Predictors of cardiac rehabilitation referral, enrolment and completion after acute myocardial infarction: an exploratory study

Cardiac rehabilitation (CR) reduces cardiovascular morbidity and mortality and increases quality of life in patients after acute myocardial infarction (AMI) [1‐3]. Therefore, CR is strongly recommended in American and European guidelines [4‐7]. Still, less than half of eligible patients with AMI or other manifestations of coronary artery disease (CAD) participate in outpatient CR [8]. Low participation rates are caused by suboptimal referral by medical professionals and unsatisfactory enrolment of patients after referral [9]. After enrolment, up to one-third of participants do not complete their CR programmes [10]. Even though low CR participation and completion rates are well studied and improved uptake has been proven beneficial in cost-benefit analyses [11, 12], detailed recommendations on strategies to improve these rates are lacking in current guidelines.

Not participating or not completing CR is associated with a wide range of factors at the patient level, healthcare professional level and (healthcare) system level [13, 14]. Examples of these factors include demographic or disease-specific factors (patient level), physician CR endorsement (professional level) and financial or geographic factors (system level). While not participating and dropout of CR are associated with a doubled risk for cardiovascular events or death [15], nonparticipation and dropout are more often seen in older and high risk patients (i.e. patients with cardiovascular risk factors or lower socioeconomic status) [16‐18]. As this may even further increase their cardiovascular risks, it is of eminent importance to understand why and at which times during a CR programme these patients drop out.

Despite a growing body of evidence on interventions to increase CR utilisation, the percentage of participating patients has not increased in Europe in the past 7 years [8]. To identify the weakest link in the CR care pathway—and thus the primary target for improvement—it is important to chart the entire pathway, and not to solely focus on CR referral or completion. Therefore, the objective of this study was to evaluate at which time (from hospital discharge to CR completion) patients drop out of CR, and to assess which patient-related characteristics can predict dropout.

In 2015 and 2016, 666 patients were hospitalised with a NSTEMI (65%) or STEMI (35%). Patients were predominantly male (66%), with a mean age of 69.0 ± 12.8 years. Twenty patients died during hospitalisation and 6 already participated in outpatient CR at the time of hospitalisation. Data of the remaining 640 patients were analysed.

In this study, nonreferral was the most important reason for not completing a CR programme after AMI. One-third of patients were not referred, and although dropout after referral was lower, more than 50% of discharged patients ultimately did not complete a CR programme. We found older age, female sex, traveling distance, NSTEMI, no coronary revascularisation and prior manifestations of CAD to be independently associated with nonreferral, which is in line with previous studies [22, 23].

The reason why several subgroups of patients are referred less often is only partly understood. Several studies have suggested that some physicians believe women or elderly patients will benefit less from CR [24, 25]. For patients with NSTEMI or without coronary revascularisation, the perceived benefit of CR may also be lower. In this case, however, limited capacity at CR facilities or lack of reimbursement for specific subgroups may also lead to lower referral rates [24], although the available evidence provides no reason for not referring these subgroups [26]. For elderly patients, transport and mobility issues may influence CR participation or willingness to be referred, as do social obligations, providing informal care or a belief that they may not benefit from CR. Widespread implementation of fully automatic referral strategies (including a mandatory justification in case a patient is not referred) may reduce selective referral, leading to improved referral rates. Besides, education of both patients and healthcare professionals on the benefits of CR may improve their motivation and attitudes towards CR and may reduce prejudices that CR is less effective in certain subgroups.

The participation rate in our cohort was 58% (371 out of 640 eligible patients started one or more therapies), which is higher than it was in recent studies [8, 18]. This may be due to differences in the definition of participation and the methods of data collection (e.g., EHR or survey data), and the possibility that participation rates have increased over time. Completion rates in our population were comparable to those in previous studies, although different definitions of completion were used [10, 18]. We found older age and worse exercise capacity at baseline to be independently associated with CR dropout. Low exercise capacity may be caused by comorbidity, which we could only partially correct for in our analyses. However, as low exercise capacity is associated with worse prognosis [27], it is even more important for these patients to participate in a multidisciplinary secondary prevention programme.

Although it seems that improving referral should have priority when aiming to increase CR utilisation, optimisation of enrolment, adherence and completion should not be overlooked. A recent Cochrane review, however, has only found limited evidence for interventions improving participation in and adherence to CR [28]. Home-based CR or cardiac telerehabilitation (CTR) is a safe and (cost-)effective alternative to centre-based CR [29], and future research should indicate whether implementation of CTR indeed leads to increased participation and completion rates. Still, budget ceilings may limit this potential increase in CR utilisation, even with full reimbursement for CR. Updated economic analyses, such as those by Frederix [11] and De Gruyter [12], are therefore needed to convince health insurers and policy makers of the benefits and return on investment of CR.

For patients discharged after AMI, the main reason for not participating in CR was nonreferral. To optimise CR utilisation, improvement of referral rates should have priority, focusing on elderly patients and women, and on patients with NSTEMI, without coronary revascularisation or with prior manifestations of CAD. To improve CR enrolment and completion, alternative models of CR, such as CTR, should be considered.