Home-Based Cardiac Rehabilitation Alone and Hybrid With Center-Based Cardiac Rehabilitation in Heart Failure: A Systematic Review and Meta-Analysis.

Background Center-based cardiac rehabilitation (CBCR) has been shown to improve outcomes in patients with heart failure (HF). Home-based cardiac rehabilitation (HBCR) can be an alternative to increase access for patients who cannot participate in CBCR. Hybrid cardiac rehabilitation (CR) combines short-term CBCR with HBCR, potentially allowing both flexibility and rigor. However, recent data comparing these initiatives have not been synthesized. Methods and Results We performed a meta-analysis to compare functional capacity and health-related quality of life (hr-QOL) outcomes in HF for (1) HBCR and usual care, (2) hybrid CR and usual care, and (3) HBCR and CBCR. A systematic search in 5 standard databases for randomized controlled trials was performed through January 31, 2019. Summary estimates were pooled using fixed- or random-effects (when I2>50%) meta-analyses. Standardized mean differences (95% CI) were used for distinct hr-QOL tools. We identified 31 randomized controlled trials with a total of 1791 HF participants. Among 18 studies that compared HBCR and usual care, participants in HBCR had improvement of peak oxygen uptake (2.39 mL/kg per minute; 95% CI, 0.28-4.49) and hr-QOL (16 studies; standardized mean difference: 0.38; 95% CI, 0.19-0.57). Nine RCTs that compared hybrid CR with usual care showed that hybrid CR had greater improvements in peak oxygen uptake (9.72 mL/kg per minute; 95% CI, 5.12-14.33) but not in hr-QOL (2 studies; standardized mean difference: 0.67; 95% CI, -0.20 to 1.54). Five studies comparing HBCR with CBCR showed similar improvements in functional capacity (0.0 mL/kg per minute; 95% CI, -1.93 to 1.92) and hr-QOL (4 studies; standardized mean difference: 0.11; 95% CI, -0.12 to 0.34). Conclusions HBCR and hybrid CR significantly improved functional capacity, but only HBCR improved hr-QOL over usual care. However, both are potential alternatives for patients who are not suitable for CBCR.

Clinical Perspective

What Is New?

Both home‐based and hybrid (combination of home‐ and center‐based) cardiac rehabilitation significantly improved functional capacity compared with usual care.

Home‐based cardiac rehabilitation also improved health‐related quality of life.

What Are the Clinical Implications?

Both home‐based and hybrid cardiac rehabilitation are reasonable alternatives for patients who cannot attend center‐based cardiac rehabilitation.

Introduction

Center‐based cardiac rehabilitation (CBCR) is safe and has been shown to improve functional capacity, cardiac function, and health‐related quality of life (hr‐QoL) in patients with heart failure (HF).1, 2, 3, 4 However, multiple barriers such as lack of transportation or conflicting schedules often result in nonparticipation in CBCR among HF patients.5, 6, 7 In these patients, home‐based cardiac rehabilitation (HBCR) can be a reasonable alternative to offer exercise‐based cardiac rehabilitation (CR).8 Although previous systematic reviews and meta‐analyses have compared outcomes of HBCR with CBCR and usual care,9, 10, 11 sample sizes were limited and at least 11 additional randomized controlled trials (RCTs) have been published since the last meta‐analysis. In addition, several studies have used a hybrid approach (hybrid CR) combining short‐term CBCR and HBCR as another alternative to either HBCR or CBCR alone, the effects of which have yet to be analyzed in a systematic fashion. Consequently, we undertook the following systematic review and meta‐analysis of RCTs to compare the outcomes of HF patients who underwent (1) HBCR versus usual care, (2) hybrid CR versus usual care, and (3) HBCR versus CBCR.

Methods

The data that support the findings of this study are available from the corresponding author on reasonable request. We conducted and report this systematic review and meta‐analysis in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta‐Analyses) guidelines.12

Eligibility Criteria

We included RCTs that evaluated exercise‐based HBCR or hybrid CR against CBCR or usual care in adult patients aged ≥18 with a diagnosis of HF with reduced ejection fraction or HF with preserved ejection fraction.

Types of Exercise Interventions

In HBCR studies, patients received an exercise training prescription of at least 3 hours of exercise per week exclusively at home or in a community setting (eg, gymnasium, senior center), without real‐time supervision by a CR professional during exercise, for at least 8 weeks but <12 months.

In CBCR studies, patients participated in exercise training sessions with real‐time supervision by a CR professional in either a hospital or physician's office–based CR center for at least 3 hours/week for ≥8 weeks up to 12 months.

In hybrid CR, patients participated in exercise training sessions of at least 3 hours/week in combined center‐ and home‐based settings for at least 8 weeks up to 12 months. Exercise training sessions can occur in a sequential fashion, with initial short‐term CBCR of ≤4 weeks followed by HBCR for at least another 4 weeks, or in a concomitant fashion in both settings where the center‐based exercise sessions did not exceed 50% of total prescribed exercise sessions.

In usual care, no exercise intervention was prescribed.

Information Sources

We performed the search on January 31, 2019, in CENTRAL (Cochrane Library; 1944 to issue 2 of 2019), MEDLINE (Ovid; 1879 to January 2019), EMBASE (Ovid; 1945 to January 2019), PsycINFO (Ovid; 1927 to January 2019), and CINAHL Plus (EBSCO; 1976 to January 2019). Further information was sought by hand‐searching the bibliographies of selected papers and through contacts with the authors of the published papers. Only studies that showed results (paper or abstract form) in English language were included.

Search

A system‐wide electronic search was performed using the following terms: heart failure and cardiac rehabilitation, rehabilitation, exercise therapy, exertion, fitness training, or exercise. Two authors (H.I.) and (M.B.) independently screened all searched articles and discarded irrelevant titles. Both authors independently reviewed articles that met the criteria. Any discrepancy was resolved after review by a third author (W.W.).

Of the 53 studies eligible for review, 22 studies were excluded after complete review. Studies were excluded if study participants had previously completed CBCR, had a different inclusion criterion other than HF (eg, chronic obstructive pulmonary disease or obstructive sleep apnea), did not report quantifiable outcomes, had duplication of data from previous studies, or had a crossover design (Table S1).

Study Outcomes

We evaluated functional capacity and hr‐QoL as primary outcomes, and hospital admissions (all‐cause and cardiac) and all‐cause mortality as secondary outcomes.

Functional capacity was measured by cardiopulmonary exercise test, exercise stress test, 6‐minute walk test or incremental shuttle walk test. When studies assessed functional capacity using both maximal stress (ie, cardiopulmonary exercise test or exercise stress test) and submaximal stress (ie, 6‐minute walk test or incremental shuttle walk test), data from maximal stress were extracted. For uniformity of data, all measures of functional capacity were converted to peak oxygen uptake by using the following formulas: Peak oxygen uptake=4.948+[0.023×6‐minute walk distance (meters)] or Peak oxygen uptake=4.19+[0.025×incremental shuttle walk distance (meters)].13, 14

For hr‐QOL, studies reported general hr‐QOL, HF‐specific hr‐QOL questionnaires, or both. For those that reported both, we pooled the scores only from HF‐specific questionnaires—Minnesota Living with Heart Failure (MLWHF),15 Kansas City Cardiomyopathy Questionnaire (KCCQ),16 Chronic Heart Failure Questionnaire (CHF Questionnaire),17 and Heart Failure Functional Status Inventory18—to be specific, when possible, for HF. For the remaining studies, we pooled the hr‐QOL tool that was utilized, general or HF‐specific. For uniformity purposes, mean differences of the scores from these distinct questionnaires were standardized by dividing them by their own standard deviation and then reported as standardized mean differences (SMDs). The scores in MLWHF were reverse coded because, unlike other hr‐QOL tools, a decrease in MLWHF score indicates better hr‐QOL.

Risk for all‐cause and cardiac‐specific hospitalizations (HF, revascularization, acute myocardial infarction) were extracted as relative risks.

All‐cause mortality was low overall, and deaths were counted if they occurred during the study period.

Data Extraction

Some studies did not report a direct comparison of the change from baseline between the arms; instead, they only compared postintervention scores. In those cases, mean change from baseline within each study arm was calculated by subtracting mean functional capacity or hr‐QoL score at entrance from the exit value. To err on the conservative side, the higher standard deviation between the entry or the exit results was selected as the standard deviation for the change from baseline within each arm.

Risk Assessment for Bias in Included Studies

We evaluated the risk of bias in studies by using Cochrane Risk of Bias Tool for RCTs.19 We reviewed studies for evidence of balance in baseline characteristics of groups. The risk of bias was assessed by each reviewer (M.B. and H.I.) independently (Table S2) for all studies.

Statistical Analysis

For continuous variables, means and standard deviations were extracted. SMD with 95% CI was calculated for functional capacity and hr‐QOL scores. For categorical variables, relative risk with 95% CI was calculated for hospitalizations and mortality outcomes if applicable. Using a fixed‐effects or DerSimonian and Laird random‐effects model (when I2>50%), results from included studies were pooled to give an overall estimate of the treatment effect20 to test 3 a priori hypotheses in HF: (1) HBCR improves functional capacity, hr‐QOL, and all‐cause hospitalization over usual care; (2) hybrid CR improves functional capacity, hr‐QOL, and all‐cause hospitalization over usual care; and (3) HBCR improves functional capacity and hr‐QOL over CBCR.

Heterogeneity in included studies was explored quantitatively using Q statistics. Funnel plots and the Egger test were performed to assess publication bias.21, 22 Subgroup analyses, based on duration of exercise training in CR, maximal or submaximal effort at functional capacity evaluation, HF with preserved versus reduced ejection fraction, or type of questionnaire used to assess hr‐QoL, were performed to assess potential sources of heterogeneity, as applicable. A 2‐sided P value <0.05 was considered significant. Stata SE v15.0 (StataCorp) was used for analysis.

Results

Description of Studies

Search results

The PRISMA flow diagram is shown in Figure 1. We identified 11 studies in addition to the 20 studies included in previous systemic reviews and meta‐analyses,10, 11 for a total of 31 studies and 1791 participants with HF. All studies but 2 had 2 comparison arms. The study by Gary et al had 4 arms: exercise only, cognitive behavior therapy only, exercise and cognitive behavior therapy, and usual care.23 Exercise only and exercise with cognitive behavior therapy arms were compared with usual care. The study by Cowie et al had 3 arms: HBCR, CBCR, and usual care. We compared HBCR with (1) CBCR and (2) usual care.24 All studies reported outcomes in patients with HF with reduced ejection fraction except for Gary et al and Lang et al, who reported outcomes for patients with HF with preserved ejection fraction.25, 26

Figure 1

Comparison of functional capacity between home‐based cardiac rehabilitation and usual care. CBCR indicates center‐based cardiac rehabilitation; RCTs, randomized controlled trials.

Exercise Training

The majority (13/18) of HBCR studies enrolled patients in the high‐risk category (based on low left ventricular ejection fraction) of the American Association of Cardiovascular Pulmonary Rehabilitation (AACVPR). Participants (in 14/18 studies) were prescribed mild‐ to moderate‐intensity aerobic exercises (mostly walking) with target heart rates at 40% to 75% of the maximal heart rate achieved during stress test or at a rate of perceived exertion of 11 to 13 on a Borg Scale. One study prescribed high‐intensity aerobic (interval) exercise training.27 Strength training and stretch exercises were prescribed in relatively fewer studies (8/18). Seven of 9 hybrid CR studies enrolled patients with high AACVPR risk, but they were also prescribed only mild‐ to moderate‐intensity exercises. The participants were able to perform the prescribed exercises in nearly all RCTs without any adverse outcomes except for 1 study in which a patient was hospitalized for hypoglycemia during exercise.28 Overall adherence to prescribed exercise sessions was 74% to 110% for HBCR and 60% to 80% for hybrid CR compared with 86% to 97% for CBCR. Program completion rates were 44% to 100% for HBCR and 31% to 100% for hybrid CR versus 72% to 100% for CBCR.

Effect of Intervention

Functional capacity

HBCR versus usual care

Eighteen studies involving 1191 participants compared functional capacity of HBCR and usual care (Table S3). On pooled assessment using a random‐effects model, functional capacity significantly increased in HBCR (678 participants) compared with usual care (563 participants; 2.39 mL/kg per minute; 95% CI, 0.28–4.49; I2=83.1%; Figure 2).20, 23, 24, 25, 26, 27, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40 Subgroup analysis based on HF subtype showed functional capacity significantly improved only in HF with reduced ejection fraction (1109 participants; 3.18 mL/kg per minute; 95% CI, 0.95–5.47) but not in HF with preserved ejection fraction (82 participants; 0.42 mL/kg per minute; 95% CI, −2.52 to 3.58). On metaregression analyses, neither duration (P=0.37) nor method of functional capacity evaluation (P=0.15) significantly explained the heterogeneity of results.

Figure 2

Functional capacity significantly improved in home‐based cardiac rehabilitation (HBCR) compared with usual care; 1 standardized mean difference (SMD)=7.02 mL/kg per minute.

Hybrid CR versus usual care

Nine studies, involving 306 participants, reported functional capacity at end of exercise training (n=155, hybrid CR; n=151, usual care; Table S4). On pooled assessment by a random‐effects model, functional capacity significantly improved at the end of the intervention (9.72 mL/kg per minute; 95% CI, 5.12–14.33; I2=90.1%; Figure 3).28, 41, 42, 43, 44, 45, 46, 47, 48 Eight studies, involving 276 participants (n=140, hybrid CR; n=136, usual care), evaluated functional capacity by maximal stress, and 1 study, involving 30 (n=15, hybrid CR; n=15, usual care), assessed functional capacity by submaximal stress. On metaregression analyses, neither method of functional capacity assessment (P=0.15) nor duration of exercise intervention (P=0.13) explained the heterogeneity of the results.

Figure 3

Comparison of functional capacity between hybrid cardiac rehabilitation (CR) and usual care. Functional capacity significantly improved in hybrid CR compared with usual care; 1 standardized mean difference (SMD)=5.23 mL/kg per minute.

HBCR versus CBCR

Five studies, involving 314 participants (n=166, HBCR; n=148, CBCR), compared functional capacity with HBCR and CBCR after exercise training (Table S5). All studies compared functional capacity by cardiopulmonary exercise test or incremental shuttle walk test at the end of an intervention of ≤3 months. Both groups experienced similar improvements in functional capacity from baseline, using the fixed‐effect model (SMD: 0.00; 95% CI, −1.91 to 1.91 mL/kg per minute; I2=0%; Figure S1).

Quality of life

HBCR versus usual care

Sixteen studies (n=576, HBCR; n=505, usual care) assessed HF‐specific hr‐QoL outcomes. Eleven studies, involving 771 participants, used the MLWHF questionnaire; 2 studies used the KCCQ; 1 study used the CHF Questionnaire; 1 study used the Heart Failure Functional Symptom Inventory; and 1 study used the Short Form 36 (SF‐36) questionnaire (Table S3). On random‐effects pooled analyses, hr‐QoL significantly increased in HBCR participants compared with usual care (SMD: 0.36; 95% CI, 0.19–0.57; I2=54.7%; Figure 4).* On metaregression analysis, different questionnaires used for assessment of hr‐QOL (P=0.77) did not significantly explain the heterogeneity of results.

Figure 4

Comparison of health‐related quality of life between home‐based cardiac rehabilitation (HBCR) and usual care. Health‐relate quality of life significantly improved with HBCR compared with usual care. CHF Questionnaire indicates Chronic Heart Failure Questionnaire; KCCQ, Kansas City Cardiomyopathy Questionnaire; MLWHF, Minnesota Living with Heart Failure; SF‐36, Short Form 36; SMD, standardized mean difference.

Hybrid CR versus usual care

Two studies, involving 102 participants (n=51, hybrid CR; n=51, usual care), reported hr‐QoL outcomes and used different questionnaires (SF‐3650 and Nottingham Health Profile,51 respectively; Table S4). In the pooled estimate of studies, there was no significant difference in hr‐QoL among hybrid CR participants compared with usual care by the random‐effects model (SMD: 0.67; 95% CI, −0.20 to 1.54; I2=70.5%; Figure S2).

HBCR versus CBCR

Only 4 studies (n=155, HBCR; n=137, CBCR) reported comparison of change in hr‐QoL between HBCR and CBCR. Three studies used the SF‐ 36 and 1 study used the MLWHF (Table S5). In a pooled fixed‐effect model, there was no significant improvement in hr‐QoL between the comparison groups (SMD: 0.11; 95% CI, −0.12 to 0.34; I2=0%; Figure S3).

All‐cause hospitalization

Four studies compared all‐cause hospitalization among 458 participants. Two studies showed 6.6% (4/61) hospitalizations in the HBCR arm versus 10.9% (6/55) in the usual care arm at 3‐month follow‐up. Two other studies showed 35.7% (61/171) in the HBCR arm versus 33.3% (57/171) in the usual care arm at 1‐year follow‐up after intervention.

Five studies compared all‐cause hospitalization among 204 participants (n=102, hybrid CR; n=102, usual care). Three studies that reported outcomes after 6 months of exercise training (hybrid CR in 3/58 versus usual care in 3/57) did not show a significant difference in hospitalizations (relative risk: 0.97; 95% CI, 0.25–3.73; I2=0) between the 2 modalities (Figure S4). Two studies reported outcomes after respective follow‐up periods of 12 months (hybrid CR versus usual care: 88.9% [16/18] versus 68.4% [13/19]) and 18 months (hybrid CR versus usual care: 53.8% [14/26] versus 57.7% [15/26]).

Only Piotrowicz et al (n=75, HBCR; n=56, CBCR) reported all‐cause hospitalizations (8‐week outcome), which was none during the 8‐week study period.

Cardiac hospitalization

Three studies, with 285 participants, reported cardiac hospitalizations at 3‐ and 12‐month follow‐up periods. Two studies reported a hospitalization rate of 4.9% (3/61) in the HBCR arm versus 7.3% (4/55) in the usual care arm after 3 months of intervention. One study reported a hospitalization rate of 13.1% (11/84) for the HBCR arm versus 12.9% (11/85) for the usual care arm at 1‐year follow‐up after intervention.

Four studies reported cardiac hospitalization outcomes among 152 participants. Three studies reported outcomes for hybrid CR at 5.2% (3/58) versus usual care at 5.3% (3/57) at 6‐month follow‐up and 66.7% (12/18) versus 63.2% (12/19), respectively, at 12‐month follow‐up.

None of the studies reported cardiac hospitalizations.

All‐cause mortality

Four studies, involving 463 participants (n=235, HBCR; n=228, usual care), reported all‐cause mortality at the end of the study period. Two studies reported mortality at the end of 3 months, and the other 2 reported mortality at the end of a 1‐year study period. The mortality rate was 8.5% (11/129) in the HBCR arm versus 7.3% (9/123) in the usual care arm at 3‐month follow‐up. Two other studies reported mortality rates of 7.5% (8/106) versus 4.8% (5/105) in the HBCR and usual care arms, respectively, at 1‐year follow‐up.

Six studies reported all‐cause mortality in 224 participants. Four studies reported mortality at 6 months, 1 reported at 12 months, and 1 reported at 18 months. Mortality rates were 5.9% (4/68) in the hybrid CR arm versus 4.5% (3/67) in the usual care arm during the 6‐month study period and 5.5% (1/18) versus 0% (0/19), respectively, at 12 month follow‐up; 3 of /14 participants in the hybrid CR arm and 3 of 15 in the usual care arm had died at 18‐month follow‐up after intervention.

None of the studies reported mortality outcomes.

Assessment of publication bias

Risk of publication bias was low for functional capacity comparisons, but potential bias may exist between HBCR and usual care for hr‐QoL. Influence analysis did not show any single study that significantly influenced the overall estimate of the effect size (Figure S5). A subgroup analysis based on sample size (n<50 versus n>50 [median: 50]) yielded similar results between the 2 subgroups (P=0.62) and suggested that sample size was unlikely to be a significant component in the interpretation of the results.

Discussion

This updated review and meta‐analysis built on previous reports by increasing the sample size (1791 versus 1290 participants) and by investigating the effects of a hybrid CR model. The results showed that both HBCR and hybrid CR significantly improved functional capacity, but only HBCR improved hr‐QoL over usual care. compared with CBCR, patients in HBCR achieve similar functional capacity and hr‐QoL outcomes (summary in Table).

Table 1

Summary of Findings

Comparison	Functional Capacity	hr‐QOL	Clinical Outcomes
HBCR vs UC	HBCR over UC (2.39 mL/kg/min; 95% CI, 0.28–4.49; I2=83.1%)	HBCR over UC (0.36; 95% CI, 0.19–0.57; I2=54.7%)	Lack of statistical power to test clinical outcomes
Hybrid CR vs UC	Hybrid CR over UC (9.72 mL/kg/min; 95% CI, 5.12–14.33; I2=93.0%)	Hybrid CR equivalent to UC (0.67; 95% CI, −0.20 to 1.54; I2=70.5%)	All‐cause hospitalization was similar (RR: 0.97; 95% CI, 0.25–3.73; I2=0.0%) between hybrid CR and UC at 6 mo; it lacks statistical power to test other clinical outcomes
HBCR vs CBCR	HBCR equivalent to CBCR (0.00 mL/kg/min, 95% CI, −1.91 to 1.91; I2=0%)	HBCR equivalent to CBCR (0.11; 95% CI, −0.12 to 0.34; I2=0%)	One study reported no outcomes in either group

CBCR indicates center‐based cardiac rehabilitation; CR, cardiac rehabilitation; HBCR, home‐based cardiac rehabilitation; hr‐QOL; health‐related quality of life; RR, relative risk; UC, usual care.

Compared with previous reports, we generated pooled estimates of improvement from baseline—an important requirement in the evaluation of CR programs, as stated by the American Heart Association and American College of Cardiology Foundation (AHA/ACCF)—as opposed to pooling only exit outcomes.52 This study overcame previous limitations by converting distinct assessment methods into unified units of comparison before pooling. Functional capacity, for example, was converted into peak oxygen uptake, and the distinct hr‐QOL questionnaires were standardized by their own standard deviations using SMDs. Given the methods we utilized, our current results are unique and the effects sizes are applicable to most assessment tools currently used in CR.

Different HBCR models have been implemented safely in the past across different single‐payer systems. The included trials required either periodic face‐to‐face visits or telephone calls to assess progress. The Stanford Coronary Rehabilitation Program and Kaiser Permanente (a private insurer) developed Multifit, an HBCR program complemented by face‐to‐face follow‐up visits with a nurse. The program improved clinical outcomes and reduced healthcare resource utilization by patients with HF.53 The Veterans Health Administration initiated a telephone‐based HBCR as an alternative to CBCR, with high patient satisfaction.54 Schopfer et al reported a higher participation rate for HBCR than CBCR among veterans.55 The clinical outcomes data presented in our current meta‐analysis showed that both HBCR alone and hybrid CR were at least as safe as CBCR and had the potential to improve clinical outcomes over usual care during short‐term follow‐up. The small sample size and distinct follow‐up periods of the studies preclude a uniform assessment of the clinical outcomes. Some data showed that mobile or web‐based platforms for intervention can also be deployed successfully.56, 57 Technological advancements such as smartphone or web‐based applications should be considered in the design of new studies to enhance the efficacy and safety of current HBCR practices.

Hybrid CR is a novel model that can provide increased monitoring opportunities during the initial exercise training phase for patients who cannot successfully or safely exercise without direct monitoring by a healthcare professional. The obstacles to exercise may not always be physical and may include psychosocial concerns.58 In hybrid CR, initial sessions at the CBCR setting can address psychosocial health in susceptible individuals, increase participation in group education sessions, and tailor the exercise regimen based on direct observation. Although the effect size of functional improvement appeared to be higher in hybrid CR versus HBCR, this has not been evaluated against HBCR or CBCR in head‐to‐head comparisons for a definitive conclusion. Future RCTs should address this knowledge gap.

Although our results showed that HBCR improved functional capacity and hr‐QOL in a fashion similar to CBCR, HBCR also has limitations. The studied HBCR models did not provide opportunities for peer support and role modeling that may come from exercising in the type of group setting that is typical of CBCR.59 Despite the AHA/ACCF recommendation that the HBCR model can be an alternative to CBCR, the Centers for Medicare and Medicaid Services and the majority of private insurers have yet to implement a reimbursement model for HBCR.60 In one study, the cost of delivery of HBCR (£196.53 [1£∼$1.25]) was similar to CBCR (£221.58) for a duration of 8 weeks (average cost of approximately £100 per 4 weeks). A recent study reported similar costs for the duration of 3 months of HBCR (£362.21).26, 49 Nonetheless, the cost of CBCR in the United States is higher than costs these reports, making HBCR models potentially financially attractive. Our results suggest that policy makers and insurers should consider a viable model of reimbursement for HBCR and/or use of a hybrid CR model, summarized in the current meta‐analysis, because the known benefits of CBCR for HF patients appeared translatable into these alternative models for patients who are not eligible for traditional CBCR.

Our analysis has limitations. Similar to other meta‐analyses, our results relied on the quality and detail of reporting, which were somewhat heterogeneous and may have contributed to the degree of heterogeneity observed in our pooled analyses. Nonetheless, subgroup and bias analyses suggest that single studies were unlikely to have altered our overall results. Next, the majority of the included studies were of patients with stable HF with reduced ejection fraction; therefore, extrapolation of the findings to patients with HF with preserved ejection should be done with caution. We performed subgroup analyses to distinguish these effects before generating the pooled estimates. Finally, most studies did not report or incompletely reported clinical outcomes or had no adverse events because of short and variable follow‐up periods. Future RCTs should be designed to assess long‐term clinical outcomes of HBCR and hybrid CR in predefined follow‐up periods.

Conclusion

In this meta‐analysis of RCTs, HBCR and hybrid CR significantly improved functional capacity compared with usual care and are potentially good alternatives for patients who are not suitable for CBCR.

Sources of Funding

Research Project Grant NIH NHLBI R01HL139795 (A.R.M.), Institutional Development Award (IDeA) from NIH NIGMS P20GM103652 (A.R.M.), Career Development Award Number 7IK2BX002527 from the United States Department of Veterans Affairs Biomedical Laboratory Research and Development Program (A.R.M.).

Disclosures

None.

Table S1. Reasons for Exclusion of Studies

Table S2. Bias Analysis of Included Studies

Table S3. Studies Comparing Home‐Based Cardiac Rehabilitation vs Usual Care

Table S4. Studies Comparing Hybrid Cardiac Rehabilitation vs Usual Care

Table S5. Studies Comparing Home‐Based and Center‐Based Cardiac Rehabilitation

Figure S1. Functional capacity comparison between home‐based and center‐based cardiac rehabilitation.

Figure S2. Health‐related quality of life comparison between hybrid cardiac rehabilitation and usual care.

Figure S3. Health‐related quality of life comparison between home‐based and center‐based cardiac rehabilitation.

Figure S4. All‐cause hospitalization comparison between hybrid cardiac rehabilitation and usual care.

Figure S5. Influence analysis of health‐related quality of life between home‐based cardiac rehabilitation and usual care.

Click here for additional data file.