JMIR Publications

ABSTRACT

Background:

Frailty affects more than 20% of older adults and increases falls, disability, and health care use. Traditional rehabilitation is limited by accessibility barriers. Interactive digital health interventions (DHIs), particularly mobile health (mHealth) programs integrating wearable activity trackers with behavioral coaching, may offer scalable alternatives.

Objective:

This systematic review and meta-analysis examined the effectiveness of interactive DHIs compared with conventional rehabilitation for physical function and physical activity in older adults with frailty or prefrailty, and whether intervention modality and behavioral coaching components influence outcomes.

Methods:

The authors searched MEDLINE, Embase, Cochrane Library, Airiti, CNKI, and WanFang through August 31, 2023, for randomized controlled trials (RCTs) involving adults aged ≥65 years with frailty or prefrailty. Primary outcomes included grip strength, Timed Up and Go (TUG), gait speed, and the Fried Frailty Index; secondary outcomes were moderate-to-vigorous physical activity (MVPA) and step count. Duplicate records were identified and removed automatically using PICO Portal (Evidence Partners). All subsequent screening of titles, abstracts, and full-text articles was performed manually by two independent reviewers without the use of automated eligibility tools; disagreements were resolved by a third reviewer. Two reviewers independently screened studies, extracted data, and assessed risk of bias (RoB 2); certainty of evidence was rated using GRADE. Random-effects meta-analyses were performed with prespecified subgroup and sensitivity analyses.

Results:

Nine RCTs (358 participants) were included. Pooling all trials, DHIs showed no clear superiority over conventional rehabilitation for primary functional outcomes, with low to very low certainty of evidence. DHIs produced a robust, clinically important increase in MVPA of 14.31 minutes per valid day (95% CI 12.45–16.16; 2 trials; I2 = 0%). Sensitivity analyses excluding high risk-of-bias trials revealed clinically meaningful improvements in gait speed (mean difference [MD] 0.11 m/s), Fried Frailty Index (MD −1.00 points), and daily step count (approximately 480 additional steps). Modality-based subgroup analyses showed that mHealth programs incorporating behavioral coaching achieved superior improvements in grip strength (MD 2.90 kgf) and TUG (MD −6.57 seconds, exceeding the minimal clinically important difference) compared with exergame-based interventions. Findings for the Fried Frailty Index, MVPA, and step count should be interpreted with caution, as each was informed by only two trials (73 participants total).

Conclusions:

DHIs may increase MVPA and may improve selected functional outcomes, including gait speed and frailty status, when implemented with sufficient fidelity, but the certainty of evidence remains low to very low. mHealth interventions integrating wearable activity tracking and behavioral coaching appear more promising than stand- alone exergames, although confirmatory trials are needed before firm clinical prioritization. Future trials should evaluate long-term outcomes, economic impacts, and implementation fidelity. Clinical Trial: PROSPERO registration number CRD42023462594; https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42023462594