Cardiovascular disease (CVD) is the leading cause of death worldwide, with approximately 80% of CVD resulting from modifiable risk factors such as physical inactivity, poor dietary habits, elevated low-density lipoprotein-cholesterol and plasma glucose levels, and smoking [1]. Following a cardiac event, cardiac rehabilitation (CR) is an effective modality that enhances recovery, reduces cardiovascular mortality and risk for hospital admissions, and improves the health-related quality of life (QoL) [2]. CR is a multi-faceted, medically supervised program that addresses established core components of guideline-directed therapy, including baseline patient assessments, nutritional counseling, risk factor modification (including management of lipids, blood pressure, weight, diabetes mellitus, and smoking), psychosocial interventions, and physical activity counseling and exercise training [3]. Although there is strong evidence supporting the benefits of CR, less than 20% of patients who are eligible participate in CR [4]. Challenges related to the low utilization of CR include the lack of referral or facilitation of enrollment, limited health insurance coverage, time and costs associated with participation and travel, and lack of access to a CR facility because of scheduling, transportation, or distance [5].

The technology for CR is advancing rapidly and has the potential to address the challenges of traditional facility-based CR programs by delivering care to patients in the convenience of their own homes with real-time, personalized support. As noted in the literature, the terminology describing this technology has not been standardized and includes telemedicine, telehealth, and eHealth [6,7]. In this review, we use the term digital health interventions to encompass technology that enables the delivery of care through means such as the use of the internet, wearable devices, and mobile apps [8,9]. Although there have been encouraging results from the use of digital health interventions for CR (eg, remote electrocardiographic monitoring and mobile or web portal tools), these developments have largely remained in the research settings and have not yet translated into widespread use in clinical practice [3]. Currently, there are gaps in understanding the comprehensiveness of digital CR programs and how successful they are in addressing the core components of CR. To help guide the development of digital CR interventions that have the potential to translate into clinical use, we have focused on the evaluation of technology used in digital interventions for CR and the comprehensiveness of these programs using the framework outlined in the scientific statement from the American Heart Association (AHA) and the American Association of Cardiovascular and Pulmonary Rehabilitation (AACVPR) for the core components of CR [3]. The specifics regarding the accreditation of CR programs are beyond the scope of this review. With the increasing need for technological advancements to revolutionize the delivery of CR care, this systematic literature review: (1) summarizes the current literature on digital interventions for CR, (2) identifies the key components of CR that have been successfully addressed through digital interventions, and (3) describes the gaps in research that need to be addressed for the sustainable implementation of digital CR interventions in clinical practice.

We designed a systematic, thematic review to answer key questions regarding the study designs to evaluate CR interventions, technology used, study size, and comprehensiveness of the investigated interventions. A full list of questions is provided in Textbox 1. Our search terms are detailed in Multimedia Appendix 1. We searched the PubMed (MEDLINE), Embase, CINAHL, and Cochrane databases for studies on digital CR published in English between January 1, 1990, and October 18, 2018. For this review, digital is defined as technology employed to deliver remote care beyond the use of telephone (eg, the delivery of care using the internet, wearable devices, and mobile apps). Telephonic-only studies, which have been addressed in the 2019 scientific statement on home-based CR from the AACVPR, AHA, and American College of Cardiology (ACC) [10], are not within the scope of this review. To determine eligibility for inclusion in this study, titles and abstracts were screened for relevance before a full-text review. The inclusion criteria for this review were as follows: (1) original research study using digital or telemedicine approaches for CR and (2) reported results for feasibility, usability, or clinical outcomes. Studies were excluded if they (1) were not full-length publications (ie, abstracts), (2) were methods papers, (3) described only the technology without any inclusion of study participants, or (4) did not include any follow-up time to study outcomes (ie, cross-sectional studies). Given the evolving terminology surrounding digital health technology, we sought external expert inputs to include articles that were not found through our primary search strategy. Papers were included if they reported original research in digital CR. Full details are presented in Figure 1.

For each study, we recorded the components of CR that were delivered as described in the AHA and AACVPR consensus statement on the core components of CR and categorized the digital intervention listed in each study as either standalone or adjunctive to conventional CR. Studies were designated as standalone interventions when the program was delivered remotely with the exception of initial in-person session(s) for onboarding or baseline or outcome assessments, as long as the rest of the intervention was remote. The quality of the articles was independently assessed by 2 evaluators using the National Institutes of Health: National Heart, Lung, and Blood Institute (NHI: NHLBI) Study Quality Assessment Tools, which include the evaluation of 14 criteria for an overall quality assessment of good, fair, or poor (Multimedia Appendices 2 and 3) [11]. Discrepancies in ratings were resolved by discussions between the evaluators to reach a consensus on the ratings. We followed the guidelines outlined in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses [12].

In total, 31 studies met the eligibility criteria and were included in this review (Table 1). The study characteristics are summarized in Table 2 and Multimedia Appendix 4. The median sample size was 98 (IQR 52.5-146), the median intervention duration was 3 months (IQR 1.6-4.4 months), and the median follow-up time was 6 months (IQR 3-6 months). The majority of these studies were conducted in Europe (12/31, 39%) and North America (8/31, 26%). A total of 22 studies (22/31, 71%) were randomized controlled trials. Of the 31 studies, 15 (15/31, 48%) were standalone digital CR interventions. The study quality was variable, with 23 studies (23/31, 74%) having a good quality, 7 (7/31, 23%) having a fair quality, and 1 (1/31, 3%) having a poor quality, according to the criteria established by the NIH: NHLBI Quality Assessment Tools (Multimedia Appendix 5).

As shown in Figure 2, the most commonly targeted CR core components were physical activity counseling (31/31, 100%), baseline assessment (30/31, 97%), and exercise training (27/31, 87%). Only about one-third of the studies addressed each of the other CR core components of nutrition counseling (11/31, 35%), psychological management (11/31, 35%), and weight management (10/31, 32%). In contrast, less than a third of the studies addressed other CR core components, with only a single study including lipid management (1/31, 3%), 2 studies including diabetes management (2/31, 6%), 7 studies including tobacco cessation (7/31, 23%), and 8 studies including blood pressure management (8/31, 26%). Smartphones/mobile devices and wearables were employed in 65% (20/31) of the studies, websites or web portals in 58% (18/31), and email-SMS communications in 35% (11/31) of the studies. The interventions were most commonly guided by physical therapists or exercise specialists (12/31, 39%), followed by CR/research team staff (11/31, 35%), and nurses (10/31, 32%). Four studies (4/31, 13%) described their interventions as fully automated or did not indicate requirement of any specific personnel [13-16]. The most commonly evaluated outcome was exercise capacity or step count (22/31, 71%). Other frequently assessed outcomes included program adherence (14/31, 45%) and QoL (14/31, 45%).

The key findings from the 31 studies are summarized in Textbox 2. Overall, program adherence was greater in patients using digital interventions than in those participating in conventional CR. Moreover, digital CR interventions were comparable to conventional CR (control) groups across multiple short-term outcomes (eg, functional capacity, physical activity, self-efficacy, program adherence, weight management, dietary habits, and QoL). However, digital CR interventions had mixed efficacy with regard to blood pressure control and mood. Of the 9 studies reporting blood pressure as an outcome of digital interventions versus conventional CR, 3 studies showed noninferiority [18,29,41], 4 found no significant impact on blood pressure [19,28,31,39], and 2 reported a better control of blood pressure in the control group as compared to the digital intervention group [25,35]. Similarly, although the majority of studies assessing mood as an outcome reported that digital interventions were noninferior to conventional CR in improving mood [22,25,29,30,33], 4 studies reported no significant improvement in anxiety or depression [19,24,25,42].

We found a paucity of studies specifying intervention components that targeted lipid management, glycemic index control, and smoking cessation. These components were often reported as secondary outcome measures, if at all. Furthermore, the research on the long-term efficacy of digital CR interventions was sparse with heterogeneous findings. Although Devi et al [25] found improvement in outcomes such as QoL, self-efficacy, and physical activity in the short term, no significant intervention effect was present on these outcomes when assessed at the 6-month follow-up. However, they noted that the intervention group demonstrated trends of improved levels of physical activity, whereas the control group did not. Reid et al [22] also reported long-term improvements in self-reported QoL and physical activity as long as 12 months from the start of the digital CR program, which was delivered over a 6-month period.

The examination of the studies by follow-up time revealed that the majority of the key outcome findings were mixed. However, all the studies reporting outcomes regarding adherence [19-21,26-30,35,37,38,40,41,43], QoL [15,17,21,22,24,25,27,29,33,35,38,39,41,42], and exercise capacity [15,17-22,24,25,27-33,35,38-42] found positive results or outcomes that were noninferior to the control group. Only studies with a follow-up period longer than 3 months reported outcomes for physical activity [14,15,22,25,31,32,36,41] and blood glucose control [29,31,35,41]. Regarding physical activity, the results were positive [14,15,22,25,31,32,36,41], whereas the results regarding blood glucose management were mixed, with positive effects [31], nonsignificant effects [29,35], or comparable results between the intervention and the control groups [41]. Similarly, the outcomes for blood pressure, depression or anxiety, and weight management were mixed (with positive effects, nonsignificant effects, or comparable results between the intervention and control groups) in both studies with shorter (3 months or less) [18,24,28,39] and longer (more than 3 months) [19,22,25,29-31,33,35,37,41,42] follow-up times. The only exceptions were as follows: one study reported a negative impact on mood at 6 months [16], 2 studies reported that the control group had better blood pressure management than the intervention group [25,35], and one study found an increase in body weight in the intervention group at 3 months [17]. For studies reporting outcomes regarding dietary habits, lipid management, and self-efficacy, studies with longer than a 3-month follow-up period reported positive or nonsignificant effects [15,16,19,25,29-32,35,37,41] or outcomes that were comparable between the intervention and the control groups [35,41], whereas studies with a follow-up period of 3 months or less found positive results [13,18,24]. The outcomes for nicotine dependence were mixed (positive or nonsignificant effects) among the studies with a short follow-up period [18,24,39], whereas one study reported no impact on smoking at 6 months [19].

The examination of more comprehensive, standalone interventions revealed that only 6 studies included in this review were standalone interventions delivering 5 or more CR components (other than a baseline assessment) [21,23-25,29,32]. The results of these studies for the majority of the key outcomes related to the CR components delivered through the interventions were heterogeneous. Of these 6 more comprehensive standalone interventions, only Brough et al [24] reported the outcomes regarding nicotine dependence, finding no impact on smoking. Devi et al [25] and Brough et al [24] found that their digital interventions had no significant impact on mood, whereas Varnfield et al [29] reported a positive effect on mood and anxiety levels in the intervention groups. Regarding dietary habits, Varnfield et al [29], Lear et al [32], and Brough et al [24] found improvement in the dietary habits of the intervention group, whereas Devi et al [25] found no significant impact. Three of these studies [24,25,29] reported an improvement in QoL in the intervention group, but Piotrowicz et al [21] found no significant difference in QoL measures between the intervention and control groups. Five of these studies reported a positive effect of digital intervention on exercise capacity [21,24,25,29,32]. In terms of weight management, Devi et al [25] and Varnfield et al [29] found a positive impact on weight in the intervention group, whereas Brough et al [24] found no impact on weight in the intervention group. Overall, there was a wide variety in the interventions delivered and outcomes reported.

This study highlights digital technology as a potential means of enhancing care and broadening access to CR through tailored interactive interventions. Our work differs from previous systematic reviews as our emphasis is on digital CR interventions with a focus on providing a systematic evaluation of the current literature to better understand the characteristics of these interventions. This study builds upon a growing body of literature supporting the use of internet-based features such as web portals and digital devices (eg, wearables) to remotely deliver CR components.

We found that digital CR was feasible and as effective as traditional CR in improving outcomes, whether as an adjunct or as an alternative to traditional CR [16,21,26,29,39,41,42]. Our findings support the conclusions of a previous study demonstrating a similar effectiveness of home- and center-based CR in improving clinical and health-related QoL outcomes in patients with myocardial infarction, myocardial revascularization, and heart failure [44]. In addition, Huang et al found that telehealth CR interventions were noninferior to center-based CR, both in the short term (12 weeks-1 year) and long term (up to 6 years), when comparing participants’ exercise capacity, all-cause mortality, and modifiable risk factors, including blood pressure, blood lipids, smoking, and weight [45]. Moreover, the AACVPR, AHA, and ACC recently released a consensus statement highlighting evidence that home- and facility-based CR can achieve similar improvements in 3-12–month clinical outcomes [10]. These developments highlight digital technology as a potential means of enhancing care and broadening access to CR through tailored interactive interventions [46,47]. However, our study and literature indicate that until digital CR is further developed and better understood, there will be a need for in-person CR sessions (Multimedia Appendix 6). In-person sessions may help digital CR by establishing baseline and monitoring progress, personalizing treatment plans, and bridging patient technology-usage challenges through technology education and deployment, especially for older users [48-50]. Currently, there are several ongoing clinical trials that are studying the efficacy of digital CR interventions [51-54].

A limitation of our study is the heterogeneity of the identified papers, thus prohibiting meta-analysis. Papers varied in CR technologies, interventions, study design, measured outcomes, and control groups. For example, when considering a traditional CR population as a control group, some studies used standard care, some used direct comparison with facility-based CR, and other feasibility studies did not have a control comparison group. Although the diversity in studies proved challenging to quantify, it reiterates the motivation of this systematic review: digital health-based CR is emerging as an alternative or adjunct to standard CR; thus, methodologies have yet to reach a consensus. Consequently, we encourage practitioners to study digital CR approaches in a collaborative environment to promote the standardization and optimization of study methods. We have summarized our key recommendations in Textbox 3.

Beyond the heterogeneity of the study designs, we identified other limitations. Although we evaluated CR program components using the AHA and AACVPR’s consensus statement on the core components of CR, the CR programs outside of the United States may differ from professional society guidelines in the United States. In addition, some studies included only minimal details about the specific components of their interventions; thus, the number of CR components delivered may have been greater than that captured in this review. Although studies reported the technology used (summarized in Multimedia Appendix 7), specific details were often sparse, limiting the evaluation of specifically what worked well and what was difficult to implement. Furthermore, the studies did not address long-term outcomes, as the maximum reported follow-up was 16 months, and most studies had fairly small sample sizes. Finally, the majority of the studies reviewed included sessions that were conducted in person for at least a portion of the overall CR intervention.

We identified several potential directions for future research. Although digital interventions have been found to successfully deliver the components of CR pertaining to physical activity, there remains a paucity of comprehensive digital CR program interventions that address risk factors such as lipid management, blood glucose level control, and smoking cessation. In addition, the long-term effectiveness of digital approaches to CR requires further evaluation [8,55]. Furthermore, studies with larger sample sizes and adequate control groups for comparison are necessary to better understand the impact of these interventions. Additional studies are also required to investigate the frequency of adverse events in patients participating in these interventions compared with traditional approaches. The adverse effects reported in the studies reviewed fell into the categories of (1) cardiac related (acute coronary syndrome, stable angina, arrhythmias, pericarditis, dyspnea, syncope, etc), (2) potential cardiac etiology (pleural effusions, cerebral ischemia, etc), (3) noncardiac related (peripheral artery disease, pneumonia, accidents, etc), and (4) death. A number of studies reported no significant difference in the rate of adverse effects between the digital intervention and control groups [14,30,33,38]. There were also studies that found that the intervention group had fewer associated adverse events than the control group; however, this did not always reach statistical significance [19,22,31,32,34]. Only one study reported that the digital CR intervention group had more adverse events during the treatment period as compared with the control group that participated in a center-based CR; however, adverse events were comparable during the postintervention follow-up period [41]. In some cases, no major adverse events occurred throughout the study [17,26-28,40,42], or the adverse effects were unrelated to the study intervention [16,21,35]. No conclusions could be drawn regarding the increased or decreased risk of adverse effects in some studies [15,18,20]. Overall, our findings highlight the need for robust digital intervention study designs with more comprehensive programs and analysis of both the short- and long-term effects.

Furthermore, as briefly mentioned in the Methods section, our team consulted with an external expert to include relevant original research on digital approaches to CR that may have been inadvertently excluded from our primary search strategy. We opted for this approach given that although the field of telehealth and telemedicine has grown rapidly over the past few decades, the adoption of common terminology remains in infancy [7]. The evolution of this terminology has been demonstrated in a bibliometric analysis by Fatehi and Wootton [6], who noted the emergence of terms such as eHealth and mHealth as well as the usage trends of the terms such as telemedicine, telehealth, and eHealth in the literature. As a result, there may be studies that include digital components, such as the usage of the internet, which our search terms failed to capture. Our experience further highlights the importance of more standardized terminology surrounding digital health interventions and an understanding of the evolving terminology to accurately review the existing literature.

Overall, we found that digital technology offers the potential to address the challenges associated with traditional, facility-based CR. If implemented on a large scale, digital CR could provide a level of impact, accessibility, affordability, cost savings, and benefits to patients not possible with conventional CR. However, so far, interventions have primarily focused on physical activity counseling and exercise training and not on the other core components of CR. In addition, our study focused on the evaluation of the technology used in digital CR and the comprehensiveness of these programs, but the intricacies of accreditation for CR programs are beyond the scope of this review. Further research is required with more comprehensive CR interventions to understand the long-term clinical impact of digital CR solutions on key cardiovascular outcomes and establish best practices for the development, delivery, and assessment of digital CR.