Effectiveness of psychological intervention as add-on to standard cardiac rehabilitation: Time to adopt new methods or keep doing more of the same?

Cardiac rehabilitation (CR) is a multidisciplinary and multifaceted effort focused on physical training, risk factor management and patient education targeted to minimise the impact of heart disease on patient outcomes and restore physical capacity and health as far as it is possible. The evidence for CR is consistent, with significant cardiovascular mortality reductions even in the era of acute revascularisation and improved pharmacotherapy.1 Hence, in the European guidelines on cardiovascular disease prevention, CR is indicated with a class Ia recommendation, which is the highest level of evidence.2,3 In this issue of the journal, Albus et al. published a systematic review and meta-analysis on CR controlled trials and controlled cohort studies to evaluate the additional benefit of psychological interventions, in comparison to exercise-based CR alone, on depression, anxiety, quality of life, cardiovascular morbidity and cardiovascular and all-cause mortality.4 The authors are to be commended for this effort and it provides much food for thought for CR researchers, cardiologists, psychologists and other mental health professionals in the field. In 20 studies of 4450 patients, they showed non-significant trends, with small to moderate effect sizes, for reducing depression and cardiovascular morbidity, with no evident associations with other outcomes. As the authors rightly point out, there is a number of limitations to these findings, including the low to moderate quality of the included studies, and heterogeneity with respect to the interventions in terms of content, duration, intensity – both in the intervention and comparator groups – along with heterogeneity in outcomes and follow-up periods. Furthermore, interpretation of the findings is especially problematic given that the psychological content of both the intervention and comparator groups is difficult to discern, define and quantify. Given the heterogeneity across almost all methodological aspects that need to be considered when designing a randomised controlled trial and the level of quality of the studies, it is important, however, that we do not jump to the conclusion that psychological interventions in the context of CR are not worthwhile. This would be tantamount to throwing away the baby with the bathwater. In 2018, Richards et al. published a systematic review and meta-analysis in this journal5 based on their 2017 Cochrane review6 on the effectiveness of psychological interventions as the sole intervention in patients with coronary heart disease. The systematic review and meta-analysis had similar problems (e.g. low to moderate quality of studies included and heterogeneity across studies, etc.) to the current systematic review and meta-analysis by Albus et al.,4 as outlined in an associated editorial.7 The review and meta-analysis by Richards et al.5 found an effect on cardiovascular mortality, but only small to moderate improvements in anxiety, depression and stress, and no effects on total mortality and major adverse cardiac events. On a broader level, the results of these two systematic reviews and meta-analyses leave the question open as to whether we see a progression in the literature, or whether we are merely producing ‘more of the same’? When we design new studies, we base the power calculation on the effect sizes of existing studies that may in fact be underpowered or not comparable with respect to methodological aspects, such as those differences found by Albus et al.4 From that point of view, we are producing ‘more of the same’ and repeating our mistakes. For example, when comparing the oldest study (i.e. the 1998 study by Black et al.) with the most recent study (i.e. the 2017 study by Peersen et al.) in the review by Albus et al.,4 there is no reduction in the risk of bias ratings, with little overall evidence that later studies have higher quality than earlier ones.7 Taken together, these results indicate that there is an urgent need for higher quality, multicentre and multicountry studies.4 This may also warrant that we look at the methodologies that we use and look to other fields of enquiry to determine if the literature and methods in CR – with or without added psychological intervention – can be enhanced. This may provide a firmer basis for determining the true effectiveness of psychological interventions in CR. The first step could be to adopt the several recommended strategies to reduce the risk of bias in health behaviour change trials. These are summarised by de Bruin and colleagues in an excellent review and case study.8 Adoption of distraction techniques, contamination bias or enhanced intervention fidelity strategies, for example, would greatly enhance future studies. A second step may be to adopt more sophisticated network meta-analytical techniques, which allow not only direct comparisons of interventions, but also indirect comparisons. Indeed, Albus et al. cited an earlier version of this technique in CR,89 but rightly point out that exercise was confounded across comparators. There are currently ongoing network meta-analytical technique, which address outcomes in cardiac patients,10 but also those undergoing CR.11 Again, however, the grouping of studies may be problematic and may lead to considerable heterogeneity, which provides a significant barrier for drawing conclusions about effectiveness. A third area of enquiry which could address the limitations of grouping may be to quantify the content and intensity in both intervention and comparator groups more accurately. This is vital when looking at exercise interventions, as it is known that their efficacy varies considerably with respect to type and intensity.12 However, usual care or standard practice often differ between settings, and if the quality of usual care is high, or if the provision of lifestyle and mood advice is brief but effective, then this may mask the effect of any additional psychological intervention if present. Such effects of standard care have clearly been demonstrated in the area of adherence to highly active antiretroviral therapy adherence interventions,13 with ongoing similar work in smoking cessation.14 Using an established taxonomy and by coding the content of both intervention and comparator groups from protocols and contacting authors to complete checklists, the authors showed that a substantial proportion of interventions (38%) could already in fact be contained within the comparator groups, but this was highly variable among studies. Thus, as standard care comparators increase in complexity and effectiveness, the effects of any (additional) intervention reduces. There are very likely similar issues within the current CR literature, which may indeed confound the findings of Albus et al.4 For example, to what degree do exercise-based CR programmes contain (brief) lifestyle or mood advice, and which intensity levels of exercise were effective for depression, beyond which psychological intervention yields little effect? An in-depth investigation of the content of both intervention and comparator groups could yield more appropriate effect sizes for exercise-based CR versus comprehensive CR. Arguably, it is only meaningful to compare intervention effects in meta-analyses after accounting for variability in comparator groups.13 This information would at a minimum enable us to better estimate the ‘real’ efficacy of psychological interventions, and in future meta-analyses we would also be able to control for this variability.13 An important but different avenue to pursue is to keep working on strategies to optimise uptake in CR (and psychological interventions) and reduce dropout, as dropout from CR is high, with up to 56%.15 Patients who are distressed and suffer from anxiety and depression are more likely to refuse enrolment in a CR programme in the first place and are also more likely to drop out.16 Potentially, this means that studies included in systematic reviews like that of Albus et al.4 tend to include patients who are highly motivated to participate and have less serious mental health problems,17 which will dilute the potential effect of CR combined with psychological intervention. While we take some time to digest the gaps and challenges in the field, as pointed out in the review and meta-analysis of Albus et al.,4 it should be clear that we do not need more meta-analyses if the individual studies will be more of the same, or if we cannot appropriately quantify the content of the intervention of comparator groups. Instead we need to focus our efforts and resources on providing high-quality, high-fidelity, multicentre, international studies. The adoption of standardised therapist-assisted internet interventions that target depression and anxiety in CR populations could be worth a try as an add-on to standard CR. Targeting depression and anxiety in our patients is paramount, as depression comprises a barrier for CR participation, increases the risk of dropout and is a risk factor in its own right.16 Internet-delivered therapy for common mental disorders has been shown to be as effective as face-to-face treatment.18–21 Internet-based programmes could be supported by the use of ecological momentary assessments that measure patients’ mood in real time that could help guide treatment and enhance patient empowerment, while also increasing ecological validity.22 Thus, it is time for CR with the add-on of psychological interventions to adopt more sophisticated study and review designs to determine its true efficacy. The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. The author(s) received no financial support for the research, authorship, and/or publication of this article.