Alternative models of cardiac rehabilitation: A systematic review

Abstract

The traditional hospital-based model of cardiac rehabilitation faces substantial challenges, such as cost and accessibility. These challenges have led to the development of alternative models of cardiac rehabilitation in recent years. The aim of this study was to identify and critique evidence for the effectiveness of these alternative models. A total of 22 databases were searched to identify quantitative studies or systematic reviews of quantitative studies regarding the effectiveness of alternative models of cardiac rehabilitation. Included studies were appraised using a Critical Appraisal Skills Programme tool and the National Health and Medical Research Council’s designations for Level of Evidence. The 83 included articles described interventions in the following broad categories of alternative models of care: multifactorial individualized telehealth, internet based, telehealth focused on exercise, telehealth focused on recovery, community- or home-based, and complementary therapies. Multifactorial individualized telehealth and community- or home-based cardiac rehabilitation are effective alternative models of cardiac rehabilitation, as they have produced similar reductions in cardiovascular disease risk factors compared with hospital-based programmes. While further research is required to address the paucity of data available regarding the effectiveness of alternative models of cardiac rehabilitation in rural, remote, and culturally and linguistically diverse populations, our review indicates there is no need to rely on hospital-based strategies alone to deliver effective cardiac rehabilitation. Local healthcare systems should strive to integrate alternative models of cardiac rehabilitation, such as brief telehealth interventions tailored to individual’s risk factor profiles as well as community- or home-based programmes, in order to ensure there are choices available for patients that best fit their needs, risk factor profile, and preferences.

Acute coronary syndrome, cardiac rehabilitation, cardiovascular diseases, coronary disease, internet, myocardial infarction, risk factors, systematic review, telemedicine

Introduction

Hospital-based cardiac rehabilitation (CR) programmes are well-established in the effective management of patients with acute coronary syndrome (ACS).^1,2 These programmes improve survival, quality of life, functional status, and cardiovascular risk profile as well as reduce hospital readmissions and psychological disorders.^2,3

Traditional models of CR generally consist of multiple phases.^4–6 The first phase begins whilst the patient is in hospital and consists of early mobilization and education. It is delivered on an individual basis and in some hospitals to groups of patients. The shorter hospital stay (now commonly 4–6 days after acute myocardial infarction, 5–7 days after coronary bypass surgery, and 1 day after percutaneous coronary intervention) makes it extremely difficult to conduct formal inpatient education programmes.^4–6

Most CR models are based upon supervised ambulatory outpatient programmes. Attendance begins soon after discharge from hospital and ends within 3 months of the acute event. Formal outpatient CR programmes vary widely in content.^4–6 Almost all contain an element of group exercise. Education is usually also delivered. Maintenance follows the ambulatory programme in which physical fitness and risk factor control are supported in a minimally supervised setting. Maintenance programmes are even more varied in content and structure than ambulatory programmes.

Although CR is an evidence-based form of secondary prevention,^{2  3} referral is suboptimal and participation rates in Australia, the USA, and Europe are low, estimated at 10–30%.^7–9 Barriers to accessing traditional CR include transport difficulties, work schedules, social commitments, lack of perceived need, and functional impairment.^10–12 Traditional CR faces substantial challenges in terms of cost and access and does not meet the needs of the majority that require secondary prevention or those patient groups most in need of risk factor reduction, such as older adults, women, ethnic groups, and low-income populations.^10–13

These challenges have led to the development of a large and diverse array of alternative models of CR.¹⁴ These programmes involve (in isolation or combination) in-person visitations, community services or home manuals with phone/electronic support for flexible and individualized management.¹⁴ Other alternatives include nurse-coordinated care, case management, telemonitoring with periodical follow up, and community-based groups with ongoing health practitioner support provided across a range of settings.¹⁴

The aim of this study was to identify and critique the evidence for the effectiveness of alternative models of CR. The review addressed three key research questions: (1) what are the different models of care for CR other than traditional hospital-based rehabilitation? (2) what are the core elements of each model of care? and (3) what are the clinical and service outcomes of each model of care?

Methods

This systematic review is an update of an unpublished review commissioned by the Heart Foundation.¹⁵ Methodologies recommended in the Joanna Briggs Institute (JBI) Manual underpinned the process of conducting and reporting this review.^16,17

Search strategy

A comprehensive electronic database search was undertaken, using keywords relating to all questions. The database searches were supplemented by pearling and manual searching of reference lists. Published literature that used a quantitative design (e.g. randomized controlled trials, RCTs, or other experimental designs, observational studies, case studies) and systematic reviews were included. Independent reviewers undertook investigation of all data sources.

The search was limited to articles published after 1999 as alternative models for CR are relatively new. Articles were included if written in full-text in English. Literature only available in abstract form was excluded, along with duplicate articles. The specific inclusion criteria for this review are outlined in Table 1.¹⁶

Table 1.

Inclusion criteria

P	Adult patients eligible for cardiac rehabilitation following an acute episode of care (e.g. post-myocardial infarction, new onset angina, post-coronary artery bypass graft, post-angioplasty)
I	Any model of care other than traditional hospital-based cardiac rehabilitation
C	Traditional hospital-based approach, a model of care compared with another model, no comparison
O	Service outcomes (e.g. drop-out rates, enrolment, provider and patient satisfaction) and clinical outcomes (angina, quality of life, functional status)
T	Short- and/or long-term

Key terms

The search used Boolean operators to combine free text terms and MeSH terms including: post myocardial infarction, cardiac rehabilitation OR rehabilitation, telehealth, angina, CABG (or post coronary artery bypass graft), non-traditional model, angioplasty, community-based programmes OR community-based health programmes OR community health programmes, post acute cardiac OR cardiac, self-directed programme, walking programme, GP-led or nurse-led programme and GP-led or nurse-led programme.

Information sources

The following 22 databases and sites were searched: Ageline, Academic Search Elite, AMED, Biomed Central Gateway, CINAHL database, CAM on Pubmed, Cochrane Library, Current Contents Connect, Digital Dissertations, EMBASE, Health Source Nursing/Academic Edition, Informit E-Library, Journal Citation Reports, MEDLINE, PubMed, PubMed Central, Psych INFO, Science Direct, Scopus, Web of Knowledge, PsycARTICLES Direct, and HighWire Press.

Study selection

Titles and abstracts were initially screened to eliminate obvious irrelevance. Potentially eligible publications were retrieved in full text. In the event that more than one publication had analysed the same data set, only the primary report was included. Two reviewers independently selected studies for inclusion and a third independent reviewer arbitrated.

Data extraction

Data was extracted from the identified publication using a tool specifically developed for this review based on JBI guidelines.¹⁶ Domains included: author and year of publication, country of origin, characteristics of participants, total population, geographical location, numbers, intervention/model of care (core component description, setting, staff delivering intervention, duration) comparison group, outcome measures, and results.¹⁶ A third reviewer checked the data when uncertainties were encountered.

Quality and risk of bias in individual studies

Included studies were appraised by two independent reviewers using the Critical Appraisal Skills Programme (CASP) tools created to appraise each of the study designed included. The CASP provides a range of critical appraisal tools to assess the methodological rigor of different research designs and reflect questions believed to be most important to healthcare providers.¹⁸

Data analysis

All studies were initially reviewed and grouped by the model of care that was used to deliver CR. As this systematic review included a large heterogeneous group of study designs and sources, the results and key information obtained from each of the related articles was synthesized in a narrative summary.¹⁶

Strength of the body of evidence

The National Health and Medical Research Council (NHMRC) levels of evidence scheme was used. These levels assess the validity of recommendations for clinical guidelines and focuses on the effectiveness of treatment.

Results

A total of 534 titles and abstracts were initially screened for relevance to the review questions. Of these, 111 full text articles were retrieved. It was determined that 83 articles fit the inclusion and exclusion criteria set for this review (Figure 1).

Figure 1.

Search results.

The articles included described interventions in the following broad categories of alternative models of CR (Table 2).

1. Multifactorial individualized telehealth delivery: These programmes addressed multiple risk factors and provided individualized assessment and risk factor modification, with the majority of patient provider contact delivered by telephone.

2. Internet-based delivery: These programmes delivered the majority of patient–provider contact for risk factor modification via the internet.

3. Telehealth interventions focused on exercise: These programmes focused on exercise, with the majority of patient–provider contact delivered by telephone. These studies often included the use of telemonitoring.

4. Telehealth interventions focused on recovery: The majority of patient–provider contact was delivered by telephone and the intervention content focused on supporting psychosocial recovery from an acute cardiac event such as myocardial infarction (MI) or coronary artery bypass graft (CABG) surgery.

5. Community- or home-based CR: The majority of patient–provider contact was delivered face-to-face, through either home visits or patient attendance at community centres (for programmes other than traditional CR).

6. Programmes specific to rural, remote, and culturally and linguistically diverse populations.

7. Multiple models of care: This group of studies addressed or reviewed multifaceted interventions across a number of these categories.

8. Complementary and alternative medicine interventions.

Table 2.

Models of care and characteristics of included studies

Publication	Level of evidence	CASP score	Country	Participants	Intervention	Outcomes
Multifactorial individualized telehealth delivery
Neubeck et al.19	II	10	Australia	3145 patients CHD	Telehealth (SR of 11 trials)	Non-significant lower all-cause mortality (RR 0.70, 95% CI 0.45–1.1)
Excluded heart failure patients severe comorbid disease	Significant favourable changes in TC (MD –0.37, 95% CI –0.56 to –0.19), HDL (MD 0.05, 95% CI 0.01 to 0.09), systolic BP (MD –4.69, 95% CI –6.47 to –2.91), smoking status (RR 0.83, 95% CI 0.7 to 0.99) were observed in meta-analysis of trials that compared telehealth interventions and usual care
74% of study participants were men
Mean age 61
Redfern et al.21	II	9	Australia	ACS patients not accessing CR	CHOICE programme	From baseline to 12 months:
Conventional care n = 72	1 hour initial consultation	Significantly reduced TC (mean ± SEM 4.0 ± 0.1 vs. 4.7 ± 0.1 mmol/l, p < 0.001), systolic BP 131.6 ± 1.8 vs. 143.9 ± 2.3 mmHg, p < 0.001), BMI 28.9 ± 0.7 vs. 31.2 ± 0.7 kg/m2, p = 0.025)
Modular care n = 72	4×10 min phone conversations over a 3-month period	Improved PA (1369.1 ± 167.2 vs. 715.1 ± 103.5 METS/kg/min, p = 0.001)
Inclusion: within 6 months of ACS diagnosis	Risk factor assessment	Fewer patients smoking
Patient selected strategy from options	Fewer CHOICE participants (21%) had three or more risk factors above recommended levels than controls (72%, p < 0.001)
Goal setting
Printed information
Redfern et al.20	II	8	Australia	ACS patients not accessing CR	CHOICE programme	From baseline to 3 months:
Conventional care n = 72	As per Redfern et al.21	Significantly greater reductions in TC (158 ± 3.9 vs. 186 ± 3.9 mg/dl, p < 0.001), systolic BP (133.5 ± 2.0 vs. 144.4 ± 2.4 mmHg, p < 0.01), BMI (28.9 ± 0.7 vs. 31.0 ± 0.7 kg/m2, p = 0.02), and PA (1,187 ± 164 vs. 636 ± 115 METS/kg/min, p < 0.01)
Modular care n = 72	Fewer patients in moderate–high risk LIPID score (40% vs. 59%, p = 0.02)
Inclusion: within 6 months of ACS diagnosis	Fewer patients smoking (6% vs. 23%, p < 0.01)
Vale et al.23	II	9	Australia	Intervention n = 398	COACH programme	From baseline to 6 months:
Usual care control n = 394	5 calls over 6-month period (length of calls determined by individual)	Significantly greater reductions in TC (mean change 21 mg/dl (0.54 mmol/l) (95% CI 16–25 mg/dl 0.42–0.65 mmol/l) in the COACH programme vs. 7 mg/dl (0.18 mmol/l) (95% CI, 3–11 mg/dl 0.07–0.29 mmol/l)
Included CABG, PCI, acute MI, or unstable angina and then discharged on medical therapy, coronary angiography with planned (elective) revascularization	Structured delivery of calls	Also, reduction in body weight, BMI, dietary intake of total fat, saturated fat, cholesterol, and anxiety level
77% were men, median age of 58.5 years	Goal setting
Individual risk factor modification
Patient could seek additional phone support
Vale et al.22	II	9	Australia	245 patients	COACH programme	From baseline to 6 months
Coaching by telephone n = 121	As per Vale et al.23	Significantly lower TC (mean 5.00, 95% CI 4.82–5.17 mmol/l vs. 5.54, 5.36–5.72 mmol/l, p < 0.0001) and LDL-C (3.11, 95% CI 2.94–3.29 mmol/l vs. 3.57, 95% CI 3.39–3.75 mmol/l, p < 0.0004)
Usual care n = 124	Coaching had no impact on TG or on HDL-C levels
Included CABG surgery, PCI	Multivariate analysis showed that being coached (p < 0.001) had an effect of equal magnitude to being prescribed lipid-lowering drug therapy (p < 0.001)
Jelinek et al.24	IV	11	Australia	656 patients, intervention only	Follow up every 6 months to 2 years	TC, target waist circumference, smoking, and PA improvements at 6 months were sustained at 24 months
45% post CABG, 48% post PCI,5% after ACS, 2% elective revascularization	COACH programme
80% men, median age 61 years	As per Vale et al.23
Fernandez et al.28	II	7	Australia	51 participants, mean age 57 years, 78% male	HeLM (health related lifestyle management system): 6-week intervention	Reduced systolic BP (120.3 ± 16.3 vs. 126.4 ± 14.6)
Inclusion: diagnosis of ACS and one or more modifiable risk factors	Three calls over 8 weeks	High levels of satisfaction with intervention
Excluded if: major comorbidity	Goal setting
Printed material
Individual risk factor modification
Hanssen et al.27	II	9	Norway	288 post MI patients (Intervention n = 156)	Weekly nurse initiated calls for first 4 weeks, then contact at 4, 8, 12, 24 weeks after discharge	At 18-month subgroup analysis:
Inclusion: diagnosis of acute MI	Individualized risk factor information	Significant difference only in physical component of SF-36 in patients over 70 years (p < 0.05)
Exclusion: coexisting severe chronic disease, CABG	Goal setting	Overall, no long-term effects despite positive short-term effects
Structured delivery of calls
Patient could seek additional phone support
Hanssen et al.26	II	9	Norway	288 post MI patients	Results at 6 months	Difference in self-reported physical health measured with SF-36 favouring intervention at 6 months (p = 0.034)
Intervention n = 156	As per Hanssen et al.27	No difference in mental health
Inclusion: diagnosis of acute MI	More participants stopped smoking in intervention group (p = 0.055)
Exclusion: coexisting severe chronic disease, CABG	Frequency of PA higher in intervention group (p = 0.004)
Hawkes et al.30	RCT protocol	n/a	Australia	Protocol only n = 550	Delivered by health ‘coaches’	SF-36, PA, and cost-effectiveness
Up to 10 × 30 minute scripted telephone sessions
Handbook for patients and educational resource to use during sessions
Woodend et al.56	II	4	USA	Intervention HF n = 62	3 months of videoconferencing with a nurse, daily transmission of weight and BP, and periodic transmission of ECG	For the patients with angina there was a 51% reduction in the number of admissions per patient with angina receiving telehome monitoring compared with those receiving usual care (p = 0.02)
Intervention angina n = 62	There was also a 61% reduction in the number of days spent in the hospital (p = 0.04)
Control HF n = 59
Control angina n = 66
Roth et al.57	III–2	8	Israel	MI patients	Telemonitoring involved a call-centre that had access to each subscriber’s medical file. Each subscriber carried a ‘cardiobeeper’ to transmit a 3 - or 12-lead ECG. Staff uses protocols for referral. Also, calls are initiated periodically for reassurance and encouragement	Mortality at 1 year was less in the telemonitoring cohort (4.4% vs. 9.7%, p = <,0.001)
Telemonitoring n = 699	Adherence to physical exercise guidelines
Control n = 3899	Cost-effectiveness will be examined as well
Not randomized
Walters et al.58	RCT protocol	n/a	Australia	Non-inferiority trial with 100 patients enrolled in intervention and 100 patients in traditional CR	Intervention involves mobile phones with a built-in accelerometer to measure exercise and WellnessDiary software to collect information on physiological risk factors
Video and teleconferencing are used for mentoring sessions aimed at behavioural modification
Mentors use a web-portal
Educational content is stored or transferred via message systems to patients’ phones
Neubeck et al.25	II	9	Australia	ACS patients not accessing CR	CHOICE programme	4-year outcomes
Conventional care n = 72	As per Redfern et al.21	1-year improvement in risk factors were maintained at 4 years
Modular care n = 72	TC (4.0 ± 0.1 vs. 4.2 ± 0.1 mmol/l, p = 0.05), systolic BP (132.2 ± 2.1 vs. 136.8 ± 2.0 mmHg, p = 0.01), and PA scores (1200 ± 209 vs. 968 ± 196 METs/week, p = 0.02)
Inclusion: within 6 months of ACS diagnosis	Proportion with three or more risk factors above national targets (20% vs. 42%, p = 0.02)
Holmes-Rovner et al.59	II	8	USA	ACS	6 sessions of telephone coaching delivered by a health educator during first 3 months after discharge	Greater self-reported PA: OR 1.53, p = 0.01), however, there was a decline after the intervention period
Intervention n = 268	Behaviour goals included: reduction or elimination of smoking, increasing PA, and eating a healthier diet	Smoking cessation, medication adherence, functional status and QoL were similar between groups at 3 and 8 months
Control n = 257	Tailored to individual goals
Hailey et al.60	SR no meta-analysis	8	Australia	Systematic review of telerehabilitation: 16 studies	Telerehabilitation	In 10 studies, TR was described as successful for home rehabilitation
Success was unclear in studies on increasing participation and on the feasibility of internet-based CR
Success was not demonstrated in studies on ‘booster intervention’, ‘psychosocial adjustment’, ‘symptom management’, ‘self-efficacy’, and ‘PA’
Mittag et al.61	II	4	Germany	Intervention n = 171	Intervention involved monthly telephone calls after 3 weeks of inpatient CR	After 12 months, patients in the intervention group ha statistically significantly, p < 0.05) lower Framingham risk scores compared with controls
Control n = 172
Internet-based delivery of cardiac rehabilitation
Southard et al.31	II	6	USA	CVD patients	Logging on 1 × 30 min/week over 6 months	Fewer CV events in intervention subjects (15.7%) than among the UC subjects (4.1%) (p = 0.053), resulting in cost savings of $1418 per patient
Intervention n = 53	Risk factor information	No statistically significant (p > 0.05) difference between groups in other risk factors measured (depression, lipids, BP, exercise, diet)
Control n = 51	Completing education modules	More weight lost in intervention group (–3.68 pounds) than in the UC group (+0.47 pounds) (p = 0.003)
Average age was 62 years (range 37–86 years)	Group discussion
Predominantly married, male, and white	Patient could seek additional support via email
Incentives
Zutz A et al.32	II	3	Canada	Internet-based	12 weeks	HDL-C, TG, TC/HDL-C ratio, exercise capacity (METs), weekly PA, and exercise specific self-efficacy were measured
Observational n = 8	Equipment intensive	No statistically significant (p > 0.05) between-group differences but intervention group had reductions in risk factors similar to reported traditional CR
Control n = 7	Exercise focused
Exclusions: depressed, smoked, had an abnormal resting ECG, had > 2 mm ST-depression or a significant arrhythmia on their exercise test ECG, or had uncontrolled diabetes were excluded	Group participation
Individual risk factor modification
Goal setting
Patient could seek additional support via email
Data entry
Electronic data transmission
Moore et al.33	III–1	6	USA	E-CHANGE group n = 7	3-month programme	6 hours more exercise in first 2 months
Traditional CR n = 18	Exercise focused	Exercise intensity was higher in the intervention group (49.8 ± 21.4 vs. 13.4 ± 17.9 METS)
Inclusion: MI, CABG surgery, and/or angioplasty, no clinical features that constitute high risk for safe participation in cardiac exercise programmes	Data entry	Energy expenditure due to moderate to high-intensity exercise was 3-times higher in the intervention group
Risk factor information
Patient initiated email communication
Education modules
Brennan et al.62	Protocol RCT	n/a	USA	Protocol of RCT reported	Internet-based information and support system for patient home recovery after CABG	Physical, social, and functional status, mood, family function, and cardiac risk factor modification
CABG n = 140
Leemrijse et al.63	Protocol RCT	n/a	Holland	Attempting to recruit 200 participants for each group	Six-months duration with contact every 4–6 weeks by telephone	Primary: BMI, waist circumference, PA, BP, TC, LDL, HDL, diet
Coached to take responsibility for achievement and maintenance of defined target levels for their individual modifiable risk factors	Secondary: blood glucose, HbA1C, self-management, smoking, medication adherence, QoL, depression, anxiety
Devi et al.64	Protocol	n/a	UK	Systematic review protocol	All internet-based interventions to promote healthy lifestyles and medicines management and reduce cardiovascular risk in patients with CHD	Clinical outcomes, cardiovascular risk factors, and HR-QOL
Clark et al.65	Pilot (IV)	n/a	Australia	24 CR participants	Internet-based self-management system that provided the case-manager with the ability to deliver education, track patient progress and have contact with the patient and carer via email and discussion boards or by telephone	11 (46%) completed the programme
Mean number of risk factors monitored was 5.
100% viewed at least one library article
33% completed all online workbooks
63% used message system
Moderate satisfaction to the usability and utility of the features of the website
All case-managers used the message system
Telehealth exercise-focused interventions
Scalvini et al.35	IV	8	Italy	47 patients who underwent CABG and/or valve replacement	Home-based exercise rehabilitation with telemedicine 15–28 days	Significant increase in 6MWT, p < 0.05)
Exclusions: insulin-dependent diabetes and/or overt chronic respiratory insufficiency	Exercise focused	Patient satisfaction 95%
Daily calls
Unscheduled phone support
Information session
Electronic data transmission
Giallauria et al.66	III–2	3	Italy	45 male	8-weeks home-based with telecardiology monitoring	Improvements in cardiovascular functional capacity, anxiety, and depression in intervention group, p < 0.05)
Intervention n = 15	Exercise focused
Patient exclusions: heart failure, residual myocardial ischaemia, severe ventricular arrhythmias, atrial fibrillation, disability, or cognitive compromise	Electronic data transmission
Körtke et al.34	III–2	3	Germany	170 cardiac surgery patients (CABG, valve replacement/reconstruction)	3 months, 3 × 30 min/week exercise	Non- randomized groups
Intervention n = 70	Electronic data transmission	Comparable improvements over time with hospital-based group, but no between-group comparison
< 60 years of age	Patient-initiated phone support
92% were male	Risk factor information
Exercise focus
Ades et al.37	III–2	3	USA	Patients within 3 months of an acute coronary event, acute MI, coronary angioplasty, CABG, cardiac transplantation (83 home programme, 50 centre based)	3 months, exercised 3 times per week	Patients in the home-based monitoring programme increased peak aerobic capacity to a similar degree as patients who exercised on site (18% vs. 23%)
Excluded if primary exertional arrhythmias or very low threshold angina	Exercise focused
Scheduled calls
Electronic data transmission
Group participation
Education programme
Dalleck et al.67	II	5	NZ	CR eligible	Telemedicine-delivered exercise and cardiologist appointments for 3 months	No significant difference, p > 0.05) in attendance, BP, cholesterol, BMI, energy expenditure, or stress between groups
Adults, MI/CABG/PCI/valve, English speaking
Intervention n = 53
Control n = 173
Antypas et al.68	RCT protocol	n/a	Norway	Study protocol of RCT	CR participants who have home internet access and mobile phone. All participants have access to a website with information regarding CR, an online discussion forum and an online activity calendar	Primary outcomes is PA over 1 year
Intervention participants also receive tailored content based on models of health behaviour through the website and text messages
Alsaleh et al.69	RCT protocol	n/a	Jordan	Protocol for RCT	Intervention involves behavioural strategies delivered to patients through individualized consultation in which participants are encouraged to set personal goals ad implement self-monitoring in addition to providing them with feedback	PA measured by the International PA Questionnaire
Clinically stable patients able to perform PA	Goals are short-term (one month) and long-term (6 months)	Secondary outcomes include BP, BMI, self-efficacy for PA (using exercise self-efficacy scale), and QoL (using Mac-New Heart Disease HR-QoL Questionnaire)
Feedback is provided through telephone calls
Periodic ‘reminder’ text messages are sent at pre-determined time intervals
Furber et al.38	II	10	Australia	CR patients	Based on social-cognitive theory and focused on increasing self-effiicacy, increasing beliefs and establishing PA goals. Intervention comprised a pedometer, a step-calendar for self-monitoring and telephone support which included goal setting and behavioural reinforcement.	After 6 weeks, improvements in PA were significantly greater, p < 0.05) in the intervention group and remained significant at 6 months.
Intervention n = 109
Control n = 113
Korzeniowska-Kubacka et al.70	III–1	4	Poland	Males after MI with preserved EF n = 62	Intervention involved exercising at home while being monitored with TeleECG	There were no significant differences, p > 0.05) between groups in workload, exercise duration, or physiological variables (BP, HR)
Intervention n = 30
Butler et al.71	II	9	Australia	Intervention n = 62	As per Furber et al.38	At 6 weeks and 6 months there were significantly greater improvements, p < 0.05) in the intervention group for PA
Telehealth recovery-focused interventions
Gallagher et al.39	II	6	Australia	Hospitalized for coronary heart disease: MI, CABG, PCI, angina	Introductory session 1–2 days before discharge and four telephone calls at 2–3 days, 1, 3, and 6 weeks post discharge	No statistically significant (p > 0.05) effect on psychosocial adjustment, anxiety, or depression
Intervention n = 93	Session content was individually tailored and included evaluation of physical and psychological status to incorporate mutual goal setting for self-management of symptoms, diet, exercise, smoking, medications, and stress response
Control n = 103
Colella40	II	3	Canada	Male CABG patients n = 61	Weekly telephone calls from a peer volunteer over 6 weeks (peer-support intervention)	6 - and 12-week follow up
Control n = 124	No statistically significant (p > 0.05) difference in depression or perceived social support
Significantly, p < 0.05) less utilization of health services by the intervention group
Barnason et al.36	II	6	USA	> 65 years after CABG n = 280	6-week (42 daily sessions)	No statistically significant (p > 0.05) differences between control and intervention group in PA and functioning or healthcare use
83% men, 86% married	Tailored information delivered via ‘Health Buddy’ device
Risk factor strategies
Unscheduled contact
Community and home-based cardiac rehabilitation
Jolly et al.72	II	11	UK	525 patients with MI, PTCA, or CABG	Heart manual delivered by nurses with additional 2-day training manual based on Health Belief Model/cognitive behavioural strategies.	No statistically significant (p > 0.05) differences in any measure between centre and home-based
Centre-based n = 262	6 weeks.	Lack of motivation to continue exercising at home centre
Home-based n = 263	Education, home-based exercise, tape-based relaxation and stress management	Direct rehabilitation costs compared: home-based programme had greater direct costs (by 25%) but when added patient costs (travel costs and travel time) there was no difference
‘Low to moderate risk’	Three home visits by nurse at 7–10 days, 6 weeks, 12 weeks (Punjabi speaking nurse where indicated)	No data available on costs in relation to hospital readmission, healthcare utilization
Telephoned at 3 weeks
Punjabi taped version of manual available
Collins et al.73	IV	11	Australia	Cost analysis study	Gym-based programme provided collaboratively by private and public organizations, and directed by a multidisciplinary team of health professionals	The average cost per patient rehabilitated was $1933 in the gym-based programme and $1169 in the home-based programme. Modest evidence of the effectiveness of either programme
People who were hospitalized after a cardiac event	The Ipswich and West Moreton Division of General Practice manage the programme and provide the majority of funds
There are three separate components of the gym-based programme: the outpatient (phase II) gym programme involving four 75 minute sessions per week, ongoing maintenance (phase Ill) gym, and walking programmes of various timeframes, these components are offered upon completion of the inpatient (phase I) programme conducted in hospital and are subject to medical assessment
Moore et al.74	IV	9	USA	n = 8 low-risk cardiac patients	CHANGE intervention	Statistically significant increase (p < 0.05) in exercise participation compared with patients who received usual teaching
5×1.5 hour small group psycho-educative sessions targeting exercise behaviour
Includes problem solving, social reinforcement, self-efficacy enhancement, relapse prevention, social support
Held weekly during last 3 weeks of trad CR programme, then at 1 and 2 months post programme completion
Gordon et al.75	II	4	USA	CAD n = 155	Good description of cardiac risk for inclusion/ exclusion	No statistically significant (p > 0.05) difference in BP, cholesterol, weight, or VO2 max between groups
Good description of cardiac programme components
Randomly assigned to 12 weeks of participation in a contemporary phase 2 cardiac rehabilitation programme (control), a physician supervised, nurse–case-managed cardiovascular risk reduction programme or a community-based cardiovascular risk reduction programme administered by exercise physiologists guided by a computerized participant management system based on national clinical guidelines
Warrington et al.76	IV	3	Australia	Pre-post experimental study	Low–medium risk cardiac patients post hospitalization n = 40	Home-based rehabilitation programme of four community nursing contacts over a 9-week period primarily aimed at individual patient education and carer support
Moderate evidence of effectiveness: statistically significant (p < 0.05) positive changes were found for measures of QoL, knowledge of angina, and exercise tolerance
Carroll et al.77	II	10	USA	RCT	Community-based collaborative peer advisor/advanced practice nurse intervention	Fewer rehospitalizations between 3 and 6 months after MI and CABG in the treatment group compared with the standard care group (Z = −3.72, p < 0.0005)
Low cardiac risk unpartnered older adults n = 247	Intervention included community-based intervention of a home visit within 72 hours and telephone calls at 2, 6, and 10 weeks from an advanced practice nurse and 12 weekly telephone calls from a peer advisor. The intervention was standard care plus the treatment for MI and CABG	A significant increase in cardiac rehabilitation participation over time (Z = 7.60, p < 0.0005)
Dalal et al.78	II	8	UK	n = 230	Comparing home-based rehabilitation (Heart Manual) with hospital-based rehabilitation after MI and to determine whether patient choice affects clinical outcomes	At 9-month follow up, no difference was seen in the change in mean depression scores between the randomized home- and hospital-based groups (MD, 0, 95% CI –1.12 to 1.12) nor mean anxiety score (−0.07, 95% CI −1.42 to 1.28), mean global MacNew score (0.14, 95% CI –0.35 to 0.62), and mean TC levels (−0.18, 95% CI −0.62 to 0.27)
Randomized n = 104	Patients could consent to randomization, or have choice between home- and hospital-based CR
Choice of arm in which to participate n = 126	Both programmes provided simple explanations about coronary heart disease, secondary prevention and stress management and similar types of aerobic exercise featured in the two interventions
Low-risk-confirmed acute MI (WHO criteria)
Ability to read English and registered with general practitioner in one of two primary care trusts
Canyon et al.79	IV	3	Australia	Comparative study (concurrent controls)	HeartBeat, a comprehensive community-based outpatient cardiac rehabilitation service. Sessions are organized by qualified health professionals and counsellors. General practitioners provide feedback to programme organizers, select suitable patients, encourage patient attendance in the programme, provide support with patient medication management and monitor and treat patient CVD risk factors. The HeartBeat programme involved one session per week for 7 weeks. The sessions involved 1 hour of exercise (walking, circuit stretches, trunk stability, posture exercises and muscle strengthening for suitable patients) followed by 1 hour of intensive education on heart conditions and how to address lifestyle risk factors for heart disease. The education sessions included input from dieticians, pharmacists, exercise physiologists, peer educators, and counsellors from the National Heart Foundation	Patients who entered cardiac rehabilitation programme were admitted to hospital less often, and spent less time in hospital
Intervention group who attended ≥ 5 (of 7) sessions n = 110
Control n = 198
Jackson et al.80	SR no meta-analysis	5	UK	SR, 16 studies	Whether self-help groups address the challenges of CHD rehabilitation and self-management	Due to the poor quality studies, the authors concluded that there was no strong evidence of effectiveness of the interventions
Aldana et al.81	III–1	3	Sweden	MI then CABG/PCI patients	Intense lifestyle modification combined with cardiac rehabilitation. Involved stress management with yoga, and specific diet requirements	3-month adherence 89%
Self-selected, not randomized and matched by income stratification	Stage 1: 12 weeks, 1st week info sessions, then 2×per week exercise, group support, stress management, and group meal with diet modifications	Rehab group adherence 85%
Intervention n = 28	Stage 2: 3–9 months, self-directed but encouraged to attend weekly support groups	19/28 adhered to diet at 6 months
Normal CR n = 28	Stage 3: Self-selected participation in monthly alumni meetings	Intervention group had greater improvement in all outcomes except physical function, and role physical.
Control n = 28	Greater reduction in stress, mental health, vitality, and social function
Robertson and Kayhko82	II	4	Finland	First-time MI	Intervention received four visits over 4 weeks	Intervention produced savings of $5716 due to reduced hospitalizations
Home-based intervention n = 32	Nurse-coordinated multidisciplinary, family-based CVD prevention programme (EUROACTION)
Wood et al.83	II	9	Europe	Hospitalized patients with CHD:	Involved eight sessions over 8 weeks including multidisciplinary assessment then group session and exercise	Statistically significantly (p < 0.05) reduced smoking and consumption of saturated fats, increased fruit and vegetable consumption, and improved evidence-based prescribing
Intervention n = 1589	No exercise for ‘high-risk’ patients	Difference in change from baseline of TC statistically significantly (p < 0.05) greater in intervention group
Control n = 1499	Focused on smoking cessation, diet improvement, promoting exercise, PA diary, pedometer)
High-risk CHD:	16 weeks in total
Intervention n = 1189
Control n = 1128
Wolkanin-Bartnik et al.84	II	3	Poland	MI patients n = 186	Minimal education intervention to improve home-based exercise	Statistically significant differences (p < 0.05) in exercise test responses and leisure time activity
Similar improvement in CHD risk factors in both groups
Wang et al.85	II	4	China	Intervention n = 68	Home-based CR with self-help manual	Follow up at 3 weeks, 3 months, and 6 months.
Usual care n = 65	Intervention group higher scores on four of the eight dimensions of the Chinese MIDAS, lower scores for anxiety but not depression (p < 0.05)
Robinson et al.86	II	5	UK	Normal CR n = 54	Low-risk patients ‘fast-tracked’ to phase IV community exercise	No statistically significantly (p > 0.05) differences between groups in incremental shuttle walking test distance or HR-QoL
Intervention n = 46	One exercise session a week was supervised and participants were asked to perform four more sessions
Smith et al.87	II	8	USA	Hospital CR n = 100	6 months of home-based exercise training for 30–50 minutes, 3 times a week at intensity 60–80% target heart rate	6-year follow up
Home-based exercise n = 98	Primarily walking but tailored to include any exercise equipment available	Statistically significant between-group differences (p < 0.05) in peak VO2 and habitual PA as assessed by PASE scores
Poortaghi et al.88	II	5	Saudi Arabia	Intervention n = 40	Home-based	Statistically significant difference (p < 0.05) in general health (GHQ-28) at 2-month follow up
Control n = 40	After end of first and second month a nurse visited the patient at home to provide education, discussed problems and practical training on measuring heart rate, detecting target heart rate and doing suitable exercises
Post CABG, MI, PCI
Matched on age, gender, comorbidities, severity of illness
Oerkild et al.89	II	7	Denmark	Intervention n = 36	Home-based rehabilitation focussing on exercise	Primary outcome was exercise capacity (VO2 and 6MWT)
Control n = 39	Physio made home visits twice with 6-week interval in order to develop a home training programme. A telephone call was made between the home visits	No statistically significant (p < 0.05) between-group differences
Mutwali et al.90	II	7	Saudi Arabia	Home-based CR n = 28	Intervention involved education before CABG, further education in hospital after CABG, then asked to walk unaided at comfortable pace for 30 mins a day for 6 months	Significantly greater (p < 0.05) improvement in HR-QoL, fasting blood glucose, TG, HDL (physical functioning, and anxiety and depression as measured with HADS
Hospital-based CR n = 21	Telephone contact was maintained every 3 weeks. Further education sessions were given including group sessions
For CABG patients
Moholdt et al.91	II	4	Norway	Home-based aerobic interval training n = 14	3×per week for 6 months. Warm up for 10 minutes, followed by four intervals of 4 minutes of high-intensity exercise, should breathe heavily with heart rate 85–95% of max, could be walking, jogging, swimming, or cycling	No statistically significant difference (p > 0.05) difference between groups in peak VO2
Hospital-based CR n = 16	No statistically significant difference (p > 0.05) in QoL
Intervention group reported good adherence to exercise
Houle et al.92	II	4	Canada	Intervention n = 32	Pedometer-based intervention to improve PA over 1 year	Increase in PA in intervention group at 3 months (p < 0.05)
Control n = 33	Also involved individualized education, face-to-face follow up	At 12-month, interaction effects (group × time) in PA and waist circumference were different between groups (p < 0.05), whereas self-efficacy expectation increased in both groups similarly (p < 0.05)
ACS patients	Asked to target 3000 steps per day
Dolansky et al.93	II	4	USA	Intervention n = 17	Intervention aimed between hospital discharge and outpatient CR.Consists of self-management instruction and exercise monitoring including:	At discharge, intervention group trended towards improved exercise self-efficacy (χ2 39.1 ± 7.4) than the control group (χ2 34.5 ± 7.0; t-test 1.9, p = 0.06)
Control n = 21	two 30-minute education sessions with family	Intervention participants had greater attendance at outpatient CR (33% compared to 11.8%; F = 7.1, p = 0.03) and trended toward increased steps walked in first week (χ2 1307 ± 652 compared to χ2 782 ± 544; t-test 1.8, p = 0.07)
For discharged patients in ‘skilled nursing facility’ or ‘home health care’ after ‘cardiac event’	daily walking programme (supervised and BP and heart rate taken before and after)
Chase et al.94	SR no meta-analysis	8	USA	Systematic review of 14 studies of interventions to improve PA	Cognitive interventions: self-efficacy enhancement, barrier management and problem-solving	Authors concluded that cognitive intervention studies reported inconsistent outcomes while behavioural interventions reported more consistent, positive findings
Behaviour interventions: self-monitoring, prompting, goal setting and feedback
Blair et al.41	SR no meta-analysis	8	Scotland	Systematic review of home- or community-based CR	22 studies	Authors concluded that there was little difference between hospital- and home-based CR in terms of reduced mortality or cardiovascular event rates
Eight compared home-based with hospital-based and the remaining compared home rehabilitation with a control group (which varied from hospital-based CR to ‘usual’ or ‘standard’ care
Young et al.95	II	6	Canada	Intervention n = 71	Intervention was a ‘disease management programme’, which involved:	Readmission days for angina, CHF, and COPD per 1000-day follow up were significantly higher in the usual care group (IDR 1.59, 95% CI 1.27–2.00, p < 0.001)
Control n = 75	A nursing checklist	All-cause readmission was also significantly higher (IDR 1.53, 95% CI 1.37–1.71, p < 0.001)
A referral criteria for specialty care
Discharge summary to family physician
Minimum 6 visits from nurse for education
Izawa et al.96	II	4	Japan	Intervention n = 52	Intervention involved using a pedometer in the period between hospital discharge and attendance at CR	Mean self-efficacy for PA score (90.5 vs. 72.7 points, p < 0.001) and mean objective PA (10,458.7 vs. 6922.5 steps/week, p < 0.001) at 12 months after MI onset were significantly higher than the control group
Control n = 51
Wu et al.97	II	5	Taiwan	Home-based exercise n = 18	60–85% max heart rate at least 3 times a week	No difference between groups in heart rate recovery (16.2 ± 4.8 beats/min)
CR n = 18	10-min warm up, 30–60-min aerobic training, and 10-min cool down
Control n = 18	Follow up at 12 weeks
CABG patients
Taylor et al.98	II	8	UK	Home-based rehabilitation n = 60	Nurse-facilitated, self-help package of 6 weeks duration (the Heart Manual)	Cost of home programme was less (MD −£30, 95% CI −£45 to −£12)
Normal CR n = 44	No difference in overall healthcare costs (MD £78, 95% CI −£1102 to £1191) or QALY (MD 0.06, 95% CI −0.15 to 0.02)
Clark et al.99	SR no meta-analysis	7	UK	SR of heart manual literature	Heart Manual (home-based CR)	Evidence from two RCTs suggests Heart Manual is as effective as normal CR on a number of psychological, behavioural, biological, service, and cost outcomes
Seven studies including two RCTs
Smith et al.100	II	7	Canada	Hospital-based CR n = 102	Exercise prescriptions were based on peak VO2 obtained during exercise testing.	At 12-month follow up for a 6-month home-based intervention, peak VO2 declined in hospital-based CR patients but not in home-based patients (p = 0.0002)
Home CR n = 96	Physical health-related QoL was higher in the intervention group
Home patients had higher habitual PA compared with patients who received CR in the hospital
Oliveira et al.101	III–1	4	Portugal	Recent MI	12 weeks of education and counselling through home visits and telephone contact	Direct between-group comparisons were not made
Home-based CR n = 15	Intervention group statistically significantly (p < 0.05) increased PA and time spent in moderate-intensity PA
Normal CR n = 15
Kodis et al.102	III–2	7	Canada	Traditional CR n = 713	Personalized exercise prescription	Retrospective analysis found that there were no statistically significant differences (p > 0.05) in peak VO2, peak workload, and peak MET levels between the groups at 6 months
Home-based exercise n = 329	Encouraged to exercise 3–5-times per week to a target heart rate
Decided by patient choice	Home-based exercise 5 times per week
Arthur et al.103	II	6	Canada	Home-based CR n = 120	Telephoned every 2 weeks and exercise logs monitored monthly	Similar improvement in peak VO2 between groups
Hospital-based CR n = 122	Home group demonstrated a greater improvement in health-related QoL (physical) by 6 months in comparison to the hospital patients (51.2 ± 6.4 vs. 48.6 ± 7.1, p = 0.004)
Post CABG
Rural, remote, and culturally and linguistically diverse population specific interventions
Dollard et al.40	I	8	Australia	No studies specifically developed for rural and remote	All home-based and categorized as:	Improved at more than 6 months: risk factors, psychological outcomes, patient satisfaction, visits to GP, ED visits, readmission, nursing intervention cost and hospital costs, mortality
14 studies outlining 11 non-conventional models identified	telephone contact only, home visits, telephone and home visits, Heart Manual
Multiple model of care categories
Cobb et al.104	SR no meta-analysis	6	USA	8 studies	Systematic review of interventions to reduce CHD risk factors	Strategies to reduce risk factors include frequent follow up, intensive diet change, individualized and group exercise, coaching, group meetings, education, and formal CR
Clark et al.2	I	10	Canada	63 RCTs	Meta-analysis of secondary prevention programmes	RR 0.83 (95% CI 0.77–0.94) for mortality overall but 0.53 (95% CI 0.35–0.81) at 24 months
Clark et al.29	I	10	Canada	46 trials (18 821 patients)	Meta-regression of programme characteristics	Pooled all-cause mortality: RR 0.87 (95% CI 0.79–0.97)
Programmes containing less than 10 hours contact: RR 0.8 (95% CI 0.68–0.95)
General practice programmes: RR 0.76 (955 CI 0.63–0.92), which was comparable with hospital-based programmes
Wong et al.46	SR no meta-analysis	9	Singapore	17 articles	One study focused on centre-based vs. no CR as well as home-based vs. no CR, nine studies compared centre-based with no CR, three studies compared centre-based with home-based, one study between inpatient vs. outpatient and four studies between home-based and no CR	Centre-based CR cost-effective
Home-based was no different to centre-based CR
No difference between inpatient and outpatient CR
Home-based cost-effective compared with no CR
Complementary and alternative medicine interventions
Manzoni et al.105	Protocol	n/a	Italy	CR patients n = 92	Follow up for 12 months	Protocol, no results
Disease-related expressive writing	HR-QoL, anxiety, depression, medical visits, CVD-related morbidities
Willmott et al.106	II	5	England	Intervention n = 88	Expressive writing	Number of medical appointments reduced, number prescribed medications reduced, more CR sessions attended, fewer cardiac-related symptoms, and lower diastolic blood pressure at 5-month follow up
Control n = 91
First MI patients
Meillier et al.107	III–2	2	Denmark	6-month extended rehabilitation programme for socially vulnerable patients. Involved extra individual nurse-led consultation, telephone follow up at 4 months, plan sent to GP, action-oriented and skills-training in diet, exercise, relaxation, smoking for up to 1.5 years at the Counselling Centre of the Danish Heart Foundation, non-cardiac specific activities at the community centre	No inequality was found in attendance and adherence between the groups
Chan et al.108	SR no meta-analysis	6	China	7 RCTs and 1 non-randomized	Chinese qigong exercise in CR programmes	Studies suggested this intervention seems to be an optimal alternative for patients unable to engage in other forms of PA
Taylor-Pillae et al.109	III–1	3	USA	n = 23 Tai chi	Wu style of Tai Chi	Better balance, perceived physical health, and tai-chi self-efficacy compared to those attending normal CR
Normal CR n = 28
Hildingh and Fridlund110	III–2	3	Sweden	MI, CABG, or PCI self-selected to receive peer support	Peer-support	Intervention group reported more health problems but scored higher on several dimensions of social support.
Peer support n = 64
No peer support n = 13
Barlow et al.111	II	6	UK	Intervention n = 95	Peer-support group for people who have completed cardiac rehabilitation	No statistically significant differences (p > 0.05) in general health status, MI-specific health status, self-efficacy, anxiety, and depression
Control n = 95
Arthur et al.112	SR no meta-analysis	6	Canada	Systematic review including studies on complementary and alternative therapies in CR	Tai Chi, acupuncture, transcendental meditation, and cheation therapy	Tai Chi, as a complement to existing exercise interventions, can be used for low and intermediate risk patients
Transcendental meditation may be used as a stress-reduction technique
There was insufficient evidence for the use of acupuncture and alternative medicines and chelation therapy

6MWT, six-minute walking test; ACS, acute coronary syndrome; BMI, body mass index; BP, blood pressure; CABG, coronary artery bypass graft; CAD, coronary artery disease; CASP, Critical Appraisal Skills Programme; CHD, coronary heart disease; COPD, chronic obstructive pulmonary disease; CR, cardiac rehabilitation; CV, cardiovascular; CVD, cardiovascular disease; ED, emergency department; HADS, Hospital Anxiety and Depression Scale; HDL, high-density lipoprotein cholesterol; HR, heart rate; IDR, incidence density ratio; LDL, low-density lipoprotein cholesterol; MD, mean difference; METs, metabolic equivalents; MI, myocardial infarction; MIDAS, Myocardial Infarction Data Acquisition Study; PA, physical activity; PASE, Physical Activity Scale for the Elderly; PCI, percutaneous coronary intervention; PTCA, percutaneous transluminal coronary angioplasty; QALY, quality-of life-adjusted life years; QoL, quality of life; RCT, randomized controlled trial; RR, relative risk; SR, Systematic review; TC, total cholesterol; TG, triglycerides; TR, Telerehabilitation; UC, usual care.

Multifactorial individualized telehealth delivery

There were 12 RCTs, one systematic review, two non-experimental studies, and two protocols for trials included in this category. Telehealth delivery of CR was effective when it addressed multiple risk factors in line with clinical guidelines, and was individualized to the patient. Telehealth delivery of CR was further enhanced by engagement with local services communication with physicians and specialists giving patients control over which risk factors to address and modes of addressing them.

Systematic review findings

A high-quality systematic review of telehealth interventions in CR¹⁹ examined 11 RCTs of CR for coronary heart disease patients. Significant favourable changes in total cholesterol (mean difference, MD, −0.37, 95% CI −0.56 to −0.19), high-density lipoprotein (MD 0.05, 95% CI 0.01 to 0.09), systolic blood pressure (BP) (MD −4.69, 95% CI −6.47 to −2.91), and smoking status (relative risk, RR, 0.83, 95% CI 0.7 to 0.99) were observed in meta-analysis of trials that compared telehealth interventions with usual care.¹⁹

Limitations of this systematic review relate to inconsistencies within the individual studies. Some or all patients in control and/or telehealth groups also participated in traditional, centre-based CR programmes. Women, the elderly, and culturally and linguistically diverse populations were under represented in the trials included in these systematic reviews. All reviews commented on the lack of cost-effectiveness data, or information about outcomes such as rehospitalization and repeat cardiac events.

Other telehealth studies

While the systematic review provided strong evidence for the effectiveness of telehealth CR, synthesis of data from all nine studies identified in this area provided information about effectiveness of specific intervention types, and combinations, on outcomes of interest. The CHOICE,^20,21 and, to a slightly lesser degree, the COACH interventions^22–24 stood out for their sustained effectiveness, significant clinical impact, and generalizability. Substantial treatment effects in total cholesterol (mean 4.0 ± 0.1 vs. 4.2 ± 0.1 mmol/l, p = 0.05), systolic BP (mean 132.2 ± 2.1 vs. 136.8 ± 2.0 mmHg, p = 0.01), physical activity scores (1200 ± 209 vs. 968 ± 196 metabolic equivalent minutes/week, p = 0.02), and proportion with three or more risk factors above national targets (20% vs. 42%, p = 0.02) resulted from the brief intervention used in the CHOICE trial.²⁵ Key features of this intervention include validated, local information on guidelines and resources to address each risk factor, individualized risk assessment, goal setting, re-evaluation and modification, active engagement of the patient in determining which risk factors to target and preferred management options to address them (summarized as doctor directed, hospital programme, individual programme, or self-help) as well as communication with local doctor and treating cardiologist.

Who may benefit, and from how much intervention?

The telephone-delivered interventions in this group were implemented in patients after ACS, percutaneous coronary intervention (PCI) and CABG, although CABG was excluded from Hanssen et al.^26,27). Telehealth interventions did not demonstrate a dose-dependent relationship with efficacy. Brief interventions (e.g. CHOICE^20,21 HeLM²⁸) demonstrated effectiveness comparable with more intensive strategies.²⁹ This may be related to the lower participant burden of brief interventions, as larger drop-out rates were observed in studies with a greater number of telephone contacts.^26,27 For example, only three participants (4%) withdrew from the intervention arm in the CHOICE trial, which lasted for 3 months, whereas 24 participants (15%) withdrew from the study by Hanssen, which involved a more intense intervention of 6-months duration.^21,26

Low-risk patients only

Clark et al.²⁹ described brief interventions as most appropriate for low-risk patients. However, Redfern and colleagues^20,21 recruited subjects from amongst those who declined to attend traditional centre-based CR, and who had greater baseline cardiovascular risk than the group who participated in centre-based exercise. Despite higher baseline risk, outcomes were not significantly different from improvements seen with centre-based CR.

Future research

Hawkes and co-investigators³⁰ report the protocol for a current RCT investigating a telehealth intervention in post-MI or PCI patients. The ‘ProActive Heart’ intervention is telephone and manual based, with individual risk factor assessment and health coaching on CHD risk factors, goal setting, and integration of local supports (social and environmental). The planned outcome measures include direct cost analysis.

Internet-based delivery of cardiac rehabilitation

Only RCTs and two non-experimental studies were included. Further, the studies were of poor quality, due to small sample sizes. Interventions required patients to communicate electronically with case managers³¹ or health professionals,^32,33 enter monitoring data, and access information.^31,33 Two studies offered opportunity for online group discussions.^31,33 Moore and Primm³³ described the use of ‘elaborate algorithms’ to tailor information to patient’s needs. Zutz and colleagues provided participants with a laptop computer which was interfaced with heart rate monitor, home blood pressure monitor, and dial-up internet access for the duration of the study.³² While no between-group differences were identified, the intervention group had reductions in risk factors similar to traditional CR.³² Southard and colleagues reported fewer cardiovascular events in the intervention compared with control groups.³¹ Of note, more data on the effectiveness of internet-based CR is expected, as three protocols for studies investigating the effectiveness of this alternative model of CR were identified.

Telehealth exercise-focused interventions with telemonitoring

Four randomized trials and five non-experimental studies were identified. The interventions involved home ECG monitoring of exercise sessions and telephonic transmission, with or without concurrent phone contact with a health professional. Most studies were conducted with high-risk patients recovering from cardiac surgery^34–36 and required patients to have the use of a stationary bicycle in their homes. The high degree of monitoring allowed close supervision of cardiac rhythm and symptoms.³⁷ In the largest RCT in this category, intervention group improvements in physical activity time (p = 0.027), physical activity sessions (p = 0.003), walking time (p = 0.013), and walking sessions (p = 0.002) were significantly greater than the control group at 6 weeks and remained significant at 6 months.³⁸

Telehealth recovery-focused interventions

Three studies were included in this category. One study³⁶ reported delivery of scripted daily information through a small screen electronic device, the Health Buddy, tailored by algorithms based on patient responses. The information addressed mostly surgical recovery, with limited information on secondary prevention of cardiovascular events. The intervention had high respondent burden and no significant improvements in physical function or healthcare use compared with usual care.³⁶

The remaining two studies involved weekly telephone calls from a nurse or peer-support person.^39,40 In the peer-support study, there were no differences in depression or perceived social support between intervention and control groups yet significantly less utilization of health services by the intervention group.⁴⁰ The study that involved telephone support from a trained nurse reported that no effect on psychosocial adjustment, anxiety, or depression was observed.³⁹

Community and home-based cardiac rehabilitation

There was consistent evidence from 22 RCTs and five non-experimental studies, which were of varying quality, sample size, and design, that community-based programmes were as effective as hospital-based programmes, and each other. For example, in a trial of 230 participants which compared home- vs. hospital-based CR, no difference was seen in the change in mean depression scores between the randomized home- and hospital-based groups (MD 0, 95% CI −1.12 to 1.12) nor mean anxiety score (MD −0.07, 95 CI −1.42 to 1.28), mean global MacNew score (MD 0.14, 95% CI −0.35 to 0.62), and mean total cholesterol levels (MD −0.18, 95% CI −0.62 to 0.27) at 9-month follow up. Also, a systematic review of eight studies identified that there was minimal difference between hospital- and home-based CR in terms of mortality or cardiovascular event rates.⁴¹

The majority of patient–provider contact in these studies was delivered face-to-face, through either home visits or patient attendance at community centres. Programmes involved multiple elements including graduated exercise, peer support, education, goal setting, and motivation. Most studies provided excellent detail on the different types of programmes tested, as well as inclusion criteria for patients into programmes. Vulnerable subgroups were addressed in the studies, with equally convincing effectiveness (e.g. elderly or lived alone). Studies considered patients with low-to-moderate and high risk of recurrent cardiac events.

Cardiac rehabilitation delivery for rural, remote, and culturally and linguistically diverse populations

Only one article of relevance to this section of the review was identified. Dollard and co-authors⁴² presented a narrative review of interventions delivered by telephone, which might be suitable for use in rural and remote regions. However, this review has been superseded by more recent comprehensive reviews of telehealth for CR that have been outlined above.¹⁹

Multiple models of care

Four systematic reviews were identified that evaluated the effectiveness of multifaceted interventions across a number of the categories outlined above. Clark et al.⁴³ reported that the relative risk reduction for mortality (RR 0.76, 95% CI 0.63–0.92) associated with general practice CR compared with hospital-based programmes. Another meta-analysis concluded that education and counselling were as effective with or without addition of supervised exercise in reducing all cause mortality (RR 0.87, 95% CI 0.76 to 0.99).² In a further systematic review, which included 18 trials of home-based CR vs. usual care and six RCTs of home- vs. centre-based CR, home-based CR was found to be as effective as the centre-based approach.⁴⁴

Also of note, shorter programmes with contact time of less than 10 hours implemented in the primary care setting were just as effective in reducing all-cause mortality as longer programmes in tertiary specialist centres (RR 0.80, 95% CI 0.68–0.95).⁴⁵ In the most recent of these reviews, it was identified that there was no difference in cost-effectiveness between home-based rehabilitation and a centre-based approach and that home-based rehabilitation was cost-effective compared with no rehabilitation.⁴⁶

Complementary therapies

A number of complementary and alternative medicine interventions have recently been examined for their effectiveness as an alternative to CR. Two systematic reviews, two RCTs, three non-experimental studies, and one protocol for a RCT were identified. Interventions include expressive writing, tai-chi, peer-support and Chinese qigong exercise. Of note, though, the quality of these studies was judged to be low (CASP scores less than 6). As such, there is no strong evidence to draw conclusions regarding effectiveness.

Discussion

In this systematic review, we have identified numerous alternative models of OCR. However, we found that only the community-based and telehealth-based individualized and multifactorial models for CR were associated with improvements in cardiovascular disease risk factor profile similar to those with the traditional hospital-based approach. Due to the high quality of these studies, the strength of the body of evidence indicates there is no need to rely on hospital-based strategies alone to deliver effective CR. As such, it is recommended that local healthcare systems should strive to integrate alternative models in order to ensure there are choices of high-quality, evidence-based CR programmes available for patients that best fit their needs, risk profile, and preferences.

In contrast, at this stage there remains insufficient evidence to support the effectiveness of internet-based delivery of CR compared with the high-quality evidence demonstrating the effectiveness of personal contact via telephone through ‘coaching’ or a written action plan.^20–23 Furthermore, as the studies involving telehealth interventions that focused solely on exercise were of low quality, produced only small changes in measured outcomes, and were associated with intensive requirements on staff and participants, the evidence at hand does not yet support implementation of these strategies.

Key features of effective interventions

Strong evidence supports the value of flexible interventions, based on individualized risk factor assessment. While patients are informed about evidence-based guidelines for their risk reduction, brief interventions which encouraged autonomy and choice had long-lasting effects. However, this needs to be supported by infrastructure: local healthcare providers (primary care physicians, community/indigenous health workers) and specialists were involved with successful programmes. Translation of patient goals into action was also facilitated by comprehensive information about local means to support lifestyle change. While long-term benefits have been evaluated from a single centre,^14,21 multicentre trials are required to provide insight into how this intervention can be adapted for a variety of metropolitan, regional, and remote settings.

Risk assessment and intervention intensity

Most of the literature, particularly for community programmes, focused on cardiac patients with low-to-moderate risk.⁴⁷ Home-based rehabilitation of the higher risk, post surgical patient was possible with telemonitoring and extensive equipment.^35,36 However, evidence of the effectiveness of these interventions is weak, due to the quality of research and the considerable costs associated with the use of sophisticated telemonitoring equipment.

Patients who decline to attend CR are known to have greater risk factors than those who attend.⁴⁸ A number of the telehealth studies in this review targeted CR non-attendees for recruitment, yet still demonstrated significant improvements.^21,25

Individualized risk assessment was a vital starting point for CR in the effective studies reviewed and serves to guide selection of appropriate level of supervision for cardiac patients at higher risk, tailor aspects of the rehabilitation programme to those risk factors which require most intervention and inform patients about their condition and engage them in the process of rehabilitation.^19,25

Gaps in the knowledge base

Areas where evidence is lacking have been observed. Few studies reported cost effectiveness or long-term information on rehospitalizations and cardiac events. While more recent trials are reflecting the real-world actual gender mix of cardiac hospital separations (around 70% male, 30% female),³⁰ women were frequently cited as under-represented in clinical trials of CR.

There was a lack of research that targeted rural, remote, or indigenous populations. No studies with Level I or II evidence were identified which specifically targeted rural, remote, or indigenous populations. As such, there is insufficient evidence in the published literature to support specific models of care for these population groups.^42,49,50 Fortunately, there is literature available to inform development of culturally specific alternative CR programmes for indigenous populations. For example, a survey of healthcare providers and indigenous cardiac patients admitted to hospital found indigenous patients reported similar barriers to accessing CR to the non-indigenous populations.⁵¹ The main barrier described to uptake was lack of referral to the programme or lack of knowledge about the programme. Another key finding of this study was the observed need to engage indigenous cardiac patients in hospital, using indigenous health workers.

Also, Parker et al.⁵⁰ described a CR education resource developed for rural health workers in Australia and reported on focus groups conducted with 60 rural health workers. The main finding was a need to better adapt patient information brochures for indigenous communities.

Similarly, a report from the NHMRC⁵² contained general statements about cultural appropriateness of engagement of cardiac patients in hospital and included recommendations on:

• cultural awareness of health workers related to patient need.

• recognition of the earlier onset of chronic health conditions for indigenous compared with Western peoples.

• putting sound and comprehensive discharge plans in place before patients leave hospital, ensuring that patients can access services once they are discharged (taking into account distance to travel, service availability, cost, etc).

• providing patients with culturally appropriate and language-appropriate literature on risk modification and lifestyle change.

Interestingly, all telehealth studies reported to date have been carried out in metropolitan areas. A significant component of the telehealth interventions includes facilitating uptake of local services and enhanced communication with local healthcare providers to reduce cardiac risk factors.^20,23 The implementation of telehealth interventions in rural and remote populations will require careful attention to these and multiple other issues of translation.

Other groups underrepresented in this body of literature include culturally and linguistically diverse populations, migrant, and refugee groups, older ethnic populations whose understanding of spoken or written English may be poor, and the vulnerable unwell whose capacity to engage with the health system (except on an emergency basis) may be limited.

Limitations

As with all systematic reviews, ours is limited by the quality of the included studies. To facilitate transparency, though, each included study was critically appraised using an appropriate CASP tool and the NHMRC level of evidence. It is also important to note that only articles published in English were included in our review. However, sensitivity testing regarding information published in languages other than English has shown that English language reviews represent a robust view of the available evidence base in health areas.^53–55

Conclusions

Numerous alternative models of care for CR other than the traditional hospital-based approach were identified. The evidence at hand indicates alternative models for CR produce similar reductions in cardiovascular disease risk factors compared with hospital-based programmes and are just as cost-effective.

Acknowledgements

The members of the review update team wish to acknowledge the contribution of Dr Kylie Johnstone and members of the team at the International Centre for Allied Health Evidence, University of South Australia. They also acknowledge the contribution and support of the South Australian Statewide Cardiology Clinical Network, Prevention and Rehabilitation Workgroup members, SA Health, the Chief Medical Officer, Professor Paddy Phillips and contributions from the South Australian Cardiovascular Health and Rehabilitation Association. The review team also acknowledges the support of Petra Lawrence, The Prince Charles Hospital, and Prof Patsy Yates, School of Nursing, Queensland University of Technology, Queensland, Australia.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.