Participation and adherence to cardiac rehabilitation programs. A systematic review

International Journal of Cardiology 223 (2016) 436–443

Contents lists available at ScienceDirect

International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard

Participation and adherence to cardiac rehabilitation programs. A systematic review Alberto Ruano-Ravina a,b,⁎, Carlos Pena-Gil c, Emad Abu-Assi c, Sergio Raposeiras c, Arnoud van 't Hof d, Esther Meindersma d, Eva Irene Bossano Prescott e, Jose Ramón González-Juanatey c a

Department of Preventive Medicine & Public Health, University of Santiago de Compostela, Santiago de Compostela, Spain Consortium for Biomedical Research in Epidemiology & Public Health (CIBER en Epidemiología y Salud Pública - CIBERESP), Spain Service of Cardiology, Clinical University Hospital of Santiago de Compostela, Santiago de Compostela, Spain d FESC Isala Heart Center, Zwolle, Netherlands e Department of Cardiology, Bispebjerg University Hospital, Copenhagen, Denmark b c

a r t i c l e

i n f o

Article history: Received 17 January 2016 Received in revised form 4 August 2016 Accepted 5 August 2016 Available online 13 August 2016 Keywords: Systematic review Cardiac rehabilitation Acute myocardial infarction Physical exercise

a b s t r a c t Acute myocardial infarction (AMI) is an important health problem. Cardiac rehabilitation (CR) programs following AMI have shown to be effective in reducing mortality. We aim to systematically review the existing literature that analyzes the factors that affect participation and adherence to cardiac rehabilitation programs. We reviewed Medline, EMBASE and Cochrane databases from 01/01/2004 to June 2016 using predefined inclusion and exclusion criteria. We classified the results into factors affecting participation and factors influencing adherence to CR programs. We included 29 studies, and there was a general agreement in those factors predicting participation and adherence to CR programs. These factors can be classified into person-related factors and aspects related to CR programs. Older participants, women, patients with comorbidities, unemployed and uncoupled persons, less educated people and those with lower income had a lower participation. A similar pattern was observed for CR adherence. Also, those potential participants who live farther from CR facilities, do not have transportation, or do not drive, attended less to CR programs. These factors were very similar when analyzing adherence to CR programs. These aspects were similar in Europe and the USA. These results clearly show that participation in CR programs follows a determined pattern that is very homogeneous in different settings. Health professionals should also be aware of patients reluctant to participate in CR programs and adapt their messages and redesign CR programs, to promote participation and adherence. © 2016 Elsevier Ireland Ltd. All rights reserved.

1. Introduction Cardiovascular diseases are the main cause of mortality and disability in developed countries [1] and acute myocardial infarction (AMI) is leading cause of cardiovascular disease mortality in certain age groups [2] [3]. Currently, survival following an AMI is relatively high. This is due to different factors. The main determinant of survival is quick access to medical assistance. Following access to medical assistance, diverse interventions are offered to patients. These interventions can be classified into pharmacologic (adequate treatment), medical (PCI or CABG), and changes in lifestyle (diet, smoking, psychosocial factors, physical inactivity), including the promotion of physical exercise. Cardiac rehabilitation (CR) has been consistently shown to reduce coronary events and mortality in those patients who have suffered an ⁎ Corresponding author at: Departament of Preventive Medicine and Public Health, School of Medicine, San Francisco Street, University of Santiago de Compostela, 15782, Santiago de Compostela, Spain. E-mail address: [email protected] (A. Ruano-Ravina).

http://dx.doi.org/10.1016/j.ijcard.2016.08.120 0167-5273/© 2016 Elsevier Ireland Ltd. All rights reserved.

AMI [4]. CR programs usually consist of a personal assessment of the patient, advice on physical activity, training exercises, nutritional advice, weight management, lipids and blood pressure control, tobacco cessation, and psychosocial management. CR should be offered to patients who have suffered an AMI [5]. CR following a cardiovascular event is a Class I recommendation of the European Society of Cardiology, the American Heart Association, and the American College of Cardiology [6]. CR also offers economic benefit. It has been estimated that CR can yield cost savings of €30,500 per patient in the first year (mainly due to return to work) and up to €14,500 per year in the following years [6]. Unfortunately, not all CR programs are performed adequately. A clinical trial performed in the UK compared patients in a CR program with those assigned to standard treatment. The study did not observe mortality differences at 9 year follow-up [7], raising the question of adequate program implementation [8]. There is great variability in the participation rates of these programs between different countries. In Germany, participation ranges between 49 and 65% [9], which is higher than the rate observed in the UK or USA.

A. Ruano-Ravina et al. / International Journal of Cardiology 223 (2016) 436–443

In Spain, there are very few multidisciplinary CR units and these programs are usually performed in a hospital setting. There are different program types corresponding to different patients' needs and the supervision of potential complications. Low participation rates in CR programs are a worrying concern for cardiologists and scientific societies. Some cardiologists do not recommend patients to take part in CR programs. Increasing participation rates in CR programs is therefore proves to be a challenge. Currently, participation is lower than expected, as many studies demonstrate [10,11]. There have been studies that compared CR participants with those who rejected participation, ultimately showing that they have different characteristics. There is also the issue of healthcare characteristics (public versus private healthcare coverage); it is possible that participation may differ between different healthcare systems. There are also some groups that have been underrepresented in CR programs such as women, ethnic minorities, and patients with other heart diseases different than AMI (such as patients following coronary revascularization or patients with heart failure) [12]. Another important issue relating to the success of CR programs is adherence to CR programs, and knowledge of dropout-prediction factors. The information on which factors impede adherence is scarce, although a substantial percentage of patients do not finish CR programs despite their relatively short length. The objective of the present research is to determine which factors influence participation and adherence rates in CR programs in patients who have suffered an AMI. We will use a systematic review of the scientific literature to identify these factors. 2. Methods We designed a systematic review of the scientific literature using the Medline (PubMed), EMBASE, and Cochrane databases. We used a combination of MeSH terms and free text words, using the following combination: ((“cardiac rehabilitation program”) OR (“cardiac rehabilitation”)) AND (“myocardial infarction” OR “Myocardial Infarction”[Mesh]). A similar approach was used for EMBASE and Cochrane. We decided to perform an exhaustive search instead of a specific search in order to obtain all relevant papers without losing any relevant information. We followed the PRISMA recommendations in performing our systematic review [13]. We limited our search to those studies performed in humans, published after 01/01/ 2004, and published in English or Spanish. The initial bibliographic search was performed on May 15th, 2015 and was updated until June 10th, 2016.

437

possible given the differences of the factors assessed in the available studies and also the heterogeneity of inclusion and exclusion criteria in each of the included studies.

3. Results 3.1. Studies included in the systematic review We found 677 studies using our search criteria, and 29 were included in the final review. Most of these studies were performed in the USA, UK, Canada, Belgium, Denmark, Germany, or Australia. We excluded one systematic review because focused on assessing the effect of different healthcare interventions to raise participation [14]. A flowchart of the search strategy results appears in Fig. 1. A description of the included studies analyzing participation and adherence can be found in Tables 1 and 2, respectively. Sample size of the included studies was highly variable. Most of the included studies had a cohort or crosssectional design. Retrospective cohort studies were based in registries and all of them contained similar information (i.e. variables collected routinely that are used to compare participants vs non participants). Some interesting variables that are not routinely collected and which may be of interest have not been analyzed in many studies. These include as civil status, depression, or living as a couple. Studies assessing adherence to CR programs usually have a lower sample size than those assessing participation. Some clinical trials were performed to assess if different interventions had an impact on adherence to CR programs. Below, we describe the impact of the different factors on participation and adherence. For each factor identified, we first comment on its impact on participation, and then its impact on adherence. 3.1.1. Gender Most studies clearly demonstrate that women participate less in CR programs than men do [8–16]. None of the included studies observed higher participation in women and only one found no differences in participation between the genders [15]. Differences in age of men and women does not appear to establish differences with respect to these programs. Regarding adherence, practically no study has directly analyzed the effect of gender on dropout. Only the studies performed by Suaya et al. and Doll et al. observed a higher adherence for men [16,17].

2.1. Inclusion and exclusion criteria We considered the following inclusion criteria: a) Regarding sample size we only included studies with more than 100 patients. For comparative studies, each group had to have at least 50 patients. For follow-up studies which assessed adherence, at least 100 patients had to have started a CR program; b) Regarding study design we included systematic reviews and meta-analyses (focused on the effect of patients' characteristics), observational studies (such as cohort studies), case–control studies, and cross-sectional studies. Clinical trials were excluded because adherence and participation in a trial setting is not comparable to adherence in everyday practice. Four clinical trials were excluded for this reason. These excluded studies analyzed the role of nursing staff in participation and adherence, and assessed the effect of a different session schedule on adherence; c) Regarding the type of patients included, we considered individuals who had coronary diseases, largely AMI. If the study included a mix of patients with different heart diseases (including AMI), it was also included; d) Regarding patients' characteristics we included patients of both sexes, with no age limitations and with any comorbidity; e) Regarding intervention, we included patients that were able to participate in a CR program; and, finally, f) Regarding the study hypothesis, we included papers that analyzed factors influencing participation or adherence even if the main objective of the study was different. 2.1.1. Exclusion criteria We excluded studies with a lower sample size than that established for the present review, studies with different design (qualitative studies, narrative studies), studies which included patients without coronary events and studies performed exclusively on patients that had not suffered AMI (i.e. patients with heart failure, coronary by-pass, and so on). We also excluded papers analyzing factors influencing indication to participate in a CR program (the so-called ‘referral studies’) as indication to participate is usually given by cardiologists and is influenced by patients' clinical characteristics, with older patients, women, or patients with comorbidities being less frequently referred to CR programs. We present the results broken down by factors affecting participation and adherence and we present a description of the studies in the tables included. Meta-analysis was not

3.1.2. Age Most studies agree that older individuals participate less than younger ones in CR programs [10,18–22]. In fact, the peak participation age is between 50 and 65. Participation declines significantly after 70 years old [18,19]. The decline in participation is even greater after 80. Some studies have observed that participants in CR programs were, on average, 10 years younger than non-participants. Only one study observed no difference in participants' age, compared to non-participants [15]. Only one study analyzed mean age for adherence, showing that patients older than 65 were more adherent than those who were younger [23]. A study by Beckie, at al., showed that younger women had a higher adherence than older women. Only women were included in this study [24]. 3.1.3. Accessibility to CR programs Accessibility plays a key role in participation in CR programs and can be measured by different means such as distance from the nearest CR center, ownership of a car, or possession of a driving license. The included studies have observed that participation is lower for those individuals living farther from the nearest CR center [10,19,20,25,26]. Only one study observed that distance did not influence participation [15]. Regarding transportation, not owning means of transportation, or not having a driving license, were problems for participation in CR programs [18,20,22,27]. Hansen et al. observed that patients with transportation difficulties had less adherence to CR programs [28].

438

A. Ruano-Ravina et al. / International Journal of Cardiology 223 (2016) 436–443

Fig. 1. Flowchart describing the papers included and excluded.

3.1.4. Employment status Most of the studies that analyzed this variable, observed that employed individuals had higher participation [15,22,29,30]. No study observed a higher participation rate in unemployed individuals. There is probably an interaction between employment status, age, and socioeconomic status; however, no study has analyzed this interaction. Employment status plays a similar role in adherence to CR programs [18, 31]. One study observed that patients left the program because they had to return to work [28]. 3.1.5. Income, socioeconomic status and health insurance A higher income is usually associated with higher participation and higher education [11,16,19,22,27,32–36]. No study has observed higher participation in patients with lower education, income, or without health insurance. The results on adherence are contradictory. A German study observed that patients of middle and low socioeconomic status had higher adherence than those with higher socioeconomic status [9]. A study by Suaya et al. observed that Medicaid patients received fewer sessions than patients not covered by Medicaid [16]. Parashar et al. observed that patients with higher education had a higher adherence overall [35]. 3.1.6. Comorbidities The included studies show that patients with fewer comorbidities had higher participation rates [10,11,17,19,27,29,34,35,37], though there are some exceptions. Smith et al. observed that patients with dyslipidemia, obesity, and diabetes participated more [38]. Dunlay et al. pointed out that among CR participants the prevalence of hypertension, diabetes, and COPD was lower than among non-participants. On the other hand, among participants there were more patients with hyperlypidemia [21]. Depression also reduces participation in CR programs. While some studies did not directly measure depression, they asked questions such as “feeling too ill to participate”, with similar implications to depression [11,27,34]. Other studies have observed higher participation in patients presenting better physical, mental, and emotional health [33]. Presence of comorbidities seems to influence drop-out from CR programs. Some studies have observed that diabetics had less adherence [18] [17,37,39].

Nevertheless, a study by Turk-Adawi observed higher adherence among diabetics [23]. Only one study observed that patients with depression experienced greater difficulty in reaching the end of the CR program [31]. 3.1.7. Civil status Practically all included studies showed that living as a couple or being married raises the chances of participation in CR programs [22, 25,26,29,32–35,38], while living alone or being single reduces them. 3.1.8. Other aspects We can divide these other aspects into those related with coronary disease, or into factors related to attitudes and patients' perceptions. Some investigations observed higher participation in patients presenting pain at admission [32], in patients with coronary by-pass [29], in patients that were referred by a cardiologist [34], and for those having higher physical function and functional status [34,35]. Lack of interest or the perception that the program was not going to be useful decreased participation [22]. Cooper et al. observed a higher participation rate in patients who thought that CR was more suitable for younger and active patients. Some patients were worried about the exercise component of CR programs [40]. Smokers had lower participation in CR programs in two studies but had displayed a higher participation in a third study [11,15,17,21]. Some studies observed that smokers had less adherence to CR programs [18,29,35]. Other factors affecting adherence are presence of previous angioplasty [35] and group sessions [23]. 4. Discussion The results of this systematic review clearly highlight that there are common aspects affecting participation and adherence to CR programs. These factors appear homogeneously in most studies, while contradictory results appear sparsely. These variables can be classified as inherent to patient characteristics or related to accessibility to CR practices. Most of them are related to patients' characteristics such as gender, age, comorbidities, disease perception, social class, or education. Regarding accessibility, proximity to a health center offering CR programs or their accessibility, plays an important role in participation and adherence.

A. Ruano-Ravina et al. / International Journal of Cardiology 223 (2016) 436–443

439

Table 1 Studies assessing factors that influence participation in cardiac rehabilitation programs. Study

Design and setting

Sample size

Patients included

Main results

Worcester et al. 2004 [18]

Cohort study, Australia

808 patients, 573 followed up

Patients who suffered AMI, cardiac surgery or angioplasty

Allen et al., 2004 [27]

Cohort study, USA

253 women (43% afroamerican)

Women who had had AMI, coronary by-pass or angioplasty. Afroamerican and caucasian women were compared.

Sundararajan et al., 2004 [19]

Cross-sectional study, Australia

12,821 patients

Individuals with AMI, angioplasty or coronary bypass

French et al., 2005 [15]

Cohort study, United Kingdom

69 participants, 85 non-participants

Patients with AMI

Smith et al., 2006 [38]

Restrospective cohort study, Canada

3536 patients, 2121 participated in CR

All patients that had received coronary bypass were automatically referred to a CR program.

Cooper et al., 2007 [40]

Cohort study, United Kingdom

130 patients

Patients with AMI

Suaya et al., 2007 [16]

Retrospective cohort study, USA

267,427 Medicare patients

Patients with AMI or coronary bypass

Nielsen et al., 2008 [32]

Cohort study, Denmark

200 patients

Individuals with primary AMI between 30 and 69

Cupples et al., 2010 [33]

Cross-sectional study, Northern Ireland

160 participants and 102 non participants

Patients with AMI or other coronary conditions

Brual et al., 2010 [34] Grace et al. [51]

Cross-sectional study, Canada

97 cardiologists were recruited and each of them recruited 25 patients. 1409 patients recruited, of them 470 took part

Patients with coronary disease

Parashar et al., 2012 [35]

Cohort study, USA

1031 non-participants and 439 participants

Patients who have suffered AMI and referred to a CR program

46.9% of patients participated in the program. Predictors of non-participation for males were: having received percutaneous angioplasty, not being able to drive, and being older than 70. For females, the only factor influencing non-participation was being older than 70. Caucasian women were older, had received more years of education, had higher income and less comorbidities. The most frequent reason for non-participation (alongside not being referred) was feeling too ill to participate and to believe it that it was not necessary, followed by being too busy and transportation issues. Those women with less income, afro-american and less educated had a lower participation. Higher participation took place between 50 and 59 years old, decreasing importantly after 70. Age, type of procedure, being male, being married, having no co-morbidities, living close to the hospital and enjoying better socio-economic status were significantly related with greater participation in CR. There were no differences in participation levels regarding illness perception. There were also no differences regarding age (with sub-analysis for those older or younger than 65) or regarding having suffered previous AMI. Mean age of participants was 48, compared to 65 for non-participants. There were no differences regarding distance to rehabilitation center. Smokers and unemployed participated less. Having a job implies a greater participation. Gender or living alone did not influence participation. Out of 18 variables included in a model to predict participation, 11 were significantly associated with it. Men under 70, who spoke English, lived in couple, had dyslipidaemia, obesity and diabetes, participated more. Those who had assisted CR before surgery doubled the chance of assisting later to more CR sessions. The variable with the greatest association with participation was living close to hospital, followed by being younger than 70 and living in couple. Participating patients believe frequently that CR is necessary and understand its function. It is less probable that those who believe that CR is more suitable for younger and more active individuals take part. Patients who are more concerned with exercise or that have barriers to access CR have a lower chance to participate. Men participated more frequently than women in CR programs (22.1% vs 14.3%) and participation was inversely related to age. Distance to the CR center was a predictor in the use of CR, with participation dropping as distance grew. Patients living at a longer distance (last quintile had a 71% lower participation than those living closer (first quintile) had less participation. Individuals with worse socio-economic status participated less in CR. . Individuals with chest pain on admission were more likely to participate. Individuals who lived alone, with lower income and inversion of T wave were less likely to participate.a Participating patients were younger and had a better socioeconomic status. Participating individuals were in good physical shape, good mental status and good emotional status, and therefore had better quality of life. Women participated less. Single individuals and those who did not live as a couple also participated less. Participation levels were associated with younger age, being employed, higher education, higher family income, less comorbidities that affect exercise, less depression symptoms, being married or in legal partnership, greater control over cardiac disease, being referred by a cardiologist, and better functional state. Participation levels diminished with distance to the rehabilitation center and with driving time to the rehabilitation center. There is also a direct relationship between driving time and less participation. Following the first month after AMI, it was shown in the multivariate analysis that women and uninsured patients participated less. Those with hypertension and previous angioplasties also participated less. Those who were married and had higher education and greater physical (continued on next page)

440

A. Ruano-Ravina et al. / International Journal of Cardiology 223 (2016) 436–443

Table 1 (continued) Study

Design and setting

Sample size

Patients included

De Vos et al., 2012 [20]

Cohort study, Belgium

226 patients

Patients with angioplasty, coronary by-pass or AMI

Beauchamp et al., 2013 [29]

Cohort study, Australia

583 individuals, 282 participating.

Patients with AMI, coronary by-pass or angioplasty

Lemstra et al., 2013 [36]

Retrospective cohort study, Canada

Dunlay et al., 2014 [21] [52]

Cohort study, USA, participants recruited between 1987 and 2010

2991 patients, 1569 taking part

Patients with AMI

McKee et al., 2014 [22]

Cohort study, Ireland

1172 patients

Patients who have suffered AMI and eligible for a CR program

van EngenVerheul et al., 2014 [10]

Cohort study, Netherlands

12,201 patients

Patients with multiple coronary syndromes, with coronary intervention or not

Mikkelsen et al., 2014 [25]

Cohort study, Denmark

682 participants

Individuals with AMI, angioplasty, or coronary by-pass

Turk-Adawi et al., 2014 [11]

Retrospective cohort study, USA

6874 patients, 4644 taking part

Individuals with different cardiac pathology (AMI, coronary by-pass and heart failure).

Armstrong et al., 2015 [39]

Retrospective cohort study, Canada

Patients who received a catheterism (n = 13,158)

Chamosa et al., 2015 [26]

Retrospective cohort study, Spain

The participation of diabetics and non-diabetics is compared in the frame of a CR program. 756 patients

Doll et al., 2015 [17] Soo Hoo et al., 2016 [30]

Retrospective cohort study, USA Prospective cohort study

a

Patients with coronary disease, angioplasty or by-pass

36,376 patients 268 patients

Patients diagnosed with AMI, other forms of acute or subacute ischemic heart disease and angina pectoris.

Patients 65 or older presenting AMI. Patients with STEMI

Main results activity tended to participate more. Men participated more than women and participation rate was of 80%. Nonparticipants were older. The most important reasons for not participating were distance to rehabilitation center, lack of confidence on how to manage the disease by themselves, lack of time and lack of transportation. These reasons were different among men and women. Among participants there were more men, they were younger and had suffered more frequently coronary by-pass than non participants. Participants were more frequently employed, had a family history of coronary disease and had a higher frequency of diabetics. Married men had a higher participation then their single counterparts. Patients from low income neighborhoods were less likely to participate than those from greater income neighborhoods. Among participants, 66.7% left the program before finishing it. Participants in CR programs were on average 11 years younger. There were less women (24% less), there was a higher percentage of smokers and a lower percentage of diabetics (11% less), more obese and with less COPD (10% less). Non participants had a longer hospital stay, were discharged more frequently to an intermediate care facility and had more comorbidities. The most frequent reasons for not participating were, in order of importance: lack of interest, aspects related with time or work, transportation and perception that the program was not going to be useful. In a multivariate model, the variables influencing participation were: age (older persons with less participation), gender (men participated more), education (those with higher education had higher participation), employment (employed participated more than unemployed) and having a private insurance (higher participation). Being married had a marginally significant effect. In all cases women participated less than men. Those younger than 70 participated more than those older than 80. Diabetics, those with lung disease or depression participated less than those without comorbidities. Patients living at more distance from the CR center participated less. The most frequent reason for not taking part was lack of time, followed by inadequate physical shape and distance to the CR center. Patients living alone also participated less and had a lower adherence. Those younger (b65 years), with a health insurance and without a depression history participated more. Smokers participated less. Those living closer to the CR center and with parking facilities had higher participation. Diabetics participated less. . Diabetics had a lower chance to start the CR program (52.3% vs 66.6%, p b 0.001). Participation was lower among women, older patients, and retired. Those living at a higher distance (more than 50 km) had also lower participation. Significant predictors of no enrollment were: age, living arrangements (living alone was associated with a much lower chance of participation), travel distance and history of previous MI (this last one only in women). Gender had no effect in the multivariate model. Smoking and diabetes did not influence on participation. Participants were younger, tended to be male, white, non-smokers and had fewer baseline comorbidities. Participants tended to be males, with employment and to have received a post-discharge health visit.

Only multivariate results are presented. These results were the only with statistically significant association.

Participation is clearly lower for women. This could be due to a lower perception of risk or benefit among women than among men. This aspect may be linked to the fact that cardiologists tend to refer men more frequently to these programs [41]. Women also state that they have more family responsibilities (caring for older persons, grandchildren) and less

available time to attend CR programs [16,20]. The same explanations might be applicable to the less adherence observed in CR programs. Participation and adherence decrease significantly with age and this effect is intriguing s older patients generally have more free time as they often do not work [11,19,21]. This may be due to in transportation or

A. Ruano-Ravina et al. / International Journal of Cardiology 223 (2016) 436–443

441

Table 2 Studies assessing adherence to a cardiac rehabilitation program. Study

Design and setting

Sample size

Included patients

Main results

Worcester et al. 2004 [18]

Cohort study, Australia

284 patients

Patients who had AMI, cardiac surgery or angioplasty

Altenhoener et al., 2005 [9]

Cohort study

536 patients, 421 finalized the program

Patients with AMI

Suaya et al., 2007 [16] Casey et al., 2008 [31] Parashar et al., 2012 [35]

Retrospective cohort study, USA Retrospective cohort study, USA Cohort study, USA

267,427 Medicare patients

Patients with AMI or coronary by-pass Patients with coronary pathology Patients who have suffered an AMI and referred to a CR program, 6 months follow-up survey after AMI

Beauchamp et al., 2013 [29] Turk-Adawi et al., 2013 [23]

Cohort study, Australia

Armstrong et al., 2014 [37]

24% left the program. In males, smokers, diabetics and being unemployed at hospital admission were more frequent in those who abandoned the program. For females, the only predictive factor of quitting was being sedentary at the time of hospital admission. Patients with lower and medium socieconomic status had a higher adherence than those with higher socieconomic status (90% vs 78%). Women, older, non-white and Medicaid patients received less CR sessions. Patients with depression and unemployed have lower probabilities of finishing a CR program. After six months from AMI, multivariate analysis showed that older participants and smokers abandoned the program more frequently. Patients with higher economic problems and previous angioplasties also left more frequently. Caucasians and those with higher education had higher adherence Participants with lower adherence had higher smoking prevalence. The result variable was defined as attending to less than half of CR sessions (more than 20). Among patient related factors, those older than 65 had higher adherence, and also white participants and those with coronary by-pass in the highest risk categories, and also diabetics. Organizational aspects associated with adherence were: attending to individual or group sessions on diet, psychological advice in group, advice on medication, advice on lifestyle and relaxation sessions. Other aspects related with participation were the presence of the medical director for more than 15 min/week in the exercise area, to assess patient satisfaction, and having adequate spaces and equipment. 84.9% of non diabetics compared to 79.6% of non diabetics completed the 12 week CR program (p b 0,0001). When men and women were compared, there were differences between diabetics and non diabetics. Drop-outs were higher for women (81.7% vs 72.1%) than for men (86.0% vs 82.5%). Patients with diabetes completed the 12 week CR program less frequently than non diabetics (41.3% vs 56.2%; p b 0.001). Patients with higher adherence tended to be male, less prevalence of comorbid conditions and were more likely to present STEMI. Those who attended more sessions had better adherence to medication.

600 patients 697 non participants and 650 participants

583 individuals, 282 participants. 4412 patients who attended a median of 21 CR sessions. Factors predicting adherence were divided in related to organization and related to patient.

Patients with AMI, coronary by-pass or angioplasty Various types of cardiac pathology (AMI, coronary by-pass and heart failure).

Retrospective cohort, Canada

Adherence of diabetics and non diabetics was compared in a CR program.

Patients with different coronary interventions (percutaneous or not). 7036 diabetics and 1,546 non diabetics

Armstrong et al., 2015 [39]

Retrospective cohort study, Canada

Patients who received a cardiac catheterism (13,158 patients)

Doll et al., 2015 [53]

Retrospective cohort study, USA

Beckie et al., 2015 [24]

Retrospective cohort study, USA

Adherence of diabetics and non diabetics was compared in a CR program. Participants were divided in those who attended more or less than 26 CR sessions. Factors influencing attendance were analyzed. Participants were divided regarding their age (younger or older than 55).

Retrospective cohort study, USA

accessibility to CR facilities or to perception of benefit from CR programs. The cardiologist should play an active role, disregarding the idea that exercise might be harmful, or that it is more suitable for younger patients. This is an argument used by older persons when rejecting participation [40]. Paradoxically, employed patients have a higher participation compared to unemployed patients [29,31]. We could expect that unemployed patients had a higher participation as they have more time to attend these programs. A possible explanation might be related with a higher perception of wellbeing for patients returning to work compared to those who are unemployed. A better mental status (with a lower chance of depression) could provide an alternative explanation for these results. Education level is also positively related to participation [35]. Patients with more comorbidities participate less [21,34], even though they would probably derive greater benefit from CR programs. Health professionals should make an effort in showing them the benefits of the program and that it will not impact negatively on its health.

Patients who have suffered AMI and attended more than 1 session of CR (11,862 patients).

Patients surviving acute coronary syndromes (252 patients).

Characteristics of younger women were different than those of older women. A higher percentage of younger women did not complete CR sessions. Younger women had a worse quality of life.

Comorbidities can also reduce accessibility to CR programs. There is very little information available regarding participation rates of specific subpopulations with diabetes or COPD who have suffered an AMI, but evidence suggests that their participation rates may be even lower. It is urgent to have studies disentangling the causes of this low participation rate and to test new interventions or accommodations of conventional CR programs directed to increase participation in these subgroups. Smokers generally have lower participation and adherence to CR programs. A recent systematic review focusing on analysis of smoking and its relationship to CR has concluded that though smokers are referred more frequently to CR programs, they attend CR programs less frequently and have higher dropout rates [42]. A lower participation rate is also observed for patients with diabetes or other comorbidities, meaning that we might say that we have ‘health adherers’ or ‘healthy participants’ in CR programs. In general, there are no important differences in participation among the different countries where research on this issue has been

442

A. Ruano-Ravina et al. / International Journal of Cardiology 223 (2016) 436–443

performed. Factors affecting participation are usually the same in different locations. Some US-based studies have shown differences depend on health coverage, as expected. In certain contexts, there may be differences in participation due to a different prevalence of older drivers, different efforts by scientific societies relating to the importance of referral, attendance of CR programs, or availability of CR centers outside a hospital setting. Similar reasoning can be applied to contextual aspects affecting adherence to CR programs. While programs are more or less homogeneous within and between countries, comorbidities distribution, accessibility issues, or age distribution of participants might vary and therefore have an effect on adherence to these programs. A further problem when assessing factors that influence participation and adherence is that CR programs have been changing quickly in the last few years. This is due mainly to changes in organizational aspects. The places where CR programs are delivered have changed progressively, from hospitals to other small health care facilities, or intermediate care facilities. Recently, some CR programs have been designed to be administered at patients' homes, being supervised through telehealth systems. The introduction of telehealth in CR will probably have an impact in participation and adherence rates, through standard programs or smartphones. Home-based monitoring might increase motivation and attendance. Some studies have shown home-based CR as effective as traditional CR [43–45]. The higher effectiveness of these systems or devices depends not only on ease-to-use but also on accessibility and the participants' skills in using these devices. Some of these changes have been carried out precisely to enhance participation and adherence to CR programs and to provide services that can be adapted to each patient. Besides the development of tablets or smartphones that can guide the CR program, some new roles are appearing in these programs such as primary care physicians, psychologists, nutritionists, and physiotherapists. These new health professionals are progressively taking charge of these programs. The timetable of CR programs is also being extended and CR centers are now closer to patients' homes. All these changes mean that factors affecting participation are likely to be changing. Perhaps this is one of the reasons why a great number of publications dedicated to participation and adherence analysis have been published since 2013. It is also important to mention that there is a large population of patients that have not been offered the possibility to be included in a CR program [46] [41]. All cardiologists should be aware about the importance of these programs, as demonstrated by different studies [47]. They should give patients the opportunity to be included in these programs. This systematic review has strengths and limitations. We have not been able to integrate the results through meta-analysis due to the heterogeneity of the studies included (different sample size, causes for rehabilitation, variables included to assess participation and adherence and so on). We decided to exclude clinical trials because adherence and participation can be higher due to this epidemiological design and also because health professionals' intervention or supervision is tighter than would be in a standard CR setting. The clinical trials excluded focused on the analysis of the effectiveness of different interventions (i.e. nursing staff, different CR schedules, or home-based programs). Regarding sample size, we chose to include studies with 100 or more patients in order to reduce the possibility of bias or unreliable results. Studies with low sample sizes could not provide accurate results when participation or adherence are stratified by gender or age groups (splitting a low sample size). Regarding publication date, we decided to include papers published from 2004 onwards because there were no papers analyzing factors affecting participation and adherence before that year and also because CR programs has become more standardized since that date. Having performed a systematic review with a strong methodology is an advantage in and of itself and allows us to compare the effect of different variables in adherence and participation. These results clearly show that participation in CR programs follows a determined pattern that is very homogeneous even when transposed

to different settings. These results show that if the program is directed to retired men living alone, the chances of participation are very low, and specific approaches should be considered for these patients to achieve satisfactory participation results. Specific approaches seem to be needed in order to increase CR participation in women. The identification of patterns predicting low participation should cause health professionals to reflect on the messages they give to patients. Health professionals should be also aware of which patients are likely to be more reluctant to participate in CR programs so that these professionals may adapt their messages accordingly. Patients have doubts regarding the role of physical exercise after having an AMI, and therefore this aspect of CR programs should be carefully explained to them. These conclusions are also applicable to adherence to CR programs, where the same pattern of participation is generally observed. Health administrations should also make efforts to strengthen participation and adherence to CR programs improving accessibility or the use of new technologies wherever possible [48]. There are also gaps in knowledge. Participation and adherence patterns may differ between Europe and USA due to different health coverage systems. Even among European countries, these patterns may vary due to different attitudes and perceptions of both patients and clinicians advice regarding CR programs. Further research in different contexts is encouraged. In this sense, qualitative studies can provide interesting inputs in specific settings and should be used to study patients' and cardiologists' attitudes towards CR programs. Finally, the use of CR programs for other heart diseases such as heart failure should be increased. Recent research has observed that it underused, though some progress has been made, particularly in Europe [49,50]. Conflict of Interest The authors report no relationships that could be construed as a conflict of interest. Funding EU-CaRE Project. This Project is funded by the Horizon 2020 Framework Program of the European Union (2015–2018; Reference 634439). Acknowledgements We are grateful to Daniel Bromberg, MPH, who has kindly edited this document. References [1] WHO. The top 10 causes of death [Internet]. (cited Oct 21 2015). Available from: http://www.who.int/mediacentre/factsheets/fs310/en/ [2] K. Smolina, F.L. Wright, M. Rayner, M.J. Goldacre, Determinants of the decline in mortality from acute myocardial infarction in England between 2002 and 2010: linked national database study, BMJ 344 (2012) d8059. [3] N.J. Pagidipati, T.A. Gaziano, Estimating deaths from cardiovascular disease: a review of global methodologies of mortality measurement, Circulation 127 (6) (Feb 12 2013) 749–756. [4] European Association of Cardiovascular Prevention and Rehabilitation Committee for Science Guidelines, EACPR, U. Corrà, M.F. Piepoli, F. Carré, P. Heuschmann, et al., Secondary prevention through cardiac rehabilitation: physical activity counselling and exercise training: key components of the position paper from the Cardiac Rehabilitation Section of the European Association of Cardiovascular Prevention and Rehabilitation, Eur. Heart J. 31 (16) (Aug 2010) 1967–1974. [5] M.F. Piepoli, U. Corrà, W. Benzer, B. Bjarnason-Wehrens, P. Dendale, D. Gaita, et al., Secondary prevention through cardiac rehabilitation: from knowledge to implementation. A position paper from the Cardiac Rehabilitation Section of the European Association of Cardiovascular Prevention and Rehabilitation, Eur. J. Cardiovasc. Prev. Rehabil. 17 (1) (Feb 2010) 1–17. [6] E. Galve, E. Alegría, A. Cordero, L. Fácila, J. Fernández de Bobadilla, C. Lluís-Ganella, et al., Update in cardiology: vascular risk and cardiac rehabilitation, Rev. Esp. Cardiol. Engl. Ed. 67 (3) (Mar 2014) 203–210. [7] R.R. West, D.A. Jones, A.H. Henderson, Rehabilitation after myocardial infarction trial (RAMIT): multi-centre randomised controlled trial of comprehensive cardiac rehabilitation in patients following acute myocardial infarction, Heart Br. Card Soc. 98 (8) (Apr 2012) 637–644.

A. Ruano-Ravina et al. / International Journal of Cardiology 223 (2016) 436–443 [8] D. Wood, Is cardiac rehabilitation fit for purpose in the NHS: maybe not, Heart Br. Card Soc. 98 (8) (Apr 2012) 607–608. [9] T. Altenhoener, A. Leppin, G. Grande, M. Romppel, Social inequality in patients' physical and psychological state and participation in rehabilitation after myocardial infarction in Germany, Int. J. Rehabil. Res. 28 (3) (Sep 2005) 251–257. [10] M. van Engen-Verheul, H. de Vries, H. Kemps, R. Kraaijenhagen, N. de Keizer, N. Peek, Cardiac rehabilitation uptake and its determinants in the Netherlands, Eur. J. Prev. Cardiol. 20 (2) (Apr 2013) 349–356. [11] K.I. Turk-Adawi, N.B. Oldridge, S.S. Tarima, W.B. Stason, D.S. Shepard, Cardiac rehabilitation enrollment among referred patients: patient and organizational factors, J. Cardiopulm. Rehabil. Prev. 34 (2) (Apr 2014) 114–122. [12] K. Rees, J. Victory, A.D. Beswick, S.C. Turner, I. Griebsch, F.C. Taylor, et al., Cardiac rehabilitation in the UK: uptake among under-represented groups, Heart Br. Card Soc. 91 (3) (Mar 2005) 375–376. [13] D. Moher, A. Liberati, J. Tetzlaff, D.G. Altman, PRISMA Group, Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement, BMJ 339 (2009) b2535. [14] A.D. Beswick, K. Rees, R.R. West, F.C. Taylor, M. Burke, I. Griebsch, et al., Improving uptake and adherence in cardiac rehabilitation: literature review, J. Adv. Nurs. 49 (5) (Mar 2005) 538–555. [15] D.P. French, R.J.P. Lewin, N. Watson, D.R. Thompson, Do illness perceptions predict attendance at cardiac rehabilitation and quality of life following myocardial infarction? J. Psychosom. Res. 59 (5) (Nov 2005) 315–322. [16] J.A. Suaya, D.S. Shepard, S.-L.T. Normand, P.A. Ades, J. Prottas, W.B. Stason, Use of cardiac rehabilitation by Medicare beneficiaries after myocardial infarction or coronary bypass surgery, Circulation 116 (15) (Oct 9 2007) 1653–1662. [17] J.A. Doll, A. Hellkamp, P.M. Ho, M.C. Kontos, M.A. Whooley, E.D. Peterson, et al., Participation in cardiac rehabilitation programs among older patients after acute myocardial infarction, JAMA Intern. Med. 175 (10) (Oct 2015) 1700–1702. [18] M.U.C. Worcester, B.M. Murphy, V.K. Mee, S.B. Roberts, A.J. Goble, Cardiac rehabilitation programmes: predictors of non-attendance and drop-out, Eur. J. Cardiovasc. Prev. Rehabil. Off. J. Eur. Soc. Cardiol. Work Groups Epidemiol. Prev. Card Rehabil. Exerc. Physiol. 11 (4) (Aug 2004) 328–335. [19] V. Sundararajan, S.J. Bunker, S. Begg, R. Marshall, H. McBurney, Attendance rates and outcomes of cardiac rehabilitation in Victoria, 1998, Med. J. Aust. 180 (6) (Mar 15 2004) 268–271. [20] C. De Vos, X. Li, I. Van Vlaenderen, O. Saka, P. Dendale, M. Eyssen, et al., Participating or not in a cardiac rehabilitation programme: factors influencing a patient's decision, Eur. J. Prev. Cardiol. 20 (2) (Apr 2013) 341–348. [21] S.M. Dunlay, B.J. Witt, T.G. Allison, S.N. Hayes, S.A. Weston, E. Koepsell, et al., Barriers to participation in cardiac rehabilitation, Am. Heart J. 158 (5) (Nov 2009) 852–859. [22] G. McKee, M. Biddle, S. O′Donnell, M. Mooney, F. O’Brien, D.K. Moser, Cardiac rehabilitation after myocardial infarction: what influences patients' intentions to attend? Eur. J. Cardiovasc. Nurs. J. Work Group Cardiovasc. Nurs. Eur. Soc. Cardiol. 13 (4) (Aug 2014) 329–337. [23] K.I. Turk-Adawi, N.B. Oldridge, S.S. Tarima, W.B. Stason, D.S. Shepard, Cardiac rehabilitation patient and organizational factors: what keeps patients in programs? J. Am. Heart Assoc. 2 (5) (Oct 2013), e000418. [24] T.M. Beckie, G. Fletcher, M.W. Groer, K.E. Kip, M. Ji, Biopsychosocial health disparities among young women enrolled in cardiac rehabilitation, J. Cardiopulm. Rehabil. Prev. 35 (2) (Apr 2015) 103–113. [25] T. Mikkelsen, K. Korsgaard Thomsen, O. Tchijevitch, Non-attendance and drop-out in cardiac rehabilitation among patients with ischaemic heart disease, Dan Med. J. 61 (10) (Oct 2014) A4919. [26] S. Chamosa, J.A. Alarcón, M. Dorronsoro, F.J. Madruga, J. Barrera, X. Arrazola, et al., Predictors of enrollment in cardiac rehabilitation programs in Spain, J. Cardiopulm. Rehabil. Prev. 35 (4) (Aug 2015) 255–262. [27] J.K. Allen, L.B. Scott, K.J. Stewart, D.R. Young, Disparities in women's referral to and enrollment in outpatient cardiac rehabilitation, J. Gen. Intern. Med. 19 (7) (Jul 2004) 747–753. [28] D. Hansen, J. Berger, P. Dendale, R. De Rybel, R. Meeusen, Training adherence in early cardiac rehabilitation: effect of exercise session duration, J. Cardiopulm. Rehabil. Prev. 29 (3) (Jun 2009) 179–182. [29] A. Beauchamp, M. Worcester, A. Ng, B. Murphy, J. Tatoulis, L. Grigg, et al., Attendance at cardiac rehabilitation is associated with lower all-cause mortality after 14 years of follow-up, Heart Br. Card Soc. 99 (9) (May 2013) 620–625. [30] S.Y. Soo Hoo, R. Gallagher, D. Elliott, Predictors of cardiac rehabilitation attendance following primary percutaneous coronary intervention for ST-elevation myocardial infarction in Australia, Nurs. Health Sci. 18 (2) (Jun 2016) 230–237. [31] E. Casey, J.W. Hughes, D. Waechter, R. Josephson, J. Rosneck, Depression predicts failure to complete phase-II cardiac rehabilitation, J. Behav. Med. 31 (5) (Oct 2008) 421–431.

443

[32] K.M. Nielsen, O. Faergeman, A. Foldspang, M.L. Larsen, Cardiac rehabilitation: health characteristics and socio-economic status among those who do not attend, Eur. J. Pub. Health 18 (5) (Oct 2008) 479–483. [33] M.E. Cupples, M.A. Tully, M. Dempster, M. Corrigan, D.O. McCall, B. Downey, Cardiac rehabilitation uptake following myocardial infarction: cross-sectional study in primary care, Br. J. Gen. Pract. J. R. Coll. Gen. Pract. 60 (575) (Jun 2010) 431–435. [34] J. Brual, S. Gravely-Witte, N. Suskin, D.E. Stewart, A. Macpherson, S.L. Grace, Drive time to cardiac rehabilitation: at what point does it affect utilization? Int. J. Health Geogr. 9 (2010) 27. [35] S. Parashar, J.A. Spertus, F. Tang, K.L. Bishop, V. Vaccarino, C.F. Jackson, et al., Predictors of early and late enrollment in cardiac rehabilitation, among those referred, after acute myocardial infarction, Circulation 126 (13) (Sep 25 2012) 1587–1595. [36] M.E. Lemstra, W. Alsabbagh, R.J. Rajakumar, M.R. Rogers, D. Blackburn, Neighbourhood income and cardiac rehabilitation access as determinants of nonattendance and noncompletion, Can. J. Cardiol. 29 (12) (Dec 2013) 1599–1603. [37] M.J. Armstrong, B.-J. Martin, R. Arena, T.L. Hauer, L.D. Austford, J.A. Stone, et al., Patients with diabetes in cardiac rehabilitation: attendance and exercise capacity, Med. Sci. Sports Exerc. 46 (5) (2014) 845–850. [38] K.M. Smith, K. Harkness, H.M. Arthur, Predicting cardiac rehabilitation enrollment: the role of automatic physician referral, Eur. J. Cardiovasc. Prev. Rehabil. Off. J. Eur. Soc. Cardiol. Work Groups Epidemiol. Prev. Card Rehabil. Exerc. Physiol. 13 (1) (Feb 2006) 60–66. [39] M.J. Armstrong, R.J. Sigal, R. Arena, T.L. Hauer, L.D. Austford, S. Aggarwal, et al., Cardiac rehabilitation completion is associated with reduced mortality in patients with diabetes and coronary artery disease, Diabetologia 58 (4) (Apr 2015) 691–698. [40] A.F. Cooper, J. Weinman, M. Hankins, G. Jackson, R. Horne, Assessing patients' beliefs about cardiac rehabilitation as a basis for predicting attendance after acute myocardial infarction, Heart Br. Card Soc. 93 (1) (Jan 2007) 53–58. [41] K.G. Aragam, M. Moscucci, D.E. Smith, A.L. Riba, M. Zainea, J.L. Chambers, et al., Trends and disparities in referral to cardiac rehabilitation after percutaneous coronary intervention, Am. Heart J. 161 (3) (Mar 2011) 544–551, e2. [42] D.E. Gaalema, A.Y. Cutler, S.T. Higgins, P.A. Ades, Smoking and cardiac rehabilitation participation: associations with referral, attendance and adherence, Prev. Med. 80 (Nov 2015) 67–74. [43] C.K. Chow, J. Redfern, G.S. Hillis, J. Thakkar, K. Santo, M.L. Hackett, et al., Effect of lifestyle-focused text messaging on risk factor modification in patients with coronary heart disease: a randomized clinical trial, JAMA 314 (12) (Sep 22 2015) 1255–1263. [44] M. Varnfield, M. Karunanithi, C.-K. Lee, E. Honeyman, D. Arnold, H. Ding, et al., Smartphone-based home care model improved use of cardiac rehabilitation in postmyocardial infarction patients: results from a randomised controlled trial, Heart Br. Card Soc. 100 (22) (Nov 2014) 1770–1779. [45] I. Frederix, N. Van Driessche, D. Hansen, J. Berger, K. Bonne, T. Alders, et al., Increasing the medium-term clinical benefits of hospital-based cardiac rehabilitation by physical activity telemonitoring in coronary artery disease patients, Eur. J. Prev. Cardiol. 22 (2) (Feb 2015) 150–158. [46] H. de Vries, H.M.C. Kemps, M.M. van Engen-Verheul, R.A. Kraaijenhagen, N. Peek, Cardiac rehabilitation and survival in a large representative community cohort of Dutch patients, Eur. Heart J. 36 (24) (Jun 21 2015) 1519–1528. [47] P.R. Lawler, K.B. Filion, M.J. Eisenberg, Efficacy of exercise-based cardiac rehabilitation post-myocardial infarction: a systematic review and meta-analysis of randomized controlled trials, Am. Heart J. 162 (4) (Oct 2011) 571–584, e2. [48] A.F. Cooper, G. Jackson, J. Weinman, R. Horne, A qualitative study investigating patients' beliefs about cardiac rehabilitation, Clin. Rehabil. 19 (1) (Jan 2005) 87–96. [49] R.-H. Tu, Z.-Y. Zeng, G.-Q. Zhong, W.-F. Wu, Y.-J. Lu, Z.-D. Bo, et al., Effects of exercise training on depression in patients with heart failure: a systematic review and metaanalysis of randomized controlled trials, Eur. J. Heart Fail. 16 (7) (Jul 2014) 749–757. [50] M.F. Piepoli, S. Binno, U. Corrà, P. Seferovic, V. Conraads, T. Jaarsma, et al., ExtraHF survey: the first European survey on implementation of exercise training in heart failure patients, Eur. J. Heart Fail. 17 (6) (Jun 2015) 631–638. [51] S.L. Grace, S. Gravely-Witte, J. Brual, G. Monette, N. Suskin, L. Higginson, et al., Contribution of patient and physician factors to cardiac rehabilitation enrollment: a prospective multilevel study, Eur. J. Cardiovasc. Prev. Rehabil. Off. J. Eur. Soc. Cardiol Work Groups Epidemiol. Prev. Card Rehabil. Exerc. Physiol. 15 (5) (Oct 2008) 548–556. [52] S.M. Dunlay, Q.R. Pack, R.J. Thomas, J.M. Killian, V.L. Roger, Participation in cardiac rehabilitation, readmissions, and death after acute myocardial infarction, Am. J. Med. 127 (6) (Jun 2014) 538–546. [53] J.A. Doll, A. Hellkamp, L. Thomas, P.M. Ho, M.C. Kontos, M.A. Whooley, et al., Effectiveness of cardiac rehabilitation among older patients after acute myocardial infarction, Am. Heart J. 170 (5) (Nov 2015) 855–864.