Abstract
Background
Digital health (DH) enables home-based cardiac rehabilitation (HBCR) for heart failure, but no prior systematic review has compared wearable-guided vs. interaction-guided DH modalities’ efficacy on cardiorespiratory fitness, exercise capacity, and quality of life.
Methods
We systematically searched PubMed, Embase, Cochrane Library, and CNKI until April 2025 for RCTs. Two investigators independently extracted data, assessed the risk of bias using the Cochrane Risk of Bias Tool 2.0, and graded evidence (GRADE).
Results
Analysis of 20 RCTs (n = 4,652) showed DH-HBCR significantly improved 6-minute walk test (6MWT: MD = 21.92 m, 95% CI 12.27–31.58), peak oxygen uptake (VO₂peak: MD = 0.96 mL/kg/min, 95% CI 0.56–1.35), and quality of life (MLHFQ: MD = -7.21, 95% CI -11.40–-3.02) versus control. Wearable-guided interventions provided superior gains in 6MWT (MD = 51.98 m, 95% CI 34.21–69.76; I² = 0%) and MLHFQ (MD = -11.14, 95% CI -14.55–-7.72; I² = 0%), while interaction-guided interventions significantly improved VO₂peak (MD = 1.00 mL/kg/min, 95% CI 0.49–1.51; I² = 15.2%) more effectively.
Conclusions
DH-mediated HBCR was associated with improved exercise capacity, cardiorespiratory fitness, and quality of life in patients with HF. Preliminary evidence suggests differential efficacy profiles between wearable-guided and interaction-guided modalities. Given the limitations of the included evidence, these findings should be considered hypothesis-generating. Future research should develop phenotype-stratified protocols and conduct head-to-head RCTs to validate these observations.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12872-026-06160-6.
Keywords: Digital health, Heart failure, Home-based cardiac rehabilitation, Systematic review, Meta-analysis
Introduction
Heart failure (HF) constitutes a clinical syndrome severely impairing quality of life with high rates of hospitalization and mortality [1–3]. As a cornerstone of comprehensive HF management, cardiac rehabilitation (CR) has demonstrated significant improvements in exercise capacity and quality of life [2, 4]. However, traditional center-based CR (CBCR) faces structural barriers including low participation rates [5] and limited healthcare accessibility [6]. Home-based cardiac rehabilitation (HBCR), as an alternative strategy to overcome geographic and resource constraints of CBCR, has proven efficacy in improving exercise capacity, cardiorespiratory fitness, and quality of life [7]. Nevertheless, conventional HBCR is persistently limited by the absence of real-time supervision, deficits in the standardization of rehabilitation protocols, and an inability to dynamically adjust regimens based on patients’ evolving clinical status [8].
Digital health (DH) technologies offer effective solutions to limitations in traditional home-based cardiac rehabilitation (HBCR) [9, 10]. Contemporary DH tools exhibit functional diversification through various modalities including wearable devices, mobile applications, and remote platforms [7]. Based on implementation mechanisms, DH-enhanced HBCR operates through two core approaches: wearable-guided interventions employ passive sensing technologies including smartwatches and implantable devices for continuous physiological monitoring and automated biofeedback [11], while interaction-guided interventions utilize communication platforms such as mobile apps and video systems to deliver dynamic guidance and behavioral support through clinician-patient dyads [12, 13]. These approaches are structurally distinguished by their technical emphasis: passive tracking of objective biometric data versus accessible professional engagement through remote communication platforms.
Prior meta-analyses investigating DH in cardiac rehabilitation report varied scopes: Jones et al. [14] (57 studies) identified that wearable-assisted HBCR improved quality of life and peak oxygen uptake (VO₂peak) in coronary artery disease and HF patients, demonstrating its potential as a CBCR alternative; Li et al. [5] (n = 1,508) confirmed mHealth-HBCR significantly enhanced 6-minute walk test (6MWT) and VO₂peak compared with usual care, while matching CBCR efficacy; Gao et al. [8] reported superior efficacy of telerehabilitation over usual care and comparable outcomes to CBCR for HF; Tang et al. [15] further validated significant improvements in 6MWT and VO₂peak through non-invasive telerehabilitation. Although existing studies affirm the effectiveness of DH-supported HBCR, no meta-analysis has systematically categorized and compared wearable-guided versus interaction-guided interventions as distinct therapeutic paradigms to evaluate their differential effects on rehabilitation outcomes. Given the escalating diversity of DH applications in HBCR practice, understanding potential disparities in the impact of different DH modalities on primary rehabilitation outcomes is crucial for optimizing resource allocation, guiding clinicians toward personalized technology strategies, and advancing precision-based rehabilitation.
This systematic review and meta-analysis compare wearable-guided versus interaction-guided DH interventions for differential effects on cardiorespiratory fitness (VO₂peak), exercise capacity (6MWT), and quality of life (Minnesota Living with Heart Failure Questionnaire, MLHFQ) in HF patients, providing evidence-based guidance for developing individualized HBCR programs.
Methods
Search strategy
This study adheres to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) statement [16] and has been registered with PROSPERO (CRD420251035626). We conducted a comprehensive search across PubMed, Embase, Cochrane Library, and CNKI using a hybrid strategy that combined Medical Subject Headings (MeSH) with free-text terms, linked by Boolean operators. Eligible studies were defined as randomized controlled trials (RCTs) published in any language from database inception through April 12, 2025. The search targeted HF patients, identified using the core MeSH term “Heart Failure” enhanced by free-text terms to capture diverse HF subtypes. Interventions included: (1) Cardiac rehabilitation, using “Cardiac Rehabilitation” as the pivotal term combined with relevant free-text descriptors; (2) Exercise-based interventions, centered on “Exercise” combined with modality-specific free terms; and (3) DH strategies, operationalized via MeSH terms (e.g., “Digital Health”) and free-text terms (e.g., “mHealth”, “telemedicine”) to encompass remote care models. Manual screening of reference lists from included studies was conducted to identify additional relevant publications. The complete search strategies for each database are detailed in Supplementary Table S1.
Inclusion criteria
The inclusion criteria were: (1) RCTs; (2) adults ≥ 18 years with clinically stable chronic HF; (3) DH-mediated HBCR interventions, encompassing either pure home-based or hybrid (center-based initiation followed by home-based) programs, versus control groups receiving usual care or CBCR; and (4) Reported at least one primary endpoint (6MWT, VO₂peak, or MLHFQ).
Exclusion criteria
Exclusion criteria were: (1) Non-randomized studies, reviews, conference abstracts, unpublished protocols, or studies with incomplete data preventing extraction of core outcomes; (2) Studies including patients aged < 18 years or with acute decompensated HF; (3) Interventions without DH components (e.g., HBCR without remote monitoring) or using concurrent hybrid rehabilitation (non-sequential combination of center-based and home-based programs); and (4) Failure to report outcomes for at least one endpoint (6MWT, VO₂peak, or MLHFQ).
Data extraction
Data extraction was independently performed by two investigators (ZZX, LJM) using a standardized Excel template, with disagreements resolved through consensus discussion or third-party adjudication (GY) when required. The extraction covered four categories: (a) basic study characteristics: first author, publication year, country, study design, and intervention duration; (b) participant attributes: total sample size, age, male proportion, HF phenotype (HFrEF/HFpEF), NYHA functional class, and left ventricular ejection fraction; (c) intervention specifications: exercise prescription parameters (frequency, intensity, type) and DH modalities used in both intervention and control groups (usual care or CBCR); and (d) Outcome metrics: We extracted baseline and post-intervention means, standard deviations, and sample sizes for three prespecified primary endpoints measured immediately post-intervention (8–24 weeks): 6MWT (6-minute walking distance per ATS guidelines, reflecting submaximal exercise capacity); VO₂peak (peak oxygen consumption during cardiopulmonary exercise testing, gold standard for cardiorespiratory fitness); and MLHFQ (validated 21-item HF-specific quality-of-life scale, 0–105, lower scores indicate better quality of life).
Digital health modality classification
Interventions were categorized into two mutually exclusive groups based on core technical pathway, dominant functional component, and primary implementation mechanism, with mixed-technology interventions classified by their dominant delivery component. The wearable-guided group centered on passive wearable sensors (e.g., smartwatches, heart rate monitors, portable ECG recorders, Polar series, Fitbit®) for continuous physiological monitoring, automated biofeedback, and real-time safety alerts, with clinician-patient interaction limited to occasional follow-up. The interaction-guided group relied on real-time or near-real-time clinician-patient communication platforms (e.g., video conferencing, telephone, mobile apps, WeChat/QQ) for dynamic exercise prescription adjustments, personalized behavioral support, and regular supervision, where wearables served only as auxiliary data collection tools.
Risk of bias and certainty assessment
Methodological quality was assessed using the Cochrane Risk of Bias Tool 2.0 (RoB 2.0) [17] across five structured domains: (1) Randomization process (2), Deviations from intended interventions (3), Missing outcome data (4), Measurement of the outcome, and (5) Selection of the reported result. Each domain was rated as “low risk”, “high risk”, or “some concerns” according to RoB 2.0 guidelines. Evidence certainty was evaluated using the GRADE framework (Grading: high/moderate/low/very low). Two reviewers (ZZX, LJM) independently performed both assessments using standardized RoB 2.0 templates. Disagreements were resolved through consensus or arbitration by a third reviewer (GY).
Data synthesis and analysis
Statistical analyses were performed in Stata 18.0 following Cochrane methodology. Continuous outcomes (6MWT, VO₂peak, MLHFQ) were pooled as mean differences (MD) with 95% confidence intervals. Pooled estimates for continuous outcomes were based on change-from-baseline values. As all studies used identical measurement scales, MD was selected to retain direct clinical interpretability for clinicians. Heterogeneity was assessed using three core statistics: Cochran’s Q test (with degrees of freedom and P-value) for the statistical significance of between-study variability; the I² statistic (≤ 50%: low-moderate; 50%–75%: moderate; >75%: high) for the proportion of variability attributable to between-study differences; and τ² (DerSimonian-Laird estimate) for the magnitude of true between-study variance. Given the anticipated clinical and methodological diversity across studies, random-effects models using the DerSimonian–Laird method were prespecified as the primary analytic approach for all meta-analyses, irrespective of the observed I² value. Fixed-effect models using the inverse-variance method were conducted as sensitivity analyses to assess the robustness of the pooled estimates. Missing standard deviations were imputed according to the Cochrane Handbook guidelines. Prespecified subgroup analyses were conducted by DH modality, control type, HF phenotype, and publication year, with between-subgroup differences tested via χ² tests. Sensitivity analyses assessed robustness by sequential study exclusion, and publication bias was evaluated using funnel plots and Egger’s test (P < 0.05 considered significant).
Results
Study selection
The systematic search retrieved 1,025 records. After duplicate removal (n = 362), 575 irrelevant studies were excluded during title and abstract screening of the remaining 663 articles. Full-text review of the resulting 88 publications led to the exclusion of 68 studies for the following reasons: non-HBCR programs; non-RCT designs; non-HF patient cohorts; unavailable primary outcomes; absence of integrated DH components; or duplicate reporting. Finally, 20 RCTs fulfilled all inclusion criteria. The selection process, compliant with PRISMA guidelines, is detailed in Fig. 1.
Fig. 1.
The PRISMA flow chart of literature screening and selection process
Characteristics of the studies
Characteristics and distribution of included studies and participants
The geographic distribution of included studies spanned 11 countries, predominantly in developed nations such as the United States, Poland, Canada, and France. China was the only represented middle-income country (n = 2), with one multinational trial. Baseline characteristics comprised 4,652 heart failure patients, including 2,364 in intervention groups and 2,288 controls (2,127 receiving usual care; 161 undergoing CBCR). Patient ages ranged from 54.3 to 77.6 years. Most trials reported male predominance (approximately 14 studies with more than 50% male representation). HF subtypes were predominantly HFrEF or unclassified phenotypes; most patients were NYHA class II-III. Complete study characteristics are detailed in Table 1.
Table 1.
General characteristics of the study and patients
| ID | Authors | Year | Country | Types of HF | Sample (n) | Age | Male (n, %) | NYHA Classification | EF (%) | Outcomes | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| INT | CON | INT | CON | INT | CON | INT | CON | INT | CON | ||||||
| 1 | Rebecca A. Gary [18] | 2004 | America | HFpEF | 16 | 16 | 67.0 ± 11.0 | 69.0 ± 11.0 | 0 | 0 | II: 6 (38.0%); III: 10 (62.5%) | II:7 (44.0%); III: 9(56%) | 53.6 ± 7.3 | 56.8 ± 9.5 | 6MWT/MLHFQ |
| 2 | Rebecca Gary [19] | 2007 | America | HFpEF | 13 | 10 | 68.0 ± 11.0 | 0 | 0 |
II: 10 (43.0%); III: 13 (57.0%) |
56.0 ± 7.0 | 6MWT/MLHFQ | |||
| 3 | O’Connor, C. M. [20] | 2009 | America, Canada, France | HFrEF | 1159 | 1172 | 59.2(51.2–67.8) | 59.3(51.1–68.2) | 812 (70.1) | 858 (73.2) | II: 723 (62.4%); III: 422 (36.4%); IV: 14 (1.2%) |
II: 754 (64.3%); III: 409 (34.9%); IV: 9 (0.8%) |
24.6 (20.0–30.0) | 24.9 (20.0-30.2) | 6MWT/VO2 peak |
| 4 | Ewa Piotrowicz [27] | 2010 | Poland | HFrEF | 77 | 75 | 56.4 ± 10.9 | 60.5 ± 8.8 | 64 (85.0) | 53 (95.0) | II: 37 (49.3%); III: 38 (50.7%) | II: 31 (55.4%) III: 25 (44.6%) | 30.2 ± 8.2 | 30.8 ± 6.7 | 6MWT/VO2peak/MLHFQ |
| 5 | Rebecca A. Gary [21] | 2011 | America | HFrEF | 12 | 12 | 59.0 ± 11.0 | 61.0 ± 10.0 | 7 (58.0) | 5 (42.0) | II: 4 (33.0%); III: 8 (67.0%) | II: 6 (50.0%); III: 6 (50.0%) | 23.0 ± 8.0 | 27.0 ± 9.0 | 6MWT/MLHFQ |
| 6 | Ewa Piotrowicz [27] | 2015 | Poland | HFrEF | 77 | 34 | 54.4 ± 10.9 | 62.1 ± 12.5 | 64 (85.0) | 31 (97.0) | II: 51 (68.0%); III: 24 (32.0%) | II: 23 (72.0%); III: 9 (28.0%) | 30.0 ± 8.0 | 34.0 ± 6.0 | 6MWT/VO2peak/MLHFQ |
| 7 | E. Smolis-Bąk, R. [29] | 2015 | Poland | HFrEF | 26 | 26 | 60.0 ± 8.5 | 65.1 ± 8.2 | NR | NR | III: 26 (100%) | III: 26 (100%) | 25.3 ± 7.4 | 24.9 ± 7.2 | VO2 peak |
| 8 | Ewa Piotrowicz [30] | 2016 | Poland | HFrEF | 46 | 23 | 54.3 ± 10.4 | 60.4 ± 11.9 | 39 (84.8) | 22 (95.7) | II: 6 (78.3%); III: 10 (21.7%) | II: 16 (69.6%); III: 7 (30.4%) | 31.0 ± 7.0 | 33.0 ± 7.0 | VO2 peak |
| 9 | Safiyari-Hafizi H [23] | 2016 | Canada | HF Phenotypes | 20 | 20 | 62.0 ± 11.0 | 62.0 ± 11.0 | 14 (70.0) | 14 (70.0) | I: 8 (40.0%); II: 10 (50.0%); III: 2 (10.0%) | I: 8 (40.0%); II: 10 (50.0%); III: 2 (10.0%) | NR | NR | VO2 peak |
| 10 | Roberto Antonicelli [22] | 2016 | Italy | HF Phenotypes | 170 | 173 | 76.2 ± 5.2 | 77.6 ± 6.0 | 103 (60.6) | 92 (53.2) | NR | NR | 47.9 ± 13.3 | 49.0 ± 13.4 | 6MWT/MLHFQ |
| 11 | Rita Hwang [31] | 2017 | Australia | HF Phenotypes | 24 | 29 | 68.0(14.0) | 67.0(11.0) | 19 (79.0) | 21 (72.0) | I: 3 (13.0%); II: 9 (37.0%); III: 12 (50.0%) | I: 2 (7.0%); II: 21 (72.0%); III: 6 (21.0%) | 36.0 (16.0) | 35.0 (17.0) | 6MWT/MLHFQ |
| 12 | Bernocchi P [24] | 2017 | Italy | HF Phenotypes | 56 | 56 | 71.0 ± 9.0 | 70.0 ± 9.5 | 50 (88.0) | 42 (75.0) | II: 25 (45.0%); III: 22 (39.0%); IV: 9 (16.0%) | II: 29 (52.0%); III: 19 (34.0%); IV: 8 (14.0%) | 44.5 ± 12.4 | 43.3 ± 13.2 | 6MWT/MLHFQ |
| 13 | Peng, X. [32] | 2018 | China | HF Phenotypes | 49 | 49 | 66.3 ± 10.5 | 28 (57.1) | 30 (61.2) | I: 11 (22.4%); II: 18 (36.7%); III: 20 (40.8%) | I: 13 (26.5%); II: 18 (36.7%); III: 18 (36.7%) | 49.4 ± 12.3 | 48.8 ± 12.2 | 6MWT | |
| 14 | Ewa Piotrowicz [33] | 2020 | Poland | HFrEF | 425 | 425 | 62.6 ± 10.8 | 62.2 ± 10.2 | 377 (88.7) | 376 (88.5) | I: 54 (12.7%); II: 293 (68.9%); III: 78 (18.4%) | I: 50 (11.8%); II: 284 (66.8%); III: 91 (21.4%) | 31.0 ± 7.0 | 30.0 ± 7.0 | 6MWT/VO2 peak/SF-36 |
| 15 | Andrade GN [25] | 2021 | Brazil | HFrEF | 11 | 12 | 59.0 ± 5.0 | 61.0 ± 7.0 | 5 (46.0) | 9 (75.0) | II: 10 (91.0%) | II: 11 (92%) | 31.0 ± 6.0 | 32.0 ± 6.0 | 6MWT/VO2peak/MLHFQ |
| 16 | PAN Haiyan [35] | 2022 | China | HFrEF | 41 | 47 | 60.7 ± 11.7 | 59.9 ± 11.9 | 27 (65.9) | 31 (65.9) | II: 17 (41.5%); III: 24 (58.5%) | II: 19 (40.4%); III: 28 (59.6%) | 32.0 ± 5.4 | 32.7 ± 4.2 | 6MWT/VO2 peak/SF-36 |
| 17 | Nagatomi Y [26] | 2022 | Japan | HF Phenotypes | 15 | 15 | 59.8 ± 10.0 | 67.7 ± 8.9 | 9 (60.0) | 7 (47.0) | II: 10 (66.7%); III: 5 (33.3%) | II: 9 (60.0%); III: 6 (40.0%) | 39.9 ± 17.8 | 44.5 ± 17.3 | 6MWT |
| 18 | Rebecca A Gary [34] | 2022 | America | HF Phenotypes | 21 | 19 | 63.0 ± 9.0 | 60.0 ± 10.0 | 13 (62.0) | 7 (37.0) | II: 11 (52.0%); III: 10 (48.0%) | II: 13 (68.0%); III: 6 (32.0%) | 34.0 ± 19.0 | 36.0 ± 14.0 | 6MWT/VO2peak/KCCQ |
| 19 | Lundgren KM [36] | 2023 | Norway | HF Phenotypes | 31 | 30 | 67.6 ± 10.9 | 67.7 ± 11.9 | 8 (25.8) | 3 (10.0) | II: 24 (77.4%); III: 7 (22.6%) | II: 26 (86.7%); III: 4 (13.3%) | 36.4 ± 12.1 | 32.4 ± 11.0 | 6MWT/VO2 peak |
| 20 | Cristine Schmidt, [37] | 2025 | Portugal | HF Phenotypes | 75 | 45 | 62.0 ± 11.0 | 79 (66.0) |
I: 20 (26.7%); II: 50(66.7%); III: 5 (6.7%) |
I: 6 (13.3%); II: 37 (82.2%); III: 2 (4.4%) | 36.0 ± 12.3 | 36.8 ± 9.6 | 6MWT/VO2peak/MLHFQ | ||
Abbreviations: INT Intervention, CON Control, HFpEF Heart Failure with Preserved Ejection Fraction, 6MWT 6-Minute Walk Test, MLHFQ Minnesota Living with Heart Failure Questionnaire, HFrEF Heart Failure with Reduced Ejection Fraction, VO2peak Peak Oxygen Uptake, HF Heart Failure, HF Heart Failure, NR Not Reported
Characteristics of interventions
All home-based cardiac rehabilitation (HBCR) interventions were delivered through digital health (DH) platforms, comprising: 9 studies [18–26] implementing wearable-guided interventions and 11 studies [27–37] utilizing interaction-guided interventions. The primary exercise modality was aerobic training (walking, cycling, treadmill), supplemented by resistance training (resistance bands, weighted exercises, dynamometer use). Selected studies integrated respiratory and functional training. Exercise frequency ranged from 2 to 7 sessions/week, with program durations of 8–24 weeks in most trials. Control groups received either usual care (UC) without structured exercise (n = 16) or CBCR with supervised exercise training (n = 4). Comprehensive intervention characteristics are detailed in Supplementary Table S2.
Risk of bias
Cochrane RoB 2.0 assessment of the 20 included studies indicated: randomization process was low risk in 7 studies, some concerns in 6, and high risk in 7; deviations from intended interventions showed some concerns in 16 studies and high risk in 4; missing outcome data presented low risk in 9 studies, some concerns in 8, and high risk in 3; measurement of the outcome had low risk in 10 studies, some concerns in 3, and high risk in 7; selection of the reported result was low risk in 9 studies and some concerns in 11 (Supplementary Figure S1–2). GRADE evaluation demonstrated very low-certainty evidence for 6MWT due to critical bias risks, substantial inconsistency, and significant publication bias. Conversely, evidence certainty was moderate for VO₂peak and MLHFQ outcomes despite identified biases, given low inconsistency and negligible publication bias.
Meta-analysis results
Effect of HBCR on 6MWT
Seventeen studies (n = 4,394) demonstrated that HBCR significantly improved 6MWT performance compared to controls (MD = 21.92 m, 95% CI 12.27–31.58; Cochran’s Q = 63.6, I² = 74.8%, τ² = 149.0). Subgroup analyses revealed: (1) By DH modality, wearable-guided interventions (8 studies, n = 2,884) showed greater improvement (MD = 51.98 m, 95% CI 34.21–69.76; I² = 0%), whereas interaction-guided interventions (9 studies, n = 1,510) had modest benefits despite high heterogeneity (MD = 12.45 m, 95% CI 3.52–21.37; I² = 76.4%); (2) By control type, HBCR outperformed usual care (13 studies, n = 4,067; MD = 28.31 m, 95% CI 17.27–39.35; I² = 80.1%) but showed comparable efficacy to CBCR (4 studies, n = 327; MD = -3.64 m, 95% CI -17.41–10.13; I² = 0%); (3) By HF phenotype, HBCR significantly improved outcomes in HFrEF (7 studies, n = 3,554; MD = 13.86 m, 95% CI 1.37–26.36; I² = 12.3%) and mixed phenotypes (8 studies, n = 789; MD = 26.78 m, 95% CI 13.06–40.50; I² = 87.1%), but not in HFpEF (2 studies, n = 51; MD = 54.86 m, 95% CI -44.85–154.56; I² = 0%); (4) By publication period, >2015 studies (11 studies, n = 1,750) demonstrated significant improvement with high heterogeneity (MD = 22.91 m, 95% CI 12.37–33.46; I² = 82.8%), whereas ≤ 2015 studies (6 studies, n = 2,644) showed non-significant effects (MD = 16.31 m, 95% CI -7.36–39.99; I² = 0%). Detailed results are presented in Fig. 2; Table 2, and Supplementary Figure S3.
Fig. 2.
Forest plots of 6MWT (m). 6MWT, 6-minute walk test
Table 2.
Results of subgroup analysis on the efficacy of home-based cardiac rehabilitation on 6MWT
| Subgroup Types | RCTs, n (participants, n) | MD (95%CI), meters | I² (%) | Cochran’s Q | P-value | τ² | Subgroup P |
|---|---|---|---|---|---|---|---|
| Types of digital health | < 0.001 | ||||||
| Wearable-guided | 8 (2,884) | 51.98 (34.21, 69.76) | 0 | 5.14 | 0.642 | 0 | |
| Interaction-guided | 9 (1,510) | 12.45 (3.52, 21.37) | 76.4 | 33.96 | < 0.001 | 86.07 | |
| Control type | < 0.001 | ||||||
| Usual care | 13 (4,067) | 28.31 (17.27, 39.35) | 80.1 | 60.2 | < 0.001 | 166.06 | |
| CBCR | 4 (327) | -3.64 (-17.41, 10.13) | 0 | 0.68 | 0.877 | 0 | |
| Types of HF | 0.311 | ||||||
| HFpEF | 2 (51) | 54.86 (-44.85, 154.56) | 0 | 0.68 | 0.409 | 0 | |
| HFrEF | 7 (3,554) | 13.86 (1.37, 26.36) | 12.3 | 6.84 | 0.336 | 47.99 | |
| HF Phenotypes | 8 (789) | 26.78 (13.06, 40.50) | 87.1 | 54.18 | < 0.001 | 223.43 | |
| Publication year | 0.656 | ||||||
| ≤2015 | 6 (2,644) | 16.31 (-7.36, 39.99) | 0 | 4.88 | 0.43 | 0 | |
| >2015 | 11 (1,750) | 22.91 (12.37, 33.46) | 82.8 | 58.14 | < 0.001 | 155.68 |
Abbreviations: 6MWT 6-Minute Walk Test, RCTs Randomized Controlled Trials, MD Mean Difference, CBCR Center-Based Cardiac Rehabilitation, HF Heart Failure, HFpEF Heart Failure with Preserved Ejection Fraction, HFrEF Heart Failure with Reduced Ejection Fraction
Effect of HBCR on VO₂Peak
Twelve studies (n = 3,878) demonstrated significantly greater improvement in VO₂peak with HBCR versus controls (MD = 0.96 mL/kg/min, 95% CI 0.56–1.35; Cochran’s Q = 11.3, I² = 3.0%, τ² = 0.02). Subgroup analyses revealed: (1) By DH modality, wearable-guided interventions (3 studies, n = 2,383) showed no significant difference (MD = 0.58 mL/kg/min, 95% CI -1.17–2.32; I² = 0%), whereas interaction-guided interventions (9 studies, n = 1,495) achieved clinically significant improvement (MD = 1.00 mL/kg/min, 95% CI 0.49–1.51; I² = 15.2%); (2) By control type, HBCR outperformed usual care (9 studies, n = 3,604; MD = 1.07 mL/kg/min, 95% CI 0.44–1.70; I² = 18.6%) but demonstrated comparable efficacy to CBCR (3 studies, n = 274; MD = 0.74 mL/kg/min, 95% CI -0.17–1.65; I² = 0%); (3) By HF phenotype, HBCR significantly improved VO₂peak in HFrEF (8 studies, n = 3,651; MD = 1.07 mL/kg/min, 95% CI 0.67–1.46; I² = 0%), but not in mixed phenotypes (4 studies, n = 227; MD = 0.52 mL/kg/min, 95% CI − 0.26–1.30; I² = 0%). No studies examined HFpEF; (4) By publication period, ≤ 2015 studies (4 studies, n = 2,621) showed significant improvement (MD = 1.32 mL/kg/min, 95% CI 0.17–2.47; I² = 0%), while >2015 studies (8 studies, n = 1,257) maintained significant effects with low heterogeneity (MD = 0.90 mL/kg/min, 95% CI 0.32–1.48; I² = 20.4%). Complete results are presented in Fig. 3; Table 3, and Supplementary Figure S4.
Fig. 3.
Forest plots of VO2peak (mL/kg per min). VO2peak, Peak Oxygen Uptake
Table 3.
Results of subgroup analysis on the efficacy of home-based cardiac rehabilitation on VO2 Peak
| Subgroup Types | RCTs, n (participants, n) | MD (95%CI), mL/kg/min | I² (%) | Cochran’s Q | P-value | τ² | Subgroup P |
|---|---|---|---|---|---|---|---|
| Types of digital health | 0.647 | ||||||
| Wearable-guided | 3 (2,383) | 0.58 (-1.17, 2.32) | 0 | 1.72 | 0.423 | 0 | |
| Interaction-guided | 9 (1,495) | 1.00 (0.49, 1.51) | 15.2 | 9.43 | 0.307 | 0.09 | |
| Control type | 0.559 | ||||||
| Usual care | 9 (3,604) | 1.07 (0.44, 1.70) | 18.6 | 9.83 | 0.277 | 0.11 | |
| CBCR | 3 (274) | 0.74 (-0.17, 1.65) | 0 | 1.26 | 0.531 | 0 | |
| Types of HF | 0.220 | ||||||
| HFrEF | 8 (3,651) | 1.07 (0.67, 1.46) | 0 | 6.84 | 0.445 | 0 | |
| HF Phenotypes | 4 (227) | 0.52 (-0.26, 1.30) | 0 | 2.99 | 0.393 | 0 | |
| Publication year | 0.524 | ||||||
| ≤2015 | 4 (2,621) | 1.32 (0.17, 2.47) | 0 | 2.11 | 0.549 | 0 | |
| >2015 | 8 (1,257) | 0.90 (0.32, 1.48) | 20.4 | 8.8 | 0.268 | 0.17 |
Abbreviations: VO2 Peak Peak Oxygen Uptake, RCTs Randomized Controlled Trials, MD Mean Difference, CBCR Center-Based Cardiac Rehabilitation, HF Heart Failure, HFpEF Heart Failure with Preserved Ejection Fraction, HFrEF Heart Failure with Reduced Ejection Fraction
Effect of HBCR on MLHFQ
Ten studies (n = 887) demonstrated significantly greater improvement in MLHFQ scores with HBCR versus controls (MD = -7.21, 95% CI -11.40–-3.02; Cochran’s Q = 16.2, I² = 44.3%, τ² = 16.2). Subgroup analyses revealed: (1) By DH modality, wearable-guided interventions (6 studies, n = 522) showed significant improvement (MD = -11.14, 95% CI -14.55–-7.72; I² = 0%), whereas interaction-guided interventions (4 studies, n = 365) demonstrated no significant change (MD = -2.68, 95% CI -8.45–3.09; I² = 41.4%); (2) By control type, HBCR significantly outperformed usual care (6 studies, n = 587; MD = -10.69, 95% CI -13.67–-7.71; I² = 0%) but was comparable to CBCR (4 studies, n = 300; MD = 0.83, 95% CI -4.60–6.25; I² = 0%); By HF phenotype, significant benefits were observed in HFpEF (2 studies, n = 51; MD = -18.49, 95% CI -36.12–-0.85; I² = 0%) and mixed HF phenotypes (4 studies, n = 571; MD = -6.47, 95% CI -12.27–-0.66; I² = 66.3%). However, the effect was not statistically significant in the HFrEF subgroup (4 studies, n = 265; MD = -6.04, 95% CI -15.85–3.77; I² = 48.3%); (4) By publication period, ≤ 2015 studies (4 studies, n = 205) showed non-significant improvement (MD = -11.07, 95% CI -24.65–2.50; I² = 49.9%), whereas >2015 studies (6 studies, n = 682) demonstrated significant improvement with moderate heterogeneity (MD = -6.90, 95% CI -11.26–-2.53; I² = 50.7%). Complete results are presented in Fig. 4; Table 4, and Supplementary Figure S5.
Fig. 4.
Forest plots of MLHFQ. MLHFQ, Minnesota Living with Heart Failure Questionnaire
Table 4.
Results of subgroup analysis on the efficacy of home-based cardiac rehabilitation on MLHFQ
| Subgroup Types | RCTs, n (participants, n) | MD (95%CI) | I² (%) | Cochran’s Q | P-value | τ² | Subgroup P |
|---|---|---|---|---|---|---|---|
| Types of digital health | 0.013 | ||||||
| Wearable-guided | 6 (522) | -11.14 (-14.55, -7.72) | 0 | 3.26 | 0.661 | 0 | |
| Interaction-guided | 4 (365) | -2.68 (-8.45, 3.09) | 41.4 | 5.12 | 0.163 | 13.97 | |
| Control type | < 0.001 | ||||||
| Usual care | 6 (587) | -10.69 (-13.67, -7.71) | 0 | 2.22 | 0.818 | 0 | |
| CBCR | 4 (300) | 0.83 (-4.60, 6.25) | 0 | 0.63 | 0.889 | 0 | |
| Types of HF | 0.43 | ||||||
| HFpEF | 2 (51) | -18.49 (-36.12, -0.85) | 0 | 0 | 0.956 | 0 | |
| HFrEF | 4 (265) | -6.04 (-15.85, 3.77) | 48.3 | 5.8 | 0.122 | 45.06 | |
| HF Phenotypes | 4 (571) | -6.47 (-12.27, -0.66) | 66.3 | 8.91 | 0.031 | 22.26 | |
| Publication year | 0.566 | ||||||
| ≤2015 | 4 (205) | -11.07 (-24.65, 2.50) | 49.9 | 5.99 | 0.112 | 93.68 | |
| >2015 | 6 (682) | -6.90 (-11.26, -2.53) | 50.7 | 10.15 | 0.071 | 13.39 |
Abbreviations: MLHFQ Minnesota Living with Heart Failure Questionnaire, RCTs Randomized Controlled Trials, MD Mean Difference, CBCR Center-Based Cardiac Rehabilitation, HF Heart Failure, HFpEF Heart Failure with Preserved Ejection Fraction, HFrEF Heart Failure with Reduced Ejection Fraction
Publication bias and sensitivity analyses
For the 6MWT, funnel plot asymmetry was observed (Supplementary Figure S6), and Egger’s test indicated potential publication bias (P = 0.012). Sensitivity analyses showed MD estimates ranging from 16.02 to 27.41 with moderate-to-high heterogeneity (I² ranging from 65.1% to 76.4%). For VO₂peak, symmetrical funnel plots (Supplementary Figure S7) and non-significant Egger’s test (P = 0.771) indicated no publication bias, with sensitivity analyses confirming stable effects (MD: 0.91–1.04) and low heterogeneity (I² = 0.0–11.8%). For MLHFQ, funnel plot symmetry (Supplementary Figure S8) and non-significant Egger’s test (P = 0.632) suggested no publication bias, with consistent sensitivity MD estimates (-8.9 to -6.15) and low heterogeneity (I² = 21.3–50.3%).
Discussion
This systematic review suggests that DH-mediated HBCR is associated with improvements in exercise capacity, cardiorespiratory fitness, and quality of life in patients with HF. However, preliminary evidence indicates potential efficacy variations between wearable-guided and interaction-guided interventions, particularly for functional improvement (6MWT) and quality of life (MLHFQ). These findings should be interpreted cautiously given the high or unclear risk of bias in several included studies and substantial heterogeneity in the 6MWT analysis. This finding aligns with prior meta-analyses: Gao et al. demonstrated the superiority of telerehabilitation over usual care and its non-inferiority to CBCR [8], while Tang et al. reported significant improvements in 6-minute walk distance and VO₂peak with noninvasive telerehabilitation compared to conventional care [15]. Crucially, existing studies conceptualized telerehabilitation as a homogeneous intervention without systematically comparing the distinct technical implementations and mechanistic pathways of wearable-guided interventions versus interaction-guided interventions. By implementing DH modality-stratified subgroup analyses, this study provides systematic evidence of these differential efficacy patterns.
Notably, substantial heterogeneity (I² = 74.8%) was observed in the pooled 6MWT analysis. Prespecified subgroup analyses (by DH modality, control type, HF phenotype, and publication year) and leave-one-out sensitivity analysis did not fully explain this variability. Plausible unmeasured contributors include differences in intervention duration, exercise prescription parameters, control group conditions, patient baseline characteristics, and healthcare systems. While the pooled 6MWT effect size should be interpreted with caution, leave-one-out analysis confirmed the overall conclusion of improved 6MWT performance with DH-mediated HBCR is robust. Importantly, the HFpEF subgroup result for 6MWT (MD = 54.86 m, 95% CI -44.85–154.56) requires extreme caution. Derived from only 2 small studies (n = 51), this estimate has insufficient statistical power and a wide confidence interval crossing the null value. This exploratory finding cannot be generalized to clinical practice and requires validation in larger, well-powered HFpEF-specific cohorts.
Importantly, the observed differences in efficacy profiles between the two modalities are based on indirect cross-study subgroup comparisons rather than head-to-head randomized trials, which represents a major limitation of this analysis. While indirect comparisons suggested differential efficacy patterns—with wearable-guided interventions exhibiting larger effect sizes for exercise capacity (6MWT, MD = 51.98 m) and quality of life (MLHFQ, MD = -11.14), and interaction-guided interventions showing a more pronounced effect on cardiorespiratory fitness (VO₂peak, MD = 1.00 mL/kg/min)—these findings should be considered preliminary and hypothesis-generating only. They may reflect inter-study heterogeneity in patient populations, exercise prescriptions, and methodological quality rather than true inherent differences between intervention modalities. These outcome differences arise from inherent operational variations: wearable-guided systems (monitors, smartwatches) utilize continuous physiological monitoring [11] to generate automated biofeedback and safety alerts, promoting activity adherence that significantly increases exercise capacity and quality of life (QoL) metrics, albeit with constrained capacity to optimize high-intensity cardiopulmonary training parameters. Conversely, interaction-guided platforms (video consultations, app-coaching) employ clinician-mediated communication cycles [12, 13] that dynamically adjust exercise prescriptions based on real-time patient data, demonstrating superior cardiopulmonary outcome optimization but reduced secondary effects on activity-sustaining behavioral reinforcement.
Understanding the distinct functional characteristics and efficacy profiles of DH modalities is essential for overcoming inherent limitations of HBCR, specifically inadequate supervision, delayed protocol adjustments, and insufficient psychosocial support, to achieve personalized care. Our analysis negates the concept of a universally optimal DH intervention, as clinical value derives from precise matching with patient-specific rehabilitation trajectories. In the included studies, wearable-guided interventions showed greater effects on 6MWT and MLHFQ, while interaction-guided interventions achieved greater improvements in VO₂peak. However, these subgroup comparisons were conducted at the study level, and the population adaptability at the individual patient level has not been verified due to the lack of patient-level data. These modalities exhibit complementary synergy: wearable-generated data actively refines interactive guidance during telehealth sessions. Future research should collect individual patient data to develop evidence-based selection criteria for optimal single-modality or hybrid HBCR delivery.
Currently, the DH-mediated HBCR model remains in its early exploratory stages. While its efficacy is well-established, its development faces significant challenges due to technical gaps and research limitations. Engagement is hampered among older adults and populations with low digital literacy due to the complexity of device operation [38, 39]. Furthermore, data security and privacy protection mechanisms require further refinement [8]. The majority of included studies have follow-up periods shorter than three months, lacking data on long-term outcomes such as mortality and readmission rates, as well as comprehensive cost-effectiveness analyses. There is also insufficient research coverage of women, patients in low- and middle-income countries, and those with poorer cardiac function (e.g., NYHA class III-IV). Future research should integrate the advantages of various DH models. This includes leveraging wearable devices for real-time data acquisition, facilitating provider-patient communication via telemedicine, and utilizing AI and big data analytics for deep data insights. This integrated approach should explore optimal intervention combinations tailored to diverse populations.
This study has several important limitations. First, the definition and purity of intervention modalities remain ambiguous. Despite classification by primary technical components, “wearable-guided” and “interaction-guided” modes may overlap in practice, necessitating stricter definitions and standardized reporting of core features in future research. Second, all subgroup analyses were performed at the study level, without examining patient-level characteristics (age, baseline function, disease severity, digital literacy), precluding individual-level intervention selection. Third, comparisons between modalities relied on indirect cross-study analyses rather than head-to-head RCTs. Observed efficacy differences may reflect inter-study variations in populations, exercise prescriptions, and methodological quality rather than inherent modality differences, and high or unclear risk of bias across multiple domains further undermines the certainty of these subgroup findings. Fourth, significant within-mode heterogeneity existed, particularly in exercise parameters, educational support, and interaction intensity in the interaction-guided group, potentially confounding core effect estimates. Notably, substantial inconsistency in the pooled 6MWT analysis, combined with critical bias and significant publication bias, resulted in this outcome being graded as very low certainty by GRADE. Fifth, generalizability is limited by predominantly male participants from high-income countries, restricting applicability to women, older adults, low-digital-literacy populations, and low- and middle-income countries. Sixth, systematic technology adherence data were lacking, precluding analysis of how adherence influenced clinical outcomes. Collectively, these limitations result in an overall moderate to very low quality of evidence, and findings should be interpreted cautiously.
Conclusions
This systematic review and meta-analysis indicate that DH-mediated HBCR is associated with improved exercise capacity, cardiorespiratory fitness, and quality of life in patients with heart failure. Preliminary subgroup analyses suggest wearable-guided interventions may particularly improve exercise capacity and quality of life, whereas interaction-guided approaches may be more effective for enhancing cardiorespiratory fitness. However, these findings are hypothesis-generating due to methodological limitations. Further research should focus on standardized protocols and head-to-head trials to inform clinical practice.
Supplementary Information
Acknowledgements
Not applicable.
Authors’ contributions
ZZX contributed to study conception and design, literature retrieval and screening, quality assessment, data analysis, and critical revisions; FHM contributed to study conception and design; LJM contributed to literature retrieval and screening, quality assessment, manuscript drafting, and critical revisions; GY contributed to literature retrieval and screening and quality assessment; TXY contributed to data analysis; PXH contributed to manuscript drafting; LHD contributed to study conception, result interpretation, and manuscript drafting. All authors approved the final manuscript and agree to be accountable for all aspects of the work.
Funding
This research was funded by the Science and Technology Program of Zhejiang Provincial Health Commission (grant number 2021KY1016) and the Project of Ningbo Leading Medical & Health Discipline (project number 2026-A29).
Data availability
As this research constitutes a systematic review and meta-analysis of existing published literature, no original datasets were generated or analyzed beyond the aggregated data extracted from the included studies. All data supporting the findings of this meta-analysis (including forest plots, pooled estimates, sensitivity analyses, and subgroup results) are fully available within this published article and its supplementary materials.
Declarations
Ethics approval and consent to participate
Not applicable.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Tanai E, Frantz S. Pathophysiology of Heart Failure. Compr Physiol. 2015;6(1):187–214. [DOI] [PubMed] [Google Scholar]
- 2.McDonagh TA, Metra M, Adamo M, Gardner RS, Baumbach A, Böhm M, et al. 2021 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure. Eur Heart J. 2021;42(36):3599–726. [DOI] [PubMed] [Google Scholar]
- 3.Becher PM, Lund LH, Coats AJS, Savarese G. An update on global epidemiology in heart failure. Eur Heart J. 2022;43(32):3005–7. [DOI] [PubMed] [Google Scholar]
- 4.Thomas RJ, Beatty AL, Beckie TM, Brewer LC, Brown TM, Forman DE, et al. Home-Based Cardiac Rehabilitation: A Scientific Statement From the American Association of Cardiovascular and Pulmonary Rehabilitation, the American Heart Association, and the American College of Cardiology. J Am Coll Cardiol. 2019;74(1):133–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Li L, Ringeval M, Wagner G, Paré G, Ozemek C, Kitsiou S. Effectiveness of home-based cardiac rehabilitation interventions delivered via mHealth technologies: a systematic review and meta-analysis. Lancet Digit health. 2025;7(4):e238–54. [DOI] [PubMed] [Google Scholar]
- 6.Nabutovsky I, Nachshon A, Klempfner R, Shapiro Y, Tesler R. Digital Cardiac Rehabilitation Programs: The Future of Patient-Centered Medicine. Telemedicine J e-health: official J Am Telemedicine Association. 2020;26(1):34–41. [DOI] [PubMed] [Google Scholar]
- 7.Harbi AS, Soh KL, Yubbu PB, Soh KG. Digital health intervention in patients undergoing cardiac rehabilitation: systematic review and meta-analysis. F1000Research. 2024;13:596. [DOI] [PMC free article] [PubMed]
- 8.Gao Y, Wang N, Zhang L, Liu N. Effectiveness of home-based cardiac telerehabilitation in patients with heart failure: a systematic review and meta-analysis of randomised controlled trials. J Clin Nurs. 2023;32(21–22):7661–76. [DOI] [PubMed]
- 9.Zwack CC, Haghani M, Hollings M, Zhang L, Gauci S, Gallagher R, et al. The evolution of digital health technologies in cardiovascular disease research. NPJ Digit Med. 2023;6(1):1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Mathews SC, McShea MJ, Hanley CL, Ravitz A, Labrique AB, Cohen AB. Digital health: a path to validation. NPJ Digit Med. 2019;2(1):38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Piotrowicz E, Piotrowicz R. Cardiac telerehabilitation: current situation and future challenges. Eur J Prev Cardiol. 2013;20(2 Suppl):12–6. [DOI] [PubMed] [Google Scholar]
- 12.Burke LE, Ma J, Azar KM, Bennett GG, Peterson ED, Zheng Y, et al. Current Science on Consumer Use of Mobile Health for Cardiovascular Disease Prevention: A Scientific Statement From the American Heart Association. Circulation. 2015;132(12):1157–213. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Meddar JM, Ponnapalli A, Azhar R, Turchioe MR, Duran AT, Creber RM. A Structured Review of Commercially Available Cardiac Rehabilitation mHealth Applications Using the Mobile Application Rating Scale. J Cardiopulm Rehabil Prev. 2022;42(3):141–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Jones AK, Yan CL, Rivera Rodriquez BP, Kaur S, Andrade-Bucknor S. Role of wearable devices in cardiac telerehabilitation: a scoping review. PLoS ONE. 2023;18(5):e0286714. [DOI] [PMC free article] [PubMed]
- 15.Tang J, Ma D, Li M, Zhang X, Zhou X, He H. Effectiveness of non-invasive remote rehabilitation in patients with heart failure after discharge: a meta-analysis. Eur J Physiotherapy. 2025;27(5):316–24.
- 16.Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ (Clinical Res ed). 2021;372:n71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Sterne JAC, Savović J, Page MJ, Elbers RG, Blencowe NS, Boutron I, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ (Clinical Res ed). 2019;366:l4898. [DOI] [PubMed] [Google Scholar]
- 18.Gary RA, Sueta CA, Dougherty M, Rosenberg B, Cheek D, Preisser J, et al. Home-based exercise improves functional performance and quality of life in women with diastolic heart failure. Heart lung: J Crit care. 2004;33(4):210–8. [DOI] [PubMed] [Google Scholar]
- 19.Gary R, Lee SY. Physical function and quality of life in older women with diastolic heart failure: effects of a progressive walking program on sleep patterns. Prog Cardiovasc Nurs. 2007;22(2):72–80. [DOI] [PubMed] [Google Scholar]
- 20.O’Connor CM, Whellan DJ, Lee KL, Keteyian SJ, Cooper LS, Ellis SJ, et al. Efficacy and safety of exercise training in patients with chronic heart failure: HF-ACTION randomized controlled trial. JAMA. 2009;301(14):1439–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Gary RA, Cress ME, Higgins MK, Smith AL, Dunbar SB. Combined aerobic and resistance exercise program improves task performance in patients with heart failure. Arch Phys Med Rehabil. 2011;92(9):1371–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Antonicelli R, Spazzafumo L, Scalvini S, Olivieri F, Matassini MV, Parati G, et al. Exercise: a new drug for elderly patients with chronic heart failure. Aging. 2016;8(5):860–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Safiyari-Hafizi H, Taunton J, Ignaszewski A, Warburton DE. The Health Benefits of a 12-Week Home-Based Interval Training Cardiac Rehabilitation Program in Patients With Heart Failure. Can J Cardiol. 2016;32(4):561–7. [DOI] [PubMed] [Google Scholar]
- 24.Bernocchi P, Vitacca M, La Rovere MT, Volterrani M, Galli T, Baratti D, et al. Home-based telerehabilitation in older patients with chronic obstructive pulmonary disease and heart failure: a randomised controlled trial. Age Ageing. 2018;47(1):82–8. [DOI] [PubMed] [Google Scholar]
- 25.Andrade GND, Umeda IIK, Fuchs ARCN, Mastrocola LE, Rossi-Neto JM, Moreira DAR, et al. Home-based training program in patients with chronic heart failure and reduced ejection fraction: a randomized pilot study. Clinics (Sao Paulo). 2021;76:e2550. [DOI] [PMC free article] [PubMed]
- 26.Nagatomi Y, Ide T, Higuchi T, Nezu T, Fujino T, Tohyama T, et al. Home-based cardiac rehabilitation using information and communication technology for heart failure patients with frailty. ESC heart Fail. 2022;9(4):2407–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Piotrowicz E, Baranowski R, Bilinska M, Stepnowska M, Piotrowska M, Wójcik A, et al. A new model of home-based telemonitored cardiac rehabilitation in patients with heart failure: effectiveness, quality of life, and adherence. Eur J Heart Fail. 2010;12(2):164–71. [DOI] [PubMed] [Google Scholar]
- 28.Piotrowicz E, Zieliłski T, Bodalski R, Rywik T, Dobraszkiewicz-Wasilewska B, Sobieszczałska-Małek M, et al. Home-based telemonitored Nordic walking training is well accepted, safe, effective and has high adherence among heart failure patients, including those with cardiovascular implantable electronic devices: A randomised controlled study. Eur J Prev Cardiol. 2015;22(11):1368–77. [DOI] [PubMed] [Google Scholar]
- 29.Smolis-Bąk E, Dąbrowski R, Piotrowicz E, Chwyczko T, Dobraszkiewicz-Wasilewska B, Kowalik I, et al. Hospital-based and telemonitoring guided home-based training programs: Effects on exercise tolerance and quality of life in patients with heart failure (NYHA class III) and cardiac resynchronization therapy. A randomized, prospective observation. Int J Cardiol. 2015;199:442–7. [DOI] [PubMed] [Google Scholar]
- 30.Piotrowicz E, Piotrowski W, Piotrowicz R. Positive Effects of the Reversion of Depression on the Sympathovagal Balance after Telerehabilitation in Heart Failure Patients. Annals noninvasive electrocardiology: official J Int Soc Holter Noninvasive Electrocardiol Inc. 2016;21(4):358–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Hwang R, Bruning J, Morris NR, Mandrusiak A, Russell T. Home-based telerehabilitation is not inferior to a centre-based program in patients with chronic heart failure: a randomised trial. J physiotherapy. 2017;63(2):101–7. [DOI] [PubMed] [Google Scholar]
- 32.Peng X, Su Y, Hu Z, Sun X, Li X, Dolansky MA, et al. Home-based telehealth exercise training program in Chinese patients with heart failure: A randomized controlled trial. Medicine. 2018;97(35):e12069. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Piotrowicz E, Pencina MJ, Opolski G, Zareba W, Banach M, Kowalik I, et al. Effects of a 9-Week Hybrid Comprehensive Telerehabilitation Program on Long-term Outcomes in Patients With Heart Failure: The Telerehabilitation in Heart Failure Patients (TELEREH-HF) Randomized Clinical Trial. JAMA Cardiol. 2020;5(3):300–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Gary RA, Paul S, Corwin E, Butts B, Miller AH, Hepburn K, et al. Exercise and Cognitive Training Intervention Improves Self-Care, Quality of Life and Functional Capacity in Persons With Heart Failure. J Appl Gerontol. 2022;41(2):486–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Haiyan P, Xuan X, Lihua C, Liang Z, Yuan L, Xiaohui W. Efficacy and safety of home-based exercise rehabilitation in chronic heart failure with reduced ejection fraction patients under the guidance of remote cardiogram monitoring. J Clin Cardiol. 2022;38(6):478–83. [Google Scholar]
- 36.Lundgren KM, Langlo KAR, Salvesen Ø, Zanaboni P, Cittanti E, Mo R, et al. Feasibility of telerehabilitation for heart failure patients inaccessible for outpatient rehabilitation. ESC heart Fail. 2023;10(4):2406–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Schmidt C, Magalhães S, Gois Basilio P, Santos C, Oliveira MI, Ferreira JP, et al. Center- vs Home-Based Cardiac Rehabilitation in Patients With Heart Failure: EXIT-HF Randomized Controlled Trial. JACC Heart failure; 2025. [DOI] [PubMed]
- 38.Berry R, Brawner CA, Kipa SG, Stevens C, Bloom C, Keteyian SJ. Telemedicine Home-Based Cardiac Rehabilitation: A CASE SERIES. J Cardiopulm Rehabil Prev. 2020;40(4):245–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Ambrosetti M, Abreu A, Corrà U, Davos CH, Hansen D, Frederix I, et al. Secondary prevention through comprehensive cardiovascular rehabilitation: From knowledge to implementation. 2020 update. A position paper from the Secondary Prevention and Rehabilitation Section of the European Association of Preventive Cardiology. Eur J Prev Cardiol. 2021;28(5):460–95. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
As this research constitutes a systematic review and meta-analysis of existing published literature, no original datasets were generated or analyzed beyond the aggregated data extracted from the included studies. All data supporting the findings of this meta-analysis (including forest plots, pooled estimates, sensitivity analyses, and subgroup results) are fully available within this published article and its supplementary materials.