of data markers indicates study weighting in random effects meta-analysis. Diamond-shaped data markers indicate overall SMD with CI for each grouping of interventions. I² statistic is reported for each group. Right side favours intervention.

Fig 6. Effect of interventions on grip strength.

A, B, D) Show Mean Difference (square data markers) with 95% CI (horizontal lines) for meta-analysis of grip strength (measured in kilogrammes), reported either as change scores or as final measurements ± SD, separated into sub-groups for clarity, for interventions using exercise plus nutrition education, exercise only, or hormone supplementation. C) Shows SMD with 95% CI for meta-analysis of mean grip strength (measured in kilogrammes) ± SD, for interventions using nutrition supplements. Size of data markers indicates study weighting in random effects meta-analysis. Diamond-shaped data markers indicate overall SMD or Mean difference with CI for each grouping of interventions. I² statistic is reported for each group. Right side favours intervention.