Rethinking the problems of adherence to medications

Introduction

Patient adherence to medication regimes has long been acknowledged to pose a significant problem in the practice of medicine, a problem which also gives rise to ethical questions. But what exactly is the problem? By getting clearer about what kinds of problematic consequences arise from poor adherence, how the problems are understood to arise, who is affected by any problems, and how they are affected, we can clarify the ethical issues. It is necessary to recognise the variety of issues that arise in different cases. This paper argues that we should distinguish two distinct ‘problems of adherence’, which produce related but distinct ethical questions. These are first, problems associated directly with adherence, and second, problems associated with measuring and understanding adherence.

Defining adherence

An introduction to the ethical issues can be gained by looking at the associated vocabulary. The problem has previously been spoken of as ‘compliance to treatment’, but this has been roundly rejected for its implications of a passive patient blindly doing what the doctor ordered. ¹ The term ‘adherence’ is the generally preferred term and refers to the degree to which a patient’s actions match agreed recommendations. ² The term ‘concordance’ describes ‘the ideal situation that exists when patient and prescriber have an informed discussion and come to an agreement about what treatment will be and how it will be managed’. ³ But a conscious discomfort with the term ‘adherence’ persists, since whatever term is chosen, it is hard to escape the implication that at root there remains the thought that patients ‘are not doing they are told’, or are ‘not doing what is best for them’. Indeed, much of the ethical debate revolves around attempts to avoid this implication, together with suspicions from some quarters that the real agenda is simply an old-fashioned ‘patient compliance’.^4–6 These issues will be discussed in further detail throughout the paper. It is necessary to look not just at what is said overtly about adherence and how to tackle it, but at what is implied by how the problem is described and explained. It is important to examine not just the formal construction of evidence, argument and claim, but the rhetorical constructions of the issue.

It is worth introducing some complexity. There is not a simple dichotomy between adherence and non-adherence; there are many different ways of failing to stick precisely to a course of treatment. Persistence, defined as the length of time between starting treatment and complete cessation, is to be distinguished from forgotten, missed or late doses, or extra doses. Other patients may fail to start a course of treatment or even fail to fill a prescription. These may all represent very different cases often with different reasons for non-adherence.

Furthermore, there is no such thing as a standard ‘non-adherent’ patient. Research has uncovered numerous reasons for non-adherence, but no one patient group escapes the problem, and it is considered that any patient may fail to be fully adherent at some time or other. Poor adherence can be a problem for many different types of disease, surprisingly even for life-threatening diseases where treatment is readily available. ⁷

What problems are attributed to poor adherence to medications?

Commonly, poor adherence to medicines is framed as meriting attention because of the serious problems which poor adherence brings. These have usually been presented as wastage of money, drugs, time and resources more generally; as lost opportunity for improving health of patients and thus of associated societal benefits; and as increased drain on health resources, for example, from hospital admissions which might have been prevented.^2,8,9 These problems are generally presented as being very severe. Indeed, it has been estimated that addressing the current poor rates of adherence to medications could potentially bring about a greater improvement in health than attempts to develop new medications.^10,11 So, historically, poor adherence has been framed as a problem of waste, inefficiency and lost health opportunities. Indeed, given the wide disparities in health outcomes that exist for those of different socioeconomic status, on the most optimistic view, addressing adherence may go some way towards addressing the relatively poor health of the worst off in society.

Worse, poor adherence may mean that health does not simply fail to improve but actually worsens. This is well recognised in certain special cases, where a patient’s poor adherence might threaten the health of third parties. For example, non-adherence to treatment for infectious diseases may promote their spread and assist the appearance of resistance organisms. ⁷ Additionally, third parties may be indirectly damaged in cases where poor adherence to rejection medication may result in the ‘waste’ of a limited resource, a transplanted organ. ¹² Poor adherence may also damage the individual’s own health, from factors such as toxic rebound, drug escalation and misdiagnosis. For example, a poorly adherent patient may be falsely labelled as having resistant hypertension. ¹³ So, it is important to recognise that the problem of adherence may give rise to different ethical issues of differing levels of urgency in different cases.

This also alerts us to a second problem of adherence: our lack of knowledge about how well patients are adhering to their medications, especially as self-reports are notoriously unreliable. This lack of knowledge is especially worrying considering that even on clinical trials, there is evidence of poor adherence. ¹⁴ Even should one consider that what is being measured is ‘intention to treat’ rather than adherence to treatment itself, a range of potentially useful information about the effects and adverse reactions of drugs is eluding us. This worry expands on consideration of the provision of adaptive licensing of drugs, where good quality information about the drugs in use cannot be obtained in conditions where the extent of adherence is unknown. ¹⁵ Thus, adherence needs to be adequately measured. There indeed have been some improvements, for example, using electronic measuring devices,^7,14 Nonetheless, there is progress still to be made.

In examining the ethical issues, we will thus need to recognise the differences between the two different questions: problems of adherence, per se, and problems about our lack of knowledge about rates of adherence and reasons for differences in adherence behaviour.

It must be noted that the problems which are cited in discussions of adherence critically frame how the resultant ethical issues are understood and addressed. For example, to claim that one is concerned that patients’ autonomy is respected may seem disingenuous, if the problem is framed as being principally about preventing wastage in the health service, or concern for third parties. Indeed, since a patient may have wider concerns than simply attaining optimal health, concern about a patient’s genuine participation in treatment decisions may be questioned if the ‘real’ agenda seems to be promoting that same patient’s health and ensuring adherence.^16,17

How do problems of adherence arise, and how might they be tackled?

It is useful briefly to examine how problems associated with adherence arise, for this also is crucial to understanding how the ethical issues are framed. First, in broad terms, within the compass of whose agency does the problem arise? The simplest account (now widely rejected, at least in theory) is that it is the patient’s problem: poor adherence occurs because the patient does not take medications as they undertook to take them. This is not only problematic in the judgemental view it takes of patients, but it also inaccurately describes a complex situation. Another account is that the problem arises from poor communication between doctor and patient. Although this might be a part of the picture, both accounts focus on individuals, and indeed only on certain aspects of these individuals. A more accurate account looks at the institutional, social and economic settings; the long supply chain; the beliefs and values that the patient and the prescriber bring to the situation; and at what happens following the writing of a prescription. ²

These factors are crucial in understanding the ethical issues and in combating simplistic and unfair attributions of responsibility for health outcomes and resource use. This widened approach also focus attention on the many different steps in the supply chain that must be considered in addressing issues of adherence.

The complexity of the tackling the problem should also be recognised. No one solution has proven effective. Recent meta-analyses raise the hopes that certain interventions may produce good results, including patient education, improved dosing schedules, improved communication and feedback to patients on their adherence patterns. ¹⁴ Adherence behaviour is highly complex and individual, and there is to date a dearth of detailed information to guide specific interventions.^18,19 This, therefore, means that tackling adherence requires improving our understanding of adherence, including detailed understanding of how patients, both as groups and individually, do in fact adhere to medication regimes.

A useful model for advancing understanding is one which distinguishes perceptions and practicalities. ² Poor adherence may be unintentional, where the patient wishes to adhere to treatment but where practical obstacles hamper perfect adherence. In other cases, poor adherence or non-adherence may be intentional, where the patient’s perceptions and beliefs about medicines inform behaviour. The division into ‘intentional’ and ‘unintentional’ may be slightly artificial but nonetheless practically useful. Quite different strategies may need to be adopted in these cases. For instance, where a patient really wants to take medicine but keeps forgetting, more specific instructions, reminders or other technology to improve adherence may be useful, but when the patient is intentionally diverging from full adherence, such techniques may be of scant use.

The beliefs which impact upon adherence may be general beliefs about medicines or specific to the particular medicine. General beliefs about medicines may relate to very widely shared concerns, for instance, about long-term use, side effects or dependence; specific beliefs relate to the particular medicine and the particular condition. A ‘necessity–concerns framework’ has been shown in meta-analysis to be effective in predicting adherence for a range of long-term conditions, where perceptions of personal need for treatment (necessity beliefs) are weighed against concerns about a range of potential adverse consequences. ²⁰

How are the associated ethical issues discussed and addressed?

A primary ethical concern, already alluded to, is that patients who are not fully adherent to medication regimes may be held responsible, blamed, stigmatised or scapegoated.^14,21 This concern focuses on what I have called the first problem of adherence that patients’ behaviour does not fully match agreed medication regimes. There are frequently expressed concerns that emphasis on achieving adherence may place undue moral burdens on patients. It has often been pointed out that patients may have good reasons for failures of adherence; a patient’s overall best interest is not necessarily served by the strategies that optimise their health.^2,16 Attempts to address adherence may be thought to pose the danger of coercion, especially where the notion of an ‘emergency’ is invoked, ²¹ or where the power differentials inherent in medical encounters are not recognised and addressed. ⁴ In illustration, proposals to use financial incentives to adherence have raised concerns that patients may find this coercive or degrading.^22,23 In reply, there has been considerable concern in the adherence debate to address these issues in recent years, as illustrated in the focus on the health beliefs of patients. ³

It is too simple, however, to see the ethical concerns as dealing only with attempts to avoid unfairly imposing moral responsibility or blame upon patients. In contrast, there are claims that patients do in fact have certain duties regarding medication. It has been argued that patients may have duties to adhere to medications once they have agreed to treatment on the basis of promise keeping to the medical professional, out of a duty to take care of their own health, or out of duties to others such as prevention of infection. ²⁴ Others argue that overemphasis on patient shared decision making may impair population health if not kept in balance. ¹⁶

It is interesting that those who are most likely to claim that patients may have duties to adhere to medication regimes seem to be philosophers or ethicists. The medical profession has perhaps striven too hard to appear not to keep blaming or stigmatising poorly adherent patients. However, overt attribution of responsibilities to patients might well be preferable, if discussion outwardly pays lip service to patient’s freedom to act whilst actually disguising an agenda in some ways more consistent with ‘compliance’. To address patient’s putative duties directly may be to accord patients more respect as moral agents with responsibilities than to see them as passively tossed on the oceans of their own ill-formed beliefs until rescued by the intervention of an expert physician. Such messages can be present even if unintended, and in quite subtle forms.

The rhetoric of adherence

If one concern is the quality of communication between professionals and patients, then it must be recognised that much communication is not directly and formally spoken. Hence, examining closely the rhetorical aspects of the debates about adherence to medications could be very important in considering if the goal is really to achieve what the patient has truly agreed, or if we are really in practice considering good old-fashioned compliance. Is concern for adherence about achieving informed consent and behaviour which translates the consent into action, or does it amount to persuasion and cajoling? There have been concerns that the rhetoric of respect for patients’ values and reasons for refusing medication or for less than perfect adherence may not match reality. For example, Segal points out that although the goal of ‘adherence’ or ‘concordance’ may be at the forefront of literature, with a nod given to a patient’s values and beliefs, the literature may focus on examples where the patient clearly has mistaken beliefs. Thus, the picture that is presented is one where the professional is the ‘expert’ who is undoubtedly right, and the patient is mistaken and needs to be persuaded. ⁵

Indeed, medical professionals may set out to achieve patient ‘concordance’, but fail to adhere to this intention. Let us look at another possible example. There is a hope that the necessity-concerns framework could be useful for better understanding medication behaviour, and indeed there seems good reason to consider this framework is a powerful advance. In their meta-analysis article, Horne et al. state the goal as being ‘informed adherence’, and it must be noted that the authors explicitly state that one result may be that the patient decides not to take the medication (hence saving waste as well as resulting in better communication). ²⁰ Notwithstanding, ‘informed adherence’ might be seen sceptically by some as implying that the goal is to make the patient realise that the professional is ‘right’.

If this promising framework is to be used widely, it would be useful to consider the implications of the wording used in addition to the overt intentions of the proponents of such a framework. Although there can be degrees of ‘necessity’, at first sight it strongly implies an absolute, and an objective reality. To talk of a patient’s concerns may be to refer to something objectively tangible, but it also implies that some of these ‘concerns’ may be brushed aside or fully addressed in the light of superior professional knowledge – to achieve the ‘informed’ adherence that is the goal. Of course, sometimes it is perfectly legitimate to address concerns, for example, to explain that there are good reasons to carry on taking certain medications even if one feels well. But if something is labelled as necessity, then there seems only one response – to tackle any obstacles to what is necessary. The rhetoric of the framework then may inadvertently weight the issue in favour of the standard practice of medicine. If only for the sake of avoiding the charge that the real goal is compliance, perhaps the presentation of the framework could be fine tuned. It might also be better to grasp the nettle as some have done, and state more explicitly where persuading patients to adopt certain courses of treatment is a goal. Indeed, as many have argued, in some instances it may be proper to communicate to patients that they have certain duties (even if these duties may be overridden by other concerns). ²⁴

A suggestion from micro-economic theory

Additionally, if taken at face value, a necessity-concerns framework implies a certain simplicity in weighing options which may not fully reflect an appreciation of the patient’s position. A suggestion that the tools of micro-economics be used to model patient’s adherence behaviour may be useful. ¹⁷ On this model, the perceived marginal benefits of treatment are balanced against the perceived marginal costs of treatment. A refinement of a simple balance model is produced, where the nearer one approaches perfect adherence, the higher the relative cost of adherence, because there is less to be gained. For instance, good disease management could be attained by 85% adherence; there may be little to be gained from improving upon 84% adherence; and nothing to be gained from achieving 90% adherence if costs are involved. Indeed, the authors claim that a perfectly compliant patient is likely to be less than fully rational, in ignoring the various costs of adhering. Although this model has yet to be empirically tested, it may further advance the goals of producing an account of adherence behaviour whilst achieving the ethical goal of more fully recognising the position of the patient.

Two problems of adherence

Furthermore, we have so far been looking at the first adherence problem. The second adherence problem – knowledge of the extent of adherence – alerts us to the gaps in our knowledge about medicines, for example, uncertainty concerning the level of adherence needed for full benefit. ¹³ If we do not have adequate information about how well patients are taking their medicines and the different ways in which they deviate from perfect adherence, then there is imperfect empirical evidence upon which to posit perfect adherence as a goal to be pursued. We are on the horns of a dilemma. We cannot advocate full adherence as a rational response to necessity beliefs about medications without further evidence of the effects of different adherence patterns to particular medications. Once patients widely realise the problem of adherence, there will be wider recognition of the gaps in our knowledge about medicine regimes. This then may feed into general health beliefs which impact upon adherence, including common scepticism about pharmaceuticals, which is indeed to this extent corroborated by the very findings of poor adherence.

So, the goal must be not just improving adherence, but improving data about adherence. This will, however, give rise to another range of ethical issues. Patient self-report is very poor, especially for those who are weakly adherent. ⁷ Removing patient blame is a good first step to addressing this problem.^2,3,20 However, other methods for gathering data about adherence may involve electronic methods and other data gathering techniques. These will, however, give rise to further ethical issues. From blame, responsibility and coercion, we journey to potential fears about surveillance and control, and tracking of behaviour. Given the value of information about adherence, we can also ask whether patients have the duty to be as truthful as possible about their medication behaviour.

Ethical complexity also exists in strategies to improve patient’s beliefs about medicines to improve adherence. Is the goal simply to improve patients’ beliefs about medicines, or to communicate beliefs which will improve adherence? Does it matter what exactly these beliefs are? In a psychiatric context, it has been argued that what matters is that patients have beliefs in the benefits of their medication, not that they have comprehensive and true beliefs about their illness and about their medication. ²⁵ But although this may be realistic when facing the challenge of psychosis, is this a sub-optimal goal in other cases? What is it to have well-formed beliefs about the necessity of one’s medications, and what are legitimate concerns about these medications?

These worries need to be explored. They are likely to be very context dependent and to vary highly according to trust and other issues. There may well be highly complex responses. A consideration of these issues is likely to benefit from attending to how an individual may feel quite happy to be observed by those who are trusted, but may feel that observation amounts to surveillance if there are concerns about authority, power and judgement.

Conclusion

There is no one ethical problem of adherence, but a range of issues arising from different problematic aspects of adherence and from different circumstances of poor adherence. Attention to exactly how the problems are presented and discussed can enhance our understanding of the potential issues of power imbalance between prescriber and patient. It is important not just that patients’ viewpoints are understood and respected, but that they are seen to be understood and respected. The very difficulties of tackling adherence can, however, be viewed positively as a great opportunity for grasping the complexity of the practice of medicine, the complexity of the causes and patterns of our beliefs and behaviour and, therefore, the complexity of how we understand and address the ethical issues.