Enhancement of Treatment Adherence Among Patients With Bipolar Disorder Martha Sajatovic, M.D. Marilyn Davies, R.N., Ph.D. Debra R. Hrouda, M.S.S.A., L.I.S.W.
Objective: Because about one-third of persons with bipolar illness take less than 30 percent of their medication and because nonadherence is associated with rehospitalization and suicide, the literature was searched to identify controlled studies of enhancement of treatment adherence among persons with bipolar disorder. Methods: Studies published up to October 2003 were evaluated. Those selected for review were controlled trials that used samples of adults with a diagnosis of bipolar disorder and that measured adherence to either mood-stabilizing medication or psychotherapy. Information was extracted on the diagnostic composition and size of the study group, the type and duration of the intervention, the method of measuring adherence, and outcomes. Results: Eleven studies met inclusion criteria. Although the literature on enhancing treatment adherence among persons with bipolar disorder is limited, the existing data are promising and demonstrate development over time in our understanding of how best to manage this illness. Interventions that have been shown to be effective include interpersonal group therapy, cognitive-behavioral therapy, group sessions for partners of persons with bipolar disorder, and patient and family psychoeducation. Effective therapies occur in the context of long-term management of illness that incorporates a good understanding of medications and their risks and benefits as well as education about illness awareness and self-management. The majority of effective therapies feature an interactional component between patients and their care providers or therapists. Conclusions: Adherence to treatment for bipolar disorder may be enhanced by interventions that address issues of appropriately taking medications to manage illness. For optimum outcomes, promotion of adherence must be integrated into the medication management of bipolar illness. (Psychiatric Services 55:264–269, 2004)
B
ipolar disorder is a severe and recurrent condition that affects nearly 2 percent of the adult population in the United States (1). It has been reported that the overall economic burden of bipolar disorder in the United States is $45 billion annually, with $7 billion in direct treatment costs (2,3), and that
the costs of medication and treatment encounters for an individual with bipolar illness exceed $17,000 annually (4). Nonadherence to treatment has been identified as a frequent cause of recurrence or relapse of bipolar disorder (5,6) that is associated with such negative consequences as rehospitalization (7) and suicide
The authors are affiliated with the department of psychiatry at Case Western Reserve University, 11100 Euclid Avenue, Cleveland, Ohio, 44106-5000 (e-mail, martha.saja
[email protected]). 264
PSYCHIATRIC SERVICES
(8). Scott and Pope (9) recently reported that one in three persons with bipolar disorder fail to take at least 30 percent of their medication. Additionally, nonadherence may be a growing problem among patients with bipolar disorder, because the prevalence of atypical forms of the disorder is increasing and more people with bipolar illness have comorbid substance use disorders (5,10). Although drug treatments for bipolar disorder have proliferated rapidly in the past decade (11), a substantial gap exists between the efficacy of known treatments and treatment effectiveness in real-world settings. Guscott and Taylor (12) have noted that the major reason for the discrepancy between efficacy and effectiveness in lithium treatment of bipolar disorder is poor treatment adherence. Interventions that enhance adherence to prescribed medication regimens for bipolar disorder are critically needed (5). Enhancing adherence is a complex clinical challenge for a number of reasons, including the large number of factors that contribute to treatment nonadherence (13), current lack of clarity in defining and monitoring adherence (5), and a paucity of research data on how to best enhance treatment adherence (5,12–14). Numerous factors appear to influence treatment adherence, including patient factors, such as younger age, single status, male gender, and low educational level (5,15,16); the symptoms of illness, such as hypomanic denial (17) and psychosis (18); and comorbid disorders, such as personality disorders (19,20) and substance use disorders (21). Side effects of medications and unfavorable personal attitudes toward
♦ http://ps.psychiatryonline.org ♦ March 2004 Vol. 55 No. 3
treatment may also have a negative impact on treatment adherence (22,23). The effectiveness of psychotherapeutic interventions to enhance adherence will hinge on the ability to address the factors that are changeable and most relevant to treatment adherence in a given individual. These factors will undoubtedly differ depending on the individual’s psychiatric symptoms, medication response patterns, age, gender, and cultural context. Although psychotherapy for bipolar disorder is known to generally improve the outcome of illness (24,25), it has been reported that interventions that focus on treatment adherence may yield positive results in this specific area (26,27). This review article summarizes published controlled studies and reports of interventions that have been evaluated with respect to enhancing treatment adherence among persons with bipolar disorder. Patterns seen across the selected studies are discussed, as are the challenges of assessing treatment adherence. Areas for future research are recommended.
Methods of Measuring Adherence in Studies of Enhancement of Adherence to Treatment for Bipolar Disorder ♦ Self-report: Adherence is evaluated via a Likert scale or by using a standardized tool such as the Medication Compliance Questionnaire (reference 30) ♦ Informant report: Patients are asked to designate an individual who can evaluate their treatment adherence ♦ Physician report: The treating physician estimates the patient’s treatment adherence ♦ Frequency of visits: The patient’s missed appointments are documented in the scheduling record ♦ Serum lithium levels: Tests are performed serially or on a single occasion ♦ Red blood cell lithium level: This test may be done along with serum levels ♦ Chart review: Notations of adherence difficulties are obtained from medical charts ♦ Composite adherence assessment: The clinician uses multiple sources to determine a composite score (reference 26) ♦ Compliance index: An independent judge evaluates all available information and assigns a compliance index rating on a 3-point ordinal scale: compliance, minor noncompliance, and major noncompliance ♦ Compliance: the patient is adhering to the regimen; minor noncompliance: the patient fails to obtain blood level tests, misses appointments without notifying the clinic, or forgets the prescribed dosages; major noncompliance: the patient deviates significantly from the treatment regimen by terminating medication against medical advice, dropping out of treatment, or taking medication too chaotically to maintain adequate blood levels ♦ Mixed methods: Adherence is measured by combinations of selected specific measures (for example, self-report plus serum levels)
Methods
Results
The objective of the review was to identify controlled published studies that evaluated enhancement of adherence in bipolar disorder. A MEDLINE search for studies published in the English language up to October 2003 was undertaken. Key search words included bipolar disorder, manic-depressive disorder, compliance, adherence, and outcome. Proceedings from recent professional meetings were also examined for abstracts evaluating treatment adherence for bipolar disorder. Citations from the articles selected for inclusion were also evaluated for additional studies. To be included in the review, studies had to have samples of adult participants with a diagnosis of bipolar disorder, to use a controlled-trial method, and to assess adherence to either mood-stabilizing medication or psychotherapy. We reviewed articles that met these criteria and extracted information on type and duration of the intervention, diagnostic composition and size of the study group, the method of measuring adherence, and adherence-related outcomes.
Measurement of treatment adherence Eleven studies met the inclusion criteria, and most—seven studies, or 64 percent—were published in the past five years. Methods of measuring adherence-related outcomes varied from self-report to plasma measurement to independent assessment of adherence patterns. The variety of methods reflects the challenges involved in accurately measuring treatment adherence of persons with bipolar illness. Seven studies used serum drug levels, either alone or with other measures, to determine medication adherence. As expected, the earlier reports focused solely on lithium adherence. Reflective of the rapid expansion of the pharmacopoeia of bipolar disorder over the past decade, the most recent study, a report by Colom and colleagues (28) published in 2003, also included an assessment of anticonvulsant medications. None of the studies evaluated use of atypical antipsychotic compounds in the context of adherence assessment. Three of the 11 studies relied solely on patient self-reports (29–31). Al-
PSYCHIATRIC SERVICES
♦ http://ps.psychiatryonline.org ♦ March 2004 Vol. 55 No. 3
though self-reports are easily attainable, it has been suggested that patients’ reports may overestimate adherence (32). However, Lam and colleagues (33) recently noted significant agreement between patients’ reports of adherence and serum levels. In that study, 91 percent of 44 patients who reported good adherence—which was defined as missing their medication fewer than three times a month—also had adequate serum levels of mood stabilizers. Cochran (26) reported six different measures of “medical compliance”: self-report, informant report, physician report, serum lithium levels, chart review, and an independently judged “compliance index.” Miklowitz and colleagues (34) used an independent assessment done by a research assistant who collected and used information from patients’ self-reports and physicians’ observations. The box on this page summarizes the methods used to evaluate treatment adherence in the studies included in this analysis. Although the studies did not specifically use medication refill records or microelectronic devices, they have also been identified 265
as reasonable additional methods to evaluate treatment adherence (5,14). Although no standard has been established for measuring treatment adherence for bipolar disorder, it has been suggested that a mixed method that combines selected measures is most likely to provide the most accurate assessment of treatment adherence among patients with bipolar disorder (5,28,33). General findings The 11 studies that met inclusion criteria are summarized in Table 1. Nine involved randomized clinical studies. The earliest randomized study, by Shakir and colleagues (35), used a mirror image design, which retrospectively evaluated lithium levels for two years before the intervention and compared them with lithium levels after the start of the intervention. The study by Weiss and colleagues (31) recruited patients in sequential blocks to receive either the experimental intervention or treatment as usual. Seven of these 11 studies focused narrowly on medication adherence. Others, such as the study by Cochran (26), viewed treatment adherence more broadly; it included maintenance of a lithium regimen, scheduled blood tests, and attendance at medical appointments. In eight of the studies, samples were relatively small, ranging from 15 to 69 patients. Larger trials were the study by Miklowitz and colleagues (34), which had a sample of 101 patients, and the most recent studies by Colom and colleagues (28), which had 120 participants, and by Lam and colleagues (33), which had 103 participants. Although participants met strict selection criteria for these studies, generalizations are difficult because of the wide disparity in selected populations. For example, Clarkin and colleagues (29) conducted a marital intervention study and selected middle-aged patients with bipolar illness who had been married for an average of 17 years. Perry and colleagues (36) selected patients who had experienced relapse in the 12 months before study entry, and Lam and associates (30) designed their cognitive therapy intervention study for a “very difficult, vulnerable group 266
of patients” who were taking mood stabilizers but who “were still relapsing.” In contrast, Harvey and Peet (37) selected lithium clinic attendees for their study who were apparently more stable than patients in the other studies. The study by Colom and colleagues (28) required participants to be outpatients who were in remission for at least six months before inclusion in the study. Similarly, the recent study by Lam and colleagues (33) recruited only individuals who were not experiencing a mood episode at study entry. The study by Weiss and colleagues (31) recruited only individuals with bipolar disorder and comorbid substance use. Of the 11 studies identified, seven (64 percent) suggested that the chosen psychosocial intervention improved treatment adherence among patients with bipolar disorder. In contrast, van Gent and Zwart (38) did not find significant improvement in medication adherence, measured by serum lithium levels, when outpatients were assigned to group therapy sessions for partners only instead of standard treatment. Perry and colleagues (36) found no difference in treatment adherence between experimental and control groups after an intervention that emphasized early recognition of bipolar symptoms. Weiss and colleagues (31) found that persons with dual diagnoses who received integrated group therapy had significant improvements in substance use compared with controls; however, no difference in medication adherence was noted. Miklowitz and colleagues (34) reported the results of a randomized trial of a nine-month program of family therapy. Although their treatment intervention was associated with lower relapse rates and less symptom severity, it had no main effect on medication adherence. However, it must be noted that in the study by Miklowitz and colleagues (34), the “average” study patient was reasonably adherent to the medication regimen. The study by Colom and colleagues (28) found that lithium adherence was improved with psychoeducation, but adherence to valproate and carbamazepine was not significantly different than that of a control group. PSYCHIATRIC SERVICES
Interventions for enhancement of treatment adherence To enhance treatment adherence among persons with bipolar disorder, the controlled trials reviewed here generally used treatments that were psychoeducational and patient focused. Nearly half the interventions involved a group-, couple-, or family-focused format. Both the intervention and the assessment of adherence may be maximized when family members or patients’ significant others are involved, although sole dependence on partners or significant others appears to produce limited results (38). As might be expected given the chronicity of bipolar disorder, most interventions in the controlled trials were moderate to long term, with a mean of 13 sessions. The intervention for persons who had bipolar disorder and a comorbid substance use disorder that was developed by Weiss and colleagues (31) consisted of 12 sessions in the initial version and was expanded to 20 sessions in the final version. The longitudinal and interactive aspect of the intervention appears critical for effectiveness (39). As has been noted by Mullen and colleagues (40), written information alone is of little benefit. Scott (41) suggested that difficulties with adherence to medication may be tackled by exploring barriers to adherence and by using cognitive and behavioral techniques to enhance adherence—for example, by challenging automatic thoughts about medications. The more recent studies use standardized or manual-driven psychoeducational interventions, with specific formats and goals that can be replicated in other treatment settings. Overall, psychoeducational interventions that focus specifically on bipolar illness appear effective in improving outcomes and treatment adherence, although improved adherence does not necessarily correlate with improved symptoms. Recent studies by Lam and colleagues (33) and Colom and colleagues (28) have minimized the effects of acute symptom state on the evaluation of psychosocial interventions in measuring bipolar outcomes by enrolling only persons whose illness is in remission
♦ http://ps.psychiatryonline.org ♦ March 2004 Vol. 55 No. 3
Table 1
Summary of studies of interventions to enhance treatment adherence among patients with bipolar disorder Study
Design
Intervention
Participants
Outcome
Comments
Shakir et al., 1979
Mirror design
Weekly interpersonal group psychotherapy
N=15 Bipolar I=12 Bipolar II=3
Lithium levels: pretreatment=.53 mEq/L; posttreatment= .94 mEq/L
Average duration of group treatment was 51 weeks; mean number of participants per session was 8.2
Cochran, 1984
Randomized Weekly modified controlled trial cognitive-behavioral intervention for six weeks
N=28; 14 to intervention, 14 to usual care
Significantly enhanced adherence immediately after intervention and at 6 months
Patients who received the intervention stopped lithium less often and were hospitalized less often
Van Gent and Zwart, 1991
Randomized Five theme-oriented controlled trial groups for partners of patients with bipolar disorder
N=26 partners of patients; 14 to intervention, 12 to control group
Nonadherence did not differ between groups
Group support was of benefit to partners. Partners in both groups reported increased well-being
Harvey and Peet, 1991
Randomized Videotaped lecture on N=59; 29 to educontrolled trial lithium, handout, and cational group, home visit 30 controls
Intervention group had Improved adherence was fewer missed lithium associated with improved doses (p
