HOW COMMON IS NON-ADHERENCE IN BD?
Adherence has been defined as “the extent to which a person’s behaviour, taking medication, following a diet, and/or executing lifestyle changes, corresponds with agreed recommendations from a health care provider”. Studies of BD have largely focused on medication non-adherence rather than other aspects of treatment-adherence. Not surprisingly, there is a considerable variation between the rates of such non-adherence ranging from universal adherence in some studies to more or less universal non-adherence in others[10-12]. Much of this variability in rates can be attributed to methodological differences across studies. Adherence has been defined and assessed differently in different studies. Studies also differ in the settings in which they have been conducted (e.g., clinics or community), in their designs (e.g., cross-sectional or longitudinal), in the patient samples included, and in the phase of illness or the duration during which non-adherence has been estimated[13-17]. Extremely high or low rates have usually been obtained from studies with small patient samples and from specialized settings, or from randomized controlled trials of medication efficacy[10-12,18-20]. If these extremes are ignored most studies report non-adherence rates from 20% to 50%, with a few reporting rates from 60% to 70% for all patients with BD[21-23]. This is very similar to the estimates of several reviews on the subject, which conclude that on the average about 40% to 50% of patients with BD (range 9%-66%) do not take their medications regularly[2,4,24-28]. These rates are essentially the same regardless of whether patients are on mood stabilizers or antipsychotics. Moreover, rates of non-adherence seem to have remained unchanged or even increased over the years despite the introduction of several new medications[1,4,13,26,29]. The fact that about half of patients with BD become non-adherent during long-term treatment puts it on par with several other chronic psychiatric and medical disorders[4,9,30-32]. Moreover, a large part of this commonly prevalent non-adherence remains undetected in real-world treatment of BD as well; clinicians appear to be particularly likely to underestimate non-adherence among their patients[17,33-36]. However, the blame for lack of detection in day-to-day practice cannot be entirely laid at the clinician’s door, because it is notoriously difficult to predict who is likely become non-adherent. Non-adherence is obviously a complex phenomenon determined by a multitude of influences. However, despite best efforts it is still unclear as to which of these factors is most critical in determining non-adherence in BD[4,13,26].
DEMOGRAPHIC AND ILLNESS-RELATED DETERMINANTS OF NON-ADHERENCE IN BD
In a seminal article about 40 years ago Jamison et al proposed four mutually interacting domains to explain non-adherence to prescribed lithium among patients with BD. These included factors related to the patient (e.g., demographic characteristics), the illness (e.g., severity), the effect of medications (e.g., side effects) and characteristics of the clinicians (e.g., relationship with patients). These determinants have been subsequently adopted by others working in the field, but some significant additions have been made in each category. For example, patient-related factors have come to include personal attitudes and beliefs about medication-taking in addition to demographic factors. Clinician-related factors has been broadened to include an environmental domain, which encompasses the influence of clinician-patient interactions and health-system related factors on adherence, as well as the influence of the family and significant others in the patient’s life[13,25,28,38,39].
In the 1980s and 1990s research on treatment non-adherence amongst those diagnosed with psychiatric disorders including BD mostly limited itself to examining demographic, clinical and medication-related factors impacting adherence[2,17,38,40]. The exclusive focus on these factors appeared to be driven by biologically and medically orientated conceptualizations of the illness, although the primacy of the biological approach had been the object of criticism for long. Unfortunately, these efforts to predict non-adherence on the basis of demographic and illness-related factors were not very successful[40-44]. This failed quest could be due to several reasons, the principal one being the equivocal and inconsistent results obtained from different studies[40,45,46]. For example, although certain demographic correlates have emerged as likely determinants in some studies[4,26,27,42], several other studies have found no association between demographic variables and non-adherence in BD[17,47-51]. Among individual demographic attributes there has been some evidence for an association of non-adherence in BD with younger age[13,27,39,52,53], minority ethnicity[7,54-58], and social disadvantage[27,34,52,56,57,59], but the evidence for such links is either limited or often contradictory[2,4,34,38,48,49,60]. The role of clinical factors in determining non-adherence in BD seems to have been similarly inconsistent and ambiguous[42,61]. Overall severity of the illness in terms of symptom-burden, greater number of episodes and prior hospitalizations appeared to influence non-adherence in some[25,27,28,35,62], but not all studies[2,16,47,51,59]. A majority of studies among patients with BD found that poor insight and denial of the illness was associated with non-adherence[4,24,26,38,63-65]. However, though it might be difficult for a patient to be adherent without a basic level of insight, simply having insight may not be sufficient to ensure adherence[1,16,36,65,66]. The presence of comorbid disorders, particularly substance use disorders has also emerged as a consistent correlate of non-adherence in BD[26,28,39,42,46], but the evidence for associations with most other clinical variables has been either scarce or inconsistently replicated[35,62,67-70]. Among medication-related factors the role of the efficacy-tolerability balance in determining adherence has been the focus of most studies in BD. A large number of studies have found that treatment side effects negatively influenced adherence in BD[7,35,46,59,70], though many of these have exclusively investigated the side effects of lithium[34,67,71-74]. On the other hand, an almost equally large number of studies and patient surveys have revealed that side effects are not associated with non-adherence in BD[2,4,26,27,39,75]. It appears that fear or concerns regarding side effects rather than their actual prevalence may be more important in determining non-adherence in BD[17,26,41,45,76,77]. The influence of treatment-efficacy on adherence has been examined less often, though some studies suggest that medications alleviating depressive symptoms are more likely to promote adherence[7,51,78-80].
The uncertain influence of clinical and demographic factors on adherence in BD could also be due to the fact that many of the studies examining this association have failed to take into account complex interactions between several such variables, which could eventually determine the risk of non-adherence in BD. For example, the higher prevalence of non-adherence during manic episodes could well be due to a lack of insight or the presence of cognitive impairment during such episodes[1,4,41]. Similarly, the greater risk of non-adherence in men could be related to the more frequent use of substances among them[81,82]. Additionally, the simple and dichotomous categorizations of demographic, clinical and treatment-related factors adopted by these studies ignored the subjective experience of medication-taking and the importance of factors such as attitudes and health-beliefs, which might underlie the link between demographic and clinical factors and non-adherence[28,81,83]. Accordingly, there is ample evidence to suggest that variables such as age, gender, severity of illness, effects of substance use, side effects and other treatment-related factors may impact adherence through their effects on subjective patient experiences such as their attitudes to medications, their knowledge of the illness and the availability of social support[2,19,51,84-88]. Moreover, while certain demographic and clinical variables such as young age, symptom-severity, substance use comorbidity and lack of insight may be useful in delineating groups at high risk for non-adherence, they do not accurately predict non-adherence at the level of the individual patient[11,40,41,45]. Finally, because many of these factors may not be amenable to change they do not provide opportunities for adherence enhancement through targeted psychosocial interventions[48,49]. These limitations of attempting to predict and target non-adherence based on demographic and illness-related factors indicates the need for an alternative perspective on treatment-adherence. The newer perspective lays greater emphasis on the patient’s point of view of medication-taking, while acknowledging that problems with adherence are likely to be determined by complex interactions between the patient, the illness, its treatment and the wider socio-cultural environment in which such treatment takes place[11,27,41,61].
A PARADIGM SHIFT TO A PATIENT-CENTRED APPROACH TO ADHERENCE IN BD
As with other chronic medical conditions, research on predictors of non-adherence in BD over the last two decades has undergone a gradual shift in thinking from an illness-centred to a patient-centred approach[40,89]. In this patient-centred paradigm, adherence is viewed as a dynamic rather than a static process which is influenced by many factors within and outside the patient[2,13,26,42,43]. At the core of this process lie the patients’ abilities to make decisions about their own treatment[11,27,90]. Patients are the “final decision-makers” who have a right to choose whether or not continue treatment based on their own beliefs, personal circumstances and their perceptions of benefits and disadvantages of treatment. This right of patients to have a say in their treatment is acknowledged and prioritized in the patient-centred approach to medication-taking[2,7,27,32,50,90]. The emphasis on the patient’s decision-making prompts a shift in the patient-clinician relationship to one in which both are equal and active collaborators. This approach is in keeping with the move away from earlier compliance-based models to those that place emphasis on concordant relationships between patients and clinicians[1,4,32,50,89,90]. While compliance-based paradigms treated patients as passive recipients of treatments and ignored the centrality of their viewpoints, the currently prevalent adherence- and concordance-based approaches place greater stress on the need for an agreement on decisions regarding treatment between patients and clinicians. The cornerstone of the concordance approach rests on open discussions of mutual views about taking medications, and a shared decision-making alliance between patients and clinicians while retaining the primacy of patients’ choices. This shift in paradigms has been further propelled by the formulation of a number of health-behaviour models, which have been used to explain non-adherence in BD with a certain degree of success[11,41,45,76,90,91]. Though consisting of divergent social, cognitive and behavioural perspectives on adherence, they give central importance to the very same elements such as patients’ attitude and health-beliefs, the treatment-alliance and factors in patients’ immediate environments influencing adherence. Finally, the move from illness-related determinants to patients’ perspectives on adherence has aided the development of several adjunctive psychosocial interventions to enhance medication-adherence in BD. These treatments, which use the framework of a collaborative alliance with patients and families to address non-adherence through educative, cognitive and behavioural means have had some success in optimizing adherence in BD[2,4,26,28,38,61].
PATIENT-CENTRED DETERMINANTS OF NON-ADHERENCE IN BD
Attitudes and beliefs regarding medications
In the research on adherence a distinction is often made between unintentional and intentional non-adherence. Unintentional non-adherence arises from personal or environmental restrictions which hamper medication-taking, while intentional non-adherence arises from patients’ views on medications which affect their willingness to take them[92,93]. In an influential study, Horne and Weinman categorized patients’ beliefs about medications into general beliefs related to the intrinsic nature of medicines and ways in which medicines are used by doctors, and specific beliefs comprised of the perceived necessity of taking medications coupled with concerns about their adverse consequences. Both general health-beliefs and specific beliefs (attitudes) regarding medications have been examined among patients with BD. Barring a few exceptions[10,27,74,80,95], the majority of such studies of patients with BD have found that both health-beliefs and attitudes to medications are associated with non-adherence in BD[2,4,13,26,28]. Some of the studies have actually concluded that adverse attitudes and health-beliefs among patients have a much greater influence on non-adherence than demographic, illness or treatment-related factors such as side effects[17,26,45,51,76]. In other studies attitudinal factors such as doubts about the need for medications, as well as concerns about their adverse effects have been found to account for a substantial proportion of variance in intentional non-adherence[11,19,37,41,87,96]. All kinds of negative attitudes have been found in these studies though the commonest ones appear to be fear of side effects and harm caused by medications[17,19,37,51,75,97], denial of severity of illness and the need for treatment[11,41,46,86,98,99], the negative impact of long-term medications on daily routines and competing priorities of life[37,85,86,98,100,101], perceived ineffectiveness of medications[19,51,69,46,98,102], fears regarding dependence, being controlled or stigmatized by taking medications[35,37,60,75,103], and missing the pleasure experienced during manic episodes because of the mood-controlling effects of medications[37,44,101,104,105]. On the other hand, some patients perceive medications to be helpful and seem to realize the necessity of taking medications to prevent relapses, hospitalizations and other negative consequences[11,69,79,98,106,107]. Attitudes towards medications among patients might be relatively independent of their demographic and clinical characteristics[50,51,108,109], or they might differ according to age and illness-related factors such as the severity of the illness and its course, comorbid substance use and side effects of medications[85,100,110-113]. Additionally, patients’ attitudes are more likely to be influenced by their knowledge of the illness, attitudes among their family members and ethno-cultural groups, the clinician-patient relationship and the overall quality of life among patients[35,60,84,101,103,114]. However, regardless of the substantial evidence in favour of attitudes and health-beliefs influencing adherence behaviour in BD, the number of studies is relatively small. Moreover, because of the cross-sectional designs and the small numbers of patients in most studies it is not possible to make any inferences about causality[41,86].
Apart from attitudes and health-beliefs the other principal influence on non-adherence in BD is the treatment-alliance between the doctors or clinicians and patients. The concept of the treatment-alliance as a collaborative and affective bond between clinicians and patients has its origin in psychoanalytic and psychotherapeutic literature[115,116]. In keeping with the research-evidence on treatment-alliance in psychotherapy and other psychiatric disorders such as schizophrenia[115-117], an effective alliance appears to have a significant influence on treatment-adherence in BD as well. Though research on the influence of treatment-alliance on adherence is relatively scarce, the more or less unequivocal finding from several studies is that a strong therapeutic alliance is associated with improved adherence among patients with BD[18,58,91,114,118-120]. A strong alliance appears to enhance treatment-adherence in BD in several ways such as fostering more positive attitudes to treatment and enhancing the acceptance of treatment among patients[1,2,13,14,26,78]. The importance of a genuinely collaborative alliance in determining adherence is also in accord with the current emphasis on the active involvement of patients in decision-making and concordance-based approaches to understanding adherence in BD[1,13,52,78,121]. Moreover, a patient-centred approach and a collaborative clinician-patient alliance appear to be essential ingredients of all psychosocial interventions designed to enhance adherence in BD[1,7]. Definitions of therapeutic alliance in psychotherapy have three common elements including the collaborative nature of the relationship, the affective bond between patients and therapists, and the patient’s and therapist’s ability to agree on treatment goals and tasks. Research among patients with BD indicates that the very same components of treatment-alliance are intimately related to adherence behaviour. Forging a successful treatment-alliance in BD begins with a two-way, reciprocal communication between the patient and the clinician. There are a number of studies among patients with BD showing that not only do clinicians tend to overestimate the degree of adherence among patients, but there is also considerable discrepancy between clinicians and their patients regarding the reasons for non-adherence as well as the meaning of non-adherence[3,24,43,76,96]. Clinicians might also fail to acknowledge the patient’s concern or distress about long-term treatment. Therefore, clinicians first need to create an atmosphere in which patients are able to freely discuss their problems about taking medications. Clinicians also need to play an active role in attempting to understand the patient’s views on illness and medication-taking. To further this open and active stance clinicians must not only listen more effectively, but also learn to value this communication with patients in order to forge effective links with them[18,26,107,122]. Empathy, compassion and skilful counselling are much valued by patients and positively associated with adherence in BD[120,123]. This bi-directional communication also forms the vehicle for imparting information about the illness and its treatment since patients frequently express the need for such information[75,106,122]. Moreover, information can be used to effectively dispel incorrect beliefs about medications, reduce feelings of stigma and foster positive attitudes to treatment among patients[50,114]. The other necessary component of an effective treatment-alliance is a genuinely collaborative relationship between the patients and clinicians. This collaborative relationship is built on respect for patients’ rights to decide about their own treatment and a shared decision-making process with patients and clinicians as active and equal partners[8,18,121]. Evidence suggests that patients place substantial emphasis on this interactional component of the treatment-alliance and that the degree of agreement between clinicians and patients on decisions regarding treatment is a high priority for patients[18,122]. Adherence is also enhanced when both patients and clinicians agree on their roles and responsibilities within the alliance. Patients also want this interactional relationship to be flexible and responsive to changes in clinical status and their treatment needs. Feelings of trust also help in building a strong treatment-alliance as studies have found that trust in medication, trust in the clinician and absence of negative treatment expectations are all associated with adherence in BD. Other elements of importance are regular contacts and reviews by clinicians, ongoing support and the long-term stability of the clinician-patient relationship[49,58,106,114,124]. Finally, factors such as attitudes of patients about the illness and its treatment, their perception of the clinician, ethnic and cultural backgrounds of patients, expectations of patients and the extent to which these have been met, and personality attributes such as locus of control have emerged as some of the more significant influences on the treatment-alliance[60,86,114,118,125,126].
Knowledge about the illness and its treatment
Another potential determinant of adherence in BD is the knowledge about the illness and its treatment among patients. Lack of such knowledge is widespread and a prevailing source of dissatisfaction among patients with BD[75,101,106,113,127,128]. Patients appear particularly unhappy with the lack of information provided on side effects and other aspects related to medication-treatment[75,86,127-129]. Although enhancing knowledge should improve adherence among patients, results of studies in BD have been somewhat equivocal in this regard. While several studies have found that inadequate knowledge of the illness among patients appears to affect treatment-adherence[19,75,80,95,102,128], quite a few others have concluded that patients’ level of knowledge is not associated with adherence in BD[10,34,51,84,118,130]. Similarly, psychosocial interventions imparting information in an effort to reduce non-adherence have met with mixed success[38,90,109,126,131,132]. There could be several reasons for these unexpected results including factors such as old age and longer duration of treatment, which influence the levels of knowledge and may act as potential confounders[10,19,84,86,130,133]. Insufficient knowledge could be due to insufficient efforts and ineffective means of imparting information by clinicians, as well as cognitive impairments, lack of insight and motivation among patients[127,128,134]. Patient’s perceptions about their need for information and the extent to which these are met also have some bearing on adherence. Provision of information is more likely to be effective only if it is tailored to the specific needs of patients.
The role of the family and significant others
Families influence patients’ medication-adherence in several ways. A disorganized or dysfunctional family environment has been associated with higher prevalence of non-adherence in BD[101,135,136]. A disturbed family atmosphere often leads to non-adherence by reducing the social support available to the patient[108,137]. Such an outcome would be more likely among patients who are more dependent on family members. This probably explains why a number of studies have found that patients with a greater external locus of control are more likely to become non-adherent[50,83,95,126]. Perceived criticism, negative affective responses and stigmatization or rejection within the family are also associated with non-adherence among patients with BD[75,138]. Finally, attitudes and health-beliefs of family members and their knowledge of the illness and treatment have been shown to have a significant influence of the patient’s own beliefs and attitudes. Accordingly, negative attitudes and improper understanding of the illness among family members may affect the patient’s decision whether or not to continue treatment[2,4,75,91,101,138].
Stigma, patient satisfaction and system-related factors
For many people with BD the stigma of having a chronic mental illness and needing to take long-term treatment for it may deter adherence. Consequently, studies among patients with BD have found their perceived sense of stigma to be associated with non-adherence[75,100,135,139]. Feelings of stigma regarding BD and its treatment also appear to promote negative attitudes towards treatment and adversely affect the treatment-alliance. Patient satisfaction with various aspects of treatment has been found to be a determinant of non-adherence in BD. Dissatisfaction with the efficacy of treatment, with lack of information, with clinicians and with the treatment-alliance have all been found to adversely affect adherence in BD[59,80,106]. On the other hand, patients who are contended with their own lives are more likely to adhere with treatment[28,103,140]. Access to treatment and affordability of treatment also constitute significant barriers to continued adherence in BD. Bhugra et al have suggested that only about half of the patients receive appropriate treatment BD because of systemic barriers to gaining access to appropriate care. This has been endorsed by the results of a number of other studies[38,49,80,101,119]. Finally, further elaboration of the concepts of shared decision-making, personal recovery and integrated or collaborative care are being increasingly applied to understand treatment-adherence in chronic medical disorders. However, as of now these concepts have been only sparingly used to explain treatment-adherence in psychiatric disorders such as schizophrenia and depression. They have not yet been widely applied to BD. A discussion of the importance of these factors is thus beyond the scope of this brief review.