For the millions of heart failure patients who cannot reliably access clinic-based rehabilitation — due to geography, disability, or healthcare capacity — the question of whether a smartphone or connected device can substitute for supervised exercise programs carries enormous practical weight. A new meta-analysis from JMIR mHealth and uHealth offers the most rigorous aggregated evidence yet on this question, with results that meaningfully shift the conversation.
Drawing on eight randomized controlled trials enrolling 1,368 patients — predominantly those with heart failure with reduced ejection fraction (HFrEF) — the analysis examined structured, exercise-focused remote, virtual, or hybrid cardiac rehabilitation (RVH-CR) delivered via mobile health platforms. The primary outcomes were peak oxygen uptake (VO2 peak) and six-minute walk distance (6MWD), the two most clinically validated markers of cardiorespiratory fitness in this population. When compared with center-based cardiac rehabilitation (CBCR), mHealth-supported RVH-CR demonstrated a statistically significant greater improvement in VO2 peak — a counterintuitive finding that challenges assumptions about supervised in-person superiority. Comparisons against usual care also favored the mHealth intervention across both primary endpoints. Secondary outcomes including health-related quality of life trended positively, and safety signals were not elevated relative to comparators. Evidence certainty was graded using the GRADE framework, providing transparency about confidence levels across outcomes.
This meta-analysis arrives at a moment when telehealth reimbursement policies remain unsettled in many health systems, making efficacy data particularly actionable for clinicians and payers. The finding that digital delivery may actually outperform center-based rehab on VO2 peak — likely reflecting better adherence enabled by convenience — aligns with emerging literature on remote exercise interventions in coronary artery disease. Key limitations include the modest trial count (eight RCTs), the concentration on HFrEF rather than preserved ejection fraction populations, and heterogeneity in the specific mHealth platforms studied. This is nonetheless a confirmatory and clinically meaningful synthesis, not merely incremental — the effect direction is consistent and the comparison set robust enough to inform practice.