Cardiac Rehabilitation for Women: A Systematic Review of Barriers and Solutions

Abstract

Cardiac rehabilitation (CR) services improve various clinical outcomes in patients with cardiovascular disease, but such services are underutilized, particularly in women. The aim of this study was to identify evidence-based barriers and solutions for CR participation in women. A literature search was carried out using PubMed, EMBASE, Cochrane, OVID/Medline, and CINAHL to identify studies that have assessed barriers and/or solutions to CR participation. Titles and abstracts were screened, and then the full-text of articles that met study criteria were reviewed. We identified 24 studies that studied barriers to CR participation in women and 31 studies that assessed the impact of various interventions to improve CR referral, enrollment, and/or completion of CR in women. Patient-level barriers included lower education level, multiple comorbid conditions, non-English native language, lack of social support, and high burden of family responsibilities. We found support for the use of automatic referral and assisted enrollment to improve CR participation. A small number of studies suggest that incentive-based strategies, as well as home-based programs, may contribute to improving CR attendance and completion rates. A systematic approach to CR referral, including automatic CR referral, may help overcome barriers to CR referral in women and should be implemented in clinical practice. However, more studies are needed to help identify the best methods to improve CR attendance and completion of CR rates in women.

Cardiac rehabilitation (CR) is a cost-effective, class 1 recommended component of the continuum of care for patients with cardiovascular disease^{1, 2} that has been shown to improve various important patient outcomes, including exercise capacity, cardiovascular risk factor control, social functioning, and psychological well-being, hospital readmission rates, and mortality rates.^{1, 3–9}

Despite this, however, CR utilization remains low,^{6, 10–13} particularly among women.^14–16 Women are substantially less likely to be referred to a CR program (odds ratio [OR], 0.68),¹⁵ to enroll in CR once referred (OR, 0.64),¹⁴ and to complete a full course of CR (OR, 0.73), as compared with men¹⁷ (OR, 0.89).¹⁸ The negative impact of this treatment gap is accentuated by the fact that clinical outcome improvements after CR are at least as great in women as in men.^17–24

Although some studies have identified gender-related barriers to CR^{14, 25, 26} and others have assessed efforts to reduce those barriers,^27–29 uncertainty exists regarding the relative strength of various strategies to improve CR utilization by women. With the objective to provide clinicians and policymakers with specific evidence-based strategies to improve CR participation in women, we carried out a systematic literature review to (1) identify gender-related barriers to CR participation, and (2) enable us to rate solutions to the gender-related gap in CR participation, according to the level of published evidence that supports them.

METHODS

To identify studies on barriers and their potential solutions to CR participation, we carried out a systematic and comprehensive search in PubMed, EMBASE, Cochrane, OVID/Medline, and CINAHL databases. An expert librarian designed the search with input from the lead investigator (M.S.P.) using key search terms (see Supplemental Appendix 1, available online at http://www.mayoclinicproceedings.org). We supplemented the search by reviewing references in the selected articles, structured reviews, and meta-analyses. We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) recommendations in performing our systematic review.³⁰

To identify studies of barriers and solutions to CR participation that are pertinent to recent gender-related health care delivery gaps, we selected randomized clinical trials, controlled clinical trials, and observational studies that had been published before October 20, 2016, without language restrictions. Studies were restricted to those whose sample included adults older than 18 years who had CR-eligible diagnoses, such as myocardial infarction, percutaneous coronary intervention, coronary artery bypass graft surgery, heart failure, heart valve surgery (replacement or repair), heart transplantation, and stable angina. International studies were included.

We excluded studies that lacked gender-specific results, although for studies on potential solutions to low CR referral, enrollment, and completion rates, we included all studies in which the percentage of women in the sample was specified, even if the results were not specified by gender, given the low number of studies in this category. Although qualitative studies can help clarify the underlying issues surrounding the delivery of health care services, we excluded them from our analysis to focus on the highest level of published evidence from clinical trials and observational studies that have studied the most common barriers to and most effective solutions for CR participation in women.

Two reviewers (M.S. and J.R.M.-I.) working independently and in duplicate analyzed the eligible studies to assess the risk of bias using a modified Ottawa classification for observational studies and the Cochrane assessment tool for randomized clinical trials (RCTs).^{31, 32} Disagreement between reviewers was resolved by consensus.

The search results were uploaded using systematic review software (Covidence). Two reviewers (M.S. and J.R.M.-I.) independently reviewed all abstracts and titles for inclusion. After abstract screening and retrieval of potentially eligible studies, the full-text publications were assessed for eligibility with excellent chance-adjusted interreviewer agreement (K statistic, 0.831169). Duplicate studies were excluded. Disagreements were resolved by group consensus. The senior author (R.J.T.) reviewed results for accuracy.

All coauthors reviewed the solutions identified from the literature search and scored each solution by the American College of Cardiology/American Heart Association guidelines for level of evidence (level A = 1 or more high-quality RCTs or meta-analyses; level B-R = 1 or more moderate quality RCTs or meta-analyses; level B-NR = 1 or more moderate quality nonrandomized studies; level C = 1 or more studies with significant limitations; level E = based on expert opinion, published evidence lacking or unclear) and strength of recommendation (class 1 = strong, benefit much greater than risk; class 2a = moderate, benefit greater than risk; class 2b = weak, benefit greater than or equal to risk; class 3 = no benefit [moderate], benefit equal to risk; class 3: harm [strong], risk greater than benefit) for each solution.³³ Coauthor responses were summarized and any items that lacked a clear majority agreement were resolved by additional discussion. Because the barriers identified did not represent the same constructs, authors agreed to not pool estimates and rather focus on summarizing the available evidence.

RESULTS

A total of 3988 articles were screened, from which we identified 54 studies that met our inclusion criteria. Of these, 24 studies^{18, 27, 34–55} were studies of CR barriers and 31 studies^{27–29, 56–83} were studies of potential solutions to improve CR participation. One study addressed both CR barriers and potential CR solutions. Of the included studies addressing CR barriers one was an RCT²⁷ and 23 were observational cohort studies.^{18, 34–55} The studies addressing potential CR solutions included 15 RCTs^{27–29, 56, 58, 59, 62, 71, 72, 74–76, 78, 82, 83} and 16 observational studies.^{57, 60, 61, 63–70, 73, 77, 79–81} The study selection process is described in more detail in Supplemental Figure 1, available online at http://www.mayoclinicproceedings.org.

Overall, studies included in our review had moderate risk of bias mainly driven by unclear allocation concealment and blinding for RCTs and/or inappropriate patient selection and follow-up for observational studies. A summary of the risk of bias of the included studies is presented in Supplemental Appendix 2 and Supplemental Figure 2, available online at http://www.mayoclinicproceedings.org.

Barriers to CR Referral, Enrollment, and Completion

We identified 24 studies related to barriers on CR participation in women.^{18, 27, 34–55} Table 1 lists the patient-, provider-, and social/ environmental-level factors that were identified, including those that are nonmodifiable (eg, age and diagnosis) and potentially modifiable (eg, transportation barriers). It is worth noting that the studies in our analysis reported barriers that were pertinent to the patient groups unique to their study cohort and local factors (eg, the ethnicity/racial mixture of the study cohort, transportation challenges, or number of available CR centers). Care should be taken in applying in one location or setting, the results of studies from different locations or settings.

TABLE 1.

Factors Associated With Lower CR Participation for Women

CR participation steps	Levels of barriers	Factor
Referral	Patient	Coronary artery bypass graft,40 percutaneous coronary intervention,40, 43 and valve surgery50 as reason for CR referral
		Older age34
	Health care provider	Lack of written referral necessary for participation from physicians38
	Social/environment	Individuals from underrepresented minority groups, particularly those with financial barriers35, 49
		Lack of CR insurance48
Enrollment	Patient	Perception of exercise as being tiring or painful18
		Lack of CR awareness18
		Multiple comorbidities18, 44
		PCI43 as reason for CR admission
		Higher exercise barriers as measured by the EBBS score43
		History of myocardial infarction39
		Older than 70 y or younger than 55 y55
	Health care provider	Lack of strong endorsement to attend to CR38, 52
		Lack of health care provider support54
	Social/environment	Education level <12 y52
		Transportation issues18
		Numerous family responsibilities18 or home- related stress42
		Lack of support system (friends, family)54
		Individuals from underrepresented minority groups, particularly those with financial barriers35
		Unemployed42
Completion/adherence	Patient	Multiple comorbidities45
		<55 y of age45
		Obesity51
		Depression as measured by the BDI-II score51
		Diabetes36
		Previous myocardial infarction39
		High level of anxiety as measured by STAI-S score27
		Current smoker27
		Physically inactive previous CR55
		Lack of history of CHD48
	Health care provider	None reported
	Social/environment	Divorced/separated27
		Transportation problems45, 46
		Long distance between CR program and place of residence46
		Numerous family responsebilities45
		Lack of CR insurance48

BDI-II = Beck Depression Inventory-ii; CHD = coronary heart disease; EBBS = Exercise Benefits and Barriers Scale; CR = cardiac rehabilitation; MI = myocardial infarction; STAI-S = State-Trait Anxiety Inventory.

Patient-level Factors

Age and^{34, 55} comorbidities,¹⁸ including musculoskeletal,⁴⁵ depression,⁵¹ diabetes,³⁶ and obesity⁵¹ among others factors, have been shown to limit participation in and completion of CR by women. A patient’s lack of information on or familiarity with CR can serve as a barrier to participation.¹⁸ Likewise, negative beliefs or perceptions about CR are associated with lower CR participation.^{18, 43}

Of note, we found discrepant results regarding the association between CR participation and a history of either percutaneous coronary intervention or coronary artery bypass graft, with 2 studies^{40, 43} reporting lower rates of CR participation in these groups and another study suggesting the opposite.⁵³ A similar discrepancy was noted with a history of depression, with one study⁵¹ reporting that depression was associated with lower rates of CR completion, whereas another study showed no significant association between depression and CR participation, for men or for women.³⁷ Of interest, obesity was found to be associated with higher CR participation rates⁴¹ but lower completion rates, as compared with nonobese patients.⁵¹

Provider-level Factors

A strong, supportive endorsement of CR by a health care professional has been shown to improve CR participation rates for women and for men.⁵⁴

Social/Environmental-level Factors

Transportation problems,^{18, 45} family obligations at home, lack of CR insurance, and financial concerns^{35, 49} are all associated with lower CR participation by women. In addition, a lack of social support from family and friends has been reported to be a barrier to CR enrollment.⁵⁴

Solutions to CR Referral, Enrollment, and Completion

A total of 31 studies were identified that assessed the impact of various interventions that were aimed at improving CR participation.^{27–29, 56–83} The percentage of participants who were women in these studies was 37.37%, with 27 of the 31 studies including less than 50% of female participants. Six studies listed results separately for women and men,^{57, 62, 68, 72, 78, 79} whereas only 3 studies assessed interventions to improve CR participation exclusively in women.^27–29

Interventions that have been reported to improve CR participation are listed in Table 2, along with the ratings for their level of evidence and strength of recommendation. Systematic approaches to CR referral (eg, automatic referral and “liaison” coaching interventions) have been found to improve CR referral in women and in men, but their impact on CR enrollment is variable.^{67, 68} Interventions aimed to help patients enroll in CR promptly after hospital discharge have been found to improve CR referral, enrollment, and completion rates.^{65, 79} In addition, interventions that increase CR awareness among patients and health providers have been shown to increase referral and enrollment rates.^{60, 62}

TABLE 2.

Summary of Evidence and Recommendations for Improving CR Referral, Enrollment, and Completion, Regardless of Gender

A. CR Referral evidence-based solutions
Intervention	Description	Articles	Proportion of women	Location	Level addressed	Level of evidence	Strength of recommendation
Early access	Integration of early cardiac access clinic increased referral rates to 8 times compared with the comparison group	Parker et al,79 2011 CCT	41%	Canada	Health care provider	B-Nonrandomized	IIa
Automatic referral	Use of the American Heart Association GWTG pathway was associated with 0.5 times higher referral rates	Mazzini et al,73 2008 observational study	33%	USA	Health care provider	B-Nonrandomized	I
	Systematic referral methods improved CR referral by 2-7 times compared with usual care	Gravely et al,68 2014 observational study	25%	Canada	Health care provider
		Grace et al,67 2007 observational study	30.84%	Canada
		Grace et al,65 2012 observational study	26.8%	Canada
		Grace et al,66 2011 observational study	26.8%	Canada
	Preapproved strategy (cardiac program leadership endorsement of a policy for referral by an allied health professional of all indicated patients) increased referral rates by 2 times that in the comparison group	Grace et al,63 2012 CCT	25%	Canada	Health care provider
Liaison referral	Liaison intervention increased referral rates 3 times that of usual care	Gravely et al,68 2014 observational study	25%	Canada	Health care provider	B-Randomized	I
		Grace et al,66 2011 observational study	26.8%	Canada
		Jolly et al,71 1998 RCT	29%	UK
Peer navigation	Bedside visit, education material, and encouragement to obtain CR referral before discharge did not have any effect on referral rates	Ali-Faisal et al,56 2016 RCT	17%	USA	Health care provider	B-Randomized	IIb
Automatic + liaison referral	Combined automatic and liaison referral resulted in increasing CR referral rates by 8 to 10 times compared with usual care	Grace et al,66 2011 observational study	26.8%	Canada	Patient + Health care provider	B-Nonrandomized	I
		Gravely et al,68 2014 observational study	25%	Canada
Increase awareness	Increase awareness of CR among health care providers through education combined with a formal referral system increased referral rates by 38 times compared with usual care	Dahhan et al,60 2015 observational study	41.3%	USA	Patient	B-Nonrandomized	I
B. CR enrollment/participation evidence-based solutions
Intervention	Description	Articles	Proportion of women	Location	Level addressed	Level of evidence	Strength of recommendation
Increase awareness	Increased awareness by patients and health care providers increased CR participation by 2 to 32 times, respectively, compared with usual care	Dankner et al,62 2015 CCT	23.83%	Israel	Patient	B-Nonrandomized	I
		Dahhan et al,60 2015 observational study	41.3%	USA
Patient navigation	Education about CR and support by phone and mail by navigator increased awareness and CR enrollment by 9 times compared with usual care	Scott et al,82 2013 RCT	33.7%	USA	Health care provider	B-Randomized	IIa
	Information about benefits of CR and then providing referral process assistance increased enrollment by 1.8 times	Pasquali et al,80 2001 observational study	38%	USA
Referred to site closer to home	Participants rereferred to a CR program closer to their home were significantly more likely to enroll than those who were not	Ali-Faisal et al,56 2016 RCT	17%	USA	Social + Environment	B-Randomized	IIa
Physician recommendation	Strong CR endorsement by a nurse or a physician increased attendance by 3 and 14 times, respectively	Johnson et al,70 2010 observational study	37.81%	Australia	Health care provider	7C	IIa
Early access	Early appointment increased participation rates from 1.6 to 4 times higher than in the comparison group	Parker et al,79 2011 CCT	41%	Canada	Health care provider	B-Nonrandomized	I
		Pack et al,78 2013 RCT	25%	USA
		Grace et al,63 2012 CCT	25%	Canada
Automatic referral	Automatic referral was associated with an increase in participation 2 to 5 times higher than in the comparison group	Gravely et al,68 2014 observational study	25%	Canada	Health care provider	B-Nonrandomized	I
		Grace et al,67 2007 observational study	30.84%	Canada
		Grace et al,65 2012 observational study	26.8%	Canada
		Grace et al,66 2011 observational study	26.8%	Canada
		Grace et al,63 2012 CCT	25%	Canada
		Mueller et al,76 2009 CCT	31%	USA
		Grace et al,64 2004 observational study	39.3%	Canada
Liaison referral	Liaison assistance increased enrollment rate compared with usual care	Grace et al,66 2011 observational study	26.8%	Canada	Health care provider	B-Nonrandomized	I
		Gravely et al,68 2014 observational study	25%	Canada
Automatic referral + liaison	Combined systematic referral and liaison strategy resulted in 4 to 6.5 times greater enrollment	Grace et al,66 2011 observational study	26.8%	Canada	Patient + Health care provider	B-Nonrandomized	I
		Gravely et al,68 2014 observational study	25%	Canada
Letter	Theory-based invitation letter improved attendance by 3-fold	Mosleh et al,75 2014 RCT	31%	Jordan	Health care provider	B-Randomized	IIa
Gender-tailored intervention	Gender-tailored behavioral intervention increased attendance by 4 to 5 sessions	Beckie and Beckstead,27 2010 RCT	100%	USA	Patient	B-Randomized	IIa
Alternative programs models	No difference in CR attendance between supervised mixed-sex, supervised women-only, or home-based CR	Grace et al,28 2016 RCT	100%	Canada	Patient + Social/ Environment	B-Randomized	IIb or C
		Midence et al,29 2016 RCT	100%	Canada
Community program	CR participation by women increased by 1.2 times vs hospital-based CR	Blackburn et al,57 2000 observational study	29%	USA	Social/ Environment	B-Nonrandomized	IIa
Home visit	Peer advisor/advanced practice nurse intervention increased participation rates at 3 mo by 3 times	Carroll et al,58 2007 RCT	66%	USA	Health care provider	B-Randomized	IIb
Nursing phone call	Nursing assessment and encouragement by phone call 2 wk before CR increased participation by 3.4 times	Harkness et al,69 2005 observational study	24.66%	Canada	Health care provider	C	IIa
One inpatient visit and 2 outpatient calls	Nursing inpatient visit and 2 additional contacts during the 10 d after hospital discharge increased CR participation rates by 2.6 times	Cossette et al,59 2012 RCT	16.9%	Canada	Health care provider	B-Randomized	IIa
Motivational interviewing	Personalized motivational interviewing increased CR participation by 2 sessions compared with the comparison group	McGrady et al,74 2014 RCT	34%	USA	Patient	B-Randomized	IIa
Home-based	Uptake was higher in home-based program compared with center-based program	Jolly et al,72 2007 RCT	23.42%	UK	Patient + Social/Environment	A	I
		Varnfield et al,83 2014 RCT	16.5%	Australia
C. CR completion/adherence evidence-based solutions
Intervention	Description	Articles	Proportion of women	Location	Level addressed	Level of evidence	Strength of recommendation
Early access	Integration of early cardiac access clinic increased CR completion rates by 1.4 times above the comparison group	Parker et al, observational study	41%	Canada	Health care provider	B –Nonrandomized	Ia
Automatic referral	No significant differences in completion rates with systematic inpatient referral compared with usual care	Grace et al,65 observational study	26.8%	Canada	Health care provider	B-Nonrandomized	I
Incentive/ motivational programs	Incentive/motivational programs increased completion rates by 1.8 times	Pack et al,77 observational study	30%	USA	Patient	B-Nonrandomized	IIa
Home based	Telephone consultation and education by webcast program did not show significant difference vs traditional program	Scane et al,81 observational study	20.5%	Canada	Patient + Social/ Environment	B- Randomized	IIa
Telemedicine supported	Telemedicine-delivered CR in a rural setting had similar adherence to conventional CR	Dalleck et al,61 observational study	42%	New Zealand	Patient + Social/ Environment	B-Nonrandomized	IIb
Smartphone/Web-based model	Smartphone-based/Web-based model improved completion rates by 1/3 compared with the traditional program83	Varnfield et al,83 RCT	16.5%	Australia	Patient + Social/ Environment	A	I

CCT = controlled clinical trial; CR = cardiac rehabilitation; GWTG = get with the guidelines; RCT = randomized clinical trial.

Interventions with a letter describing a strong support for CR participation by a health care provider, posthospital home-based visits, and/or telephone calls have also been found to help improve CR participation.^{58, 59, 69, 75} Home-based CR delivery may contribute to improving participation and completion rates⁷² for men and for women, based on the results of a meta-analysis.^{84, 85} Tele-health delivery models may also help increase CR enrollment and completion rates, particularly in women, who appear to be greater adopters of tele-health health care delivery tools and interventions.^{86, 87}

DISCUSSION

This is the first study to our knowledge to have systematically reviewed CR participation barriers for women and provide a practical summary of specific and effective evidencebased interventions to facilitate their participation with the last goal of decreasing the high burden of cardiovascular disease among women.

We identified 24 studies^{18, 27, 34–55} that ascertained significant and unique barriers for women associated with reduced referral, enrollment, and participation in women. The barriers identified reflect a complex array of demographic, socioeconomic, medical, and societal challenges that impede the delivery of CR services for all patients, but particularly for women. Many of the barriers appear to be either nonmodifiable (eg, age and gender) or are not easily modifiable (eg, socioeconomic status, financial constraints, and education level). Others, such as a low awareness of CR or a lack of strong physician recommendation or referral to CR, appear to be more amenable to corrective action.

We also identified 31 published studies^{27–29, 56–83} that analyzed the impact of various interventions on CR referral, enrollment, and completion. Several types of interventionsdprimarily systematic approaches to CR referral, enrollment, and completiond were considered to be of highest strength of recommendation and level of evidence, as noted in Table 2. It is worth noting that these highly rated, systematic approaches to improving CR delivery can help to reduce most, if not all, types of barriers to CR participation, including both modifiable and nonmodifiable barriers. Systematic approaches, such as automated CR referral systems,^{64–68, 73, 76} liaison/nurse-to-patient contact,^{56, 66, 68, 80, 82} and early posthospital enrollment,^{63, 78, 79} are feasible and effective ways to overcome referral bias, improve awareness of CR, and address other barriers that are particularly challenging for women in need of CR. The use of systematic approaches to CR enrollment would help to address the current common problem of delayed enrollment,⁷⁸ an important issue since each day of enrollment delay translates into a 1% decrement in the likelihood of CR enrollment.⁸⁸ At the same time, such approaches may play an essential role in addressing the lack of support or endorsement that have been reported as barriers to CR participation among women.

Furthermore, incentive programs,^{74, 77} flexible hours, and the use of alternative delivery CR models are additional strategies that can help women once they enroll in CR to avoid “dropping out,” and instead receive a “full dose” of CR services. The impact of such interventions can have a considerable impact, potentially tripling the level of CR participation and the related health benefits for women (and for men) in need of CR services.⁸⁹ Alternative models of CR delivery that offer more flexible and personalized treatment options, including the use of home-based and/or smartphone-based CR models, may be even more ideally suited for women than for men.^{86, 87} Women-only CR programs have been tested in a relatively small number of studies, and may play a role in helping to provide more flexible options to women in need of effective CR services. Further research is warranted in exploring the impact of these alternative delivery models on CR outcomes in both women and men. In addition, further studies are needed that explore the barriers and potential solutions to CR participation for women, related to their specific psychosocial and health perception parameters.^{27, 51} Additional work is also needed to explore the impact of larger-scale, public awareness campaigns, such as the Go Red For Women campaign from the American Heart Association on CR awareness and participation among women.⁹⁰ Finally, the interplay between gender and ethnicity warrants further investigation, particularly because women in underrepresented minority groups experience various health care disparities, including the underutilization of CR services.⁹¹

Study Limitations

Our review of barriers to CR participation for women was admittedly focused on quantitative research studies and could be enhanced by the addition of qualitative studies that may further elucidate some CR barriers, such as the lack of peer support.⁹² Our assessment of interventions aimed at improving CR participation is limited by the relatively low availability of published studies on gender-based solutions to CR referral, enrollment, and participation. However, the studies in our analysis did include a significant number of women, sufficient to provide evidence for the impact of various interventions on CR participation in women.

CONCLUSION

A wide, complex variety of modifiable and nonmodifiable barriers exist that limit CR participation in women. High-quality systematic approaches to improving CR participation can help to overcome these barriers and are strongly recommended. New delivery models for CR, such as home-based and/or smartphone-based CR, appear to be promising approaches to help improve CR delivery to women, but further research is needed in this important area.

ARTICLE HIGHLIGHTS.

• Fewer women than men participate in cardiac rehabilitation because of a complex and unique array of demographic, socioeconomic, medical, and societal challenges faced by women.

• Systematic approaches to cardiac rehabilitation referral, enrollment, and participation have been shown to improve cardiac rehabilitation participation and are recommended to help overcome the gender gap in cardiac rehabilitation participation.

• Further research is warranted on the impact of novel cardiac rehabilitation delivery models, such as home-based cardiac rehabilitation, on the gender gap in cardiac rehabilitation participation.

Abbreviations and Acronyms

CR

cardiac rehabilitation

OR

odds ratio

RCT

randomized clinical trial