Understanding medication adherence in patients with heart failure: commentary

This invited commentary refers to ‘A predictive model for medication adherence in older adults with heart failure’ by EH Oh et al. https://doi.org/10.1093/eurjcn/zvae021 Adherence to and persistence with prescribed medications remain a substantial problem in patients with cardiovascular conditions such as heart failure (HF). Studies have highlighted a number of contributors affecting medication adherence in patients with HF including psychosocial conditions such as depression, number and type of medications, health literacy, and social support.1,2,3 Understanding the factors that influence medication adherence is important because these become targets for intervention. In the paper by Oh et al.,4 the authors focused on an Information-Motivation-Behavioural skills (IMB) model for medication adherence that has been previously tested in patients with other long-term conditions. Their model included information (measured by knowledge about HF), extrinsic and intrinsic motivation (social support and depressive symptoms), and behavioural skills (assessed by self-efficacy for appropriate medication use) in 153 adults with HF in South Korea. Medication adherence was measured using the Korean version of the Adherence to Refills and Medication Scale (ARMS-K). In their analyses, adherence was moderately or strongly correlated with depressive symptoms, knowledge, self-efficacy, and social support, and explained 44.3% of the variability in adherence scores. When left ventricular ejection fraction (LVEF) was added to the model, 64.4% of the variability was explained. What has this study added to our knowledge of medication adherence in patients with HF and will the results help us to support better adherence? The IMB model and the measures used demonstrated a relationship to medication adherence in this older sample of patients with HF, and the identified associated factors of knowledge, social support, depressive symptoms, and self-efficacy can be addressed in practice. However, there are limitations to the study as acknowledged by the authors: it was cross-sectional reflecting a single point in time so causality cannot be assumed, a convenience sample was recruited, and measures were self-reported with potential for bias in responses. In addition, the sample generally had good adherence with a low standard deviation on the ARMS-K, and reported fairly high self-efficacy, social support, and knowledge. A more diverse sample may have provided different results, although the variables assessed have been shown in other studies to affect adherence. Two important issues should be considered. First of all, what factors were not considered that have been shown to affect adherence? An important factor not analysed was the number and type of medications. This older sample of patients with HF in the study took a mean of eight medications (±4.57), which often means remembering to take various combinations of drugs at multiple times of the day. Few patients (<10%) reported side effects of the medications, which likely encouraged adherence. An analysis from the Norwegian Patient Registry 2014–2020 (n = 54 899 patients with HF) found that adherence to a single HF drug was high in the year following initiation, but patients started on dual and triple therapies had very low levels of adherence (42% and 5%, respectively).2 This finding has substantial implications given the recommendation for rapid up-titration of recommended medications for patients after acute HF hospitalization.5,6 Rapid up-titration may also interact with other factors as patients may have had little time to absorb information, develop confidence about self-management and may not have been assessed for conditions such as depression. Simplifying medication regimens is recommended as a means of improving compliance. In other cardiovascular conditions such as hypertension and coronary heart disease, a fixed dose combination pill or ‘polypill’ has been shown to improve adherence. Among 2004 participants with established cardiovascular disease (CVD) or at risk for CVD those randomly assigned to a polypill had higher adherence compared to usual care (86% vs. 65%) over 15 months. The effect was intensified in those who had lower adherence at baseline so that at 15 months, the rates of adherence for this subgroup were 77% for polypill and 23% for usual care.7 The SECURE trial demonstrated that patients randomized within 6 months post-myocardial infarction to a polypill had higher adherence at 6 and 24 months compared to usual care, and importantly had a significantly lower risk of major adverse cardiovascular effects.8 In HF, ensuring that patients are on all of the recommended therapies is essential, and in future, a polypill may be useful. A recent analysis of the Asian-HF Registry estimated that 70% of patients would be eligible for a fixed dose combination pill based on specific criteria.9 Even if a fixed dose combination pill is not available in heart failure, simplifying the medication regimen (especially for older multimorbid patients) may decrease side effects and drug interactions as well as improve adherence. The complexity of adherence and the multiple factors that may affect it is well illustrated in a European Association of Preventive Cardiology consensus statement.10 They advocate addressing five components: the patient (including psychosocial and cognitive status, self-efficacy, and health literacy), disease or condition(s) related factors such as symptoms and comorbidities, the provider (e.g. communication and time pressures), the therapy and its complexity (multiple medications or polypill, burden of treatment), and the healthcare system (including informal care givers). A case in point is that the model in this study explained more variability in adherence with the addition of LVEF. The consensus statement advocates a multidisciplinary approach identifying patients at risk of non-adherence and implementing strategies for follow-up across time. The multitude of factors that can affect adherence may seem daunting and it is likely that interventions will need to be multifactorial. This relates to the second point, nursing and allied health professional (AHP) research needs to move from cross-sectional observational studies to longitudinal modelling and trials testing interventions to improve adherence. Identifying patients most at risk for non-adherence is important, and studies such as this one by Oh and colleagues are useful in that regard. However, we should be mindful of the multitude of variables that can affect adherence as highlighted by Pedretti et al.,10 and not only focus on patient factors. We need to keep in mind that the healthcare system, providers, the therapy, HF, and associated multimorbidities can impact adherence. Addressing these variables may require complex interventions that includes implementing assessment of these factors in practice and testing methods to affect the behaviour. Complex interventions are challenging to develop and implement and need careful planning and analysis (process and outcome) in order to determine if, how, and why it works, e.g. which components of the intervention and their interaction, delivered in what way by whom are responsible for the outcomes. However, these challenges can be met, and these are the types of health services research in which nurses and AHPs can excel. The importance of adherence, especially given the recent drug trials in heart failure across all levels of LVEF, cannot be overstated and should be a focus for intervention by nurses and AHPs.