Journal of Affective Disorders 116 (2009) 51–55
Contents lists available at ScienceDirect
Journal of Affective Disorders j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j a d
Understanding medication non-adherence in bipolar disorders using a Necessity-Concerns Framework Jane Clatworthy a, Richard Bowskill b,c, Rhian Parham a, Tim Rank c, Jan Scott d, Rob Horne a,⁎ a b c d
Centre for Behavioural Medicine, School of Pharmacy, University of London, UK Brighton and Sussex Medical School, UK Sussex Partnership NHS Trust, UK Department of Psychiatry, University of Newcastle upon Tyne, UK
a r t i c l e
i n f o
Article history: Received 16 July 2008 Received in revised form 6 November 2008 Accepted 6 November 2008 Available online 19 December 2008 Keywords: Patient compliance Treatment beliefs Attitudes Bipolar disorder Adherence
a b s t r a c t Background: Medication non-adherence is a major problem in patients with severe mental disorders and is associated with poor clinical outcomes and high resource utilization. This study examined the utility of the Necessity-Concerns Framework for understanding patient attitudes towards and levels of adherence with medications prescribed for bipolar disorders. Method: A convenience sample of 223 individuals currently prescribed medication for bipolar disorders, recruited by advertisement in a Manic Depression Fellowship newsletter, completed the Beliefs about Medication Questionnaire and the Medication Adherence Report Scale. Results: Low adherence was reported by 30% (n = 64) and was predicted by greater doubts about personal need for treatment (OR = .50; 95% CI: .31–.82) and stronger concerns about potential negative effects (OR = 2.00; 95% CI: 1.20–3.34). These predictors were independent of current mood state, illness and demographic characteristics. Limitations: Participants were a potentially biased sample of volunteers who had been recruited through a patient organisation newsletter. However, clinical characteristics and adherence rates in this study were similar to those reported in other studies conducted in Europe and the USA. Conclusions: The Necessity-Concerns Framework is a useful theoretical model for understanding key attitudes towards medication in bipolar disorders. Interventions to facilitate adherence should elicit and address patients' beliefs about medication. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Non-adherence with medication in bipolar disorder is prevalent (Lingam and Scott, 2002; World Health Organisation, 2003), costly (Durrenberger et al., 1999; Knapp et al., 2004) and associated with poor clinical outcome (Keck et al.,
⁎ Corresponding author. Centre for Behavioural Medicine, Department of Practice and Policy, School of Pharmacy, University of London, Mezzanine Floor, BMA House, Tavistock Square, London WC1H 9JP, UK. Tel: +44 020 7874 1281; fax: +44 020 7387 5693. E-mail address: [email protected]
(R. Horne). 0165-0327/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jad.2008.11.004
1998; Scott and Pope, 2002). Whilst certain demographic and clinical features may identify some patients who as a group may be at high risk of non-adherence (e.g. young males; Maarbjerg et al., 1988) these general characteristics do not help clinicians to identify accurately which speciﬁc individuals on their caseload are at risk of becoming non-adherent (Horne, 2007). Treatment studies frequently suggest that the side-effect proﬁle of medications is the main cause of nonadherence. However, this is overly simplistic, as nonadherence rates have not changed since the introduction of the ﬁrst psychotropic medications in the 1950s, despite vast numbers of new compounds being marketed (Tacchi and Scott, 2005). Furthermore, when patients are asked directly,
J. Clatworthy et al. / Journal of Affective Disorders 116 (2009) 51–55
side-effects are ranked only 7th as the most important reason for stopping their medications (Morselli and Elgie, 2003). Increasing evidence from other branches of medicine and recent research in psychiatry has indicated that patient beliefs regarding the ‘threat’ posed to them by the disorder and their views about medications (e.g. that they are all potentially harmful) are associated with adherence (Clatworthy et al., 2007; Scott, 2002). However, application of these concepts in day to day psychiatric practice is rare. A major barrier to introducing procedures to identify such beliefs has been the lack of a simple, theoretically driven assessment tool that provides reliable and valid results that are clinically meaningful and readily interpretable by nonresearch staff. An approach which shows promise in a range of persistent medical disorders and in depression (Aikens et al., 2005; Brown et al., 2005; Horne and Weinman, 1999) is the Necessity-Concerns Framework (Horne and Weinman, 1999). This suggests that patients' motivation to begin and continue treatment is inﬂuenced by their beliefs about treatment and how they judge their personal need for treatment relative to their concerns about potential adverse effects. This study is the ﬁrst to estimate how medication adherence is inﬂuenced by the way individuals with bipolar disorders balance perceived need against concerns. 2. Method
2.2.3. Adherence The Medication Adherence Report Scale (MARS; Horne, 1997) is a 5-item self-report measure which has been validated against electronic adherence monitors (Cohen et al., 2008) and has demonstrated good psychometric qualities in a range of illnesses (George et al., 2005; Horne and Weinman, 2002; Mardby et al., 2007). Although self-report ratings of adherence have been validated against more objective adherence measures (Garber et al., 2004; Haynes et al., 1980; Morisky et al., 1986; Scott and Pope, 2002) there is a widely held concern that individuals will over-report their level of adherence because of their desire for social conformity. To reduce social acquiescence, the MARS is prefaced with the following text: “Many people ﬁnd a way of using their medicines which suits them. This may differ from the instructions on the label or from what their doctor had said. Here are some ways in which people have said they use their medicines. For each statement please tick the box that best applies to you”. Participants indicate how often they engage in each of ﬁve non-adherent behaviours (e.g. “I take less than instructed”) on a 1–5 likert scale (always to never). The item scores are summed to indicate overall level of adherence. Patients completed a MARS for each of their medications prescribed for bipolar disorder and a mean score was calculated for each participant. Consistent with previous research (Bowskill et al., 2007; George et al., 2005; Mardby et al., 2007), MARS scores were then dichotomised to give low adherence (LAd, MARS score ≤ 21) and high adherence sub-groups (HAd, MARS score N21).
2.1. Sample With ethical approval, the study methodology was piloted with a small sample (n = 16) of individuals with bipolar disorders. Following minor revisions to the protocol, a convenience sample of individuals prescribed medication for bipolar disorders was recruited via an advertisement placed in a Manic Depression Fellowship: The Bipolar Organisation (MDF) newsletter, distributed to up to 4500 individuals in Spring 2005. Questionnaire booklets were sent to the 259 service users who contacted the research ofﬁce expressing an interest in participating in the project. 2.2. Measures 2.2.1. Background information The booklet asked about demographic and clinical details including age at ﬁrst diagnosis, total number of previous admissions and currently prescribed psychotropic medications. 2.2.2. Current symptoms The Beck Depression Inventory (BDI) (Beck et al., 1961) and the Altman Self-Rating Mania Scale (ASRM) (Altman et al., 1997) were used to measure depressive and manic symptoms respectively. For each questionnaire item, participants are provided with a series of statements (rated 0–3 in the BDI and 0–4 in the ASRM) and are asked to select the statement that best describes how they have been over the preceding week. The 21-item version of the BDI was used, and a total score of ≥14 was regarded as the cut-off for mild depression. The 5-item ASRM was used to measure manic symptoms and a total score of ≥6 was regarded as the cut-off for hypomania.
2.2.4. Treatment beliefs The Beliefs about Medication Questionnaire: Speciﬁc Version (BMQ-Speciﬁc; Horne et al., 1999) is an 11-item questionnaire comprising two scales: a 5-item Necessity scale that assesses perceived personal need for the medication and a 6-item Concerns scale that assesses concerns about potential adverse effects such as dependence or side effects. Participants indicate on a ﬁve point likert scale their agreement with a series of statements, e.g. “Without this medication I would be very ill” (Necessity) or “I sometimes worry about long-term effects of this medicine” (Concerns). Mean Necessity and Concerns scores, ranging from 1–5, were calculated for each participant. Each individual was categorized into one of four groups according to whether their scores on the Necessity and Concerns scales were above or below three (the midpoint score) for each of these scales. The four subgroups represent different attitudes towards medication (Aikens et al., 2005), namely Skeptical (low Necessity, high Concerns), Ambivalent (high Necessity, high Concerns), Indifferent (low Necessity, low Concerns) and Accepting (high Necessity, low Concerns). 2.3. Analysis Statistical analyses were undertaken using SPSS Version 15 (SPSS, 2006). In order to reduce the impact of random missing data, scales were pro-rated so that missing values were replaced with the individual's mean scale score providing the participant had answered at least 80% of the items in the scale. Demographic, clinical and health belief variables were compared between HAd and LAd groups using student's t-tests, Mann–Whitney U tests and chi-square
J. Clatworthy et al. / Journal of Affective Disorders 116 (2009) 51–55
Table 1 Sample characteristics and comparison of high (HAd) and low (LAd) adherers.
Female (n) Married/ with partner (n) Mean age Mean age at diagnosis Median no. hospital admissions Median no. medications prescribed No. prescribed each medication Lithium Anti-convulsants Anti-depressants Atypical anti-psychotics Benzodiazepines/night sedation Typical anti-psychotics Mean BDI score Mean ASRM scale score Mean BMQ Necessity score Mean BMQ Concerns score a
Total sample (n = 223)
HAd (n = 148)a
LAd (n = 64) a
143 (64%) 98 (44%) 48 years (SD = 11.2) 34 years (SD = 10.3) 3 (range = 0–15) 3 (range = 0–9)
94 (65%) 66 (45%) 48 years (SD = 11.0) 34 years (SD = 10.1) 3 (range = 0–13) 3 (range = 1–9)
38 (60%) 26 (41%) 47 years (SD = 11.6) 35 years (SD = 11.5) 2.5 (range = 0–15) 2.5 (range = 1–5)
.54 .85 .33 .36 .37 .27
117 (52%) 101 (45%) 92 (41%) 90 (40%) 52 (23%) 34 (15%) 14.4 (SD = 10.4) 3.8 (SD = 3.5) 3.74 (SD = .7) 2.89 (SD = .8)
87 (59%) 74 (50%) 65 (44%) 62 (42%) 32 (22%) 20 (14%) 13.7 (SD = 10.1) 3.8 (SD = 3.8) 3.8 (SD = .7) 2.7 (SD = .8)
29 (45%) 26 (41%) 26 (41%) 28 (44%) 19 (30%) 14 (22%) 16.1 (SD = 11.2) 4.3 (SD = 3.4) 3.5 (SD = .7) 3.2 (SD = .7)
.07 .21 .66 .80 .21 .13 .15 .38 .01 .00
11 subjects could not be classiﬁed due to insufﬁcient data on adherence.
analyses. Key demographic and clinical characteristics plus those variables shown to have a statistically signiﬁcant associations with adherence (set a priori at p b .05) were then entered into a logistic regression analysis. Chi-squared analysis was used to compare adherence status (HAd v LAd) across the Indifferent, Skeptical, Ambivalent and Accepting attitude groups. 3. Results Completed questionnaires were returned by 223 (86%) of the 259 individuals who were sent study information As shown in Table 1, the mean age of the sample was 48 years (SD = 11.1), 143 (64%) were female and 98 (44%) were married or living with a partner. Data on past psychiatric history revealed that the mean duration of illness was 14.0 years (SD = 10.6) and 90% (n = 195) had one or more hospitalizations. Thirty percent (n = 64) were classiﬁed as low adherers (LAd) according to their MARS scores. The mean score on the BDI was 14.4 (SD = 10.4) and on the ASRM was 3.8 (SD = 3.5) suggesting moderate levels of sub-syndromal symptoms. Forty-nine percent (n = 99) of participants scored above the BDI cut-off for mild depression, whilst 27% (n = 59) scored above the ASRM cut-off for possible hypomania. The median number of medications prescribed per person for
bipolar disorder was three. The most commonly prescribed medications were lithium (52% of participants), anti-convulsant mood stabilizers (45% of participants), anti-depressants (41% participants) and atypical anti-psychotics (40% participants); fewer than 25% participants were prescribed benzodiazepines or typical anti-psychotics. As shown in Table 1, there were no statistically signiﬁcant differences between HAd and LAd groups on any of the demographic, clinical or treatment variables, although there was a trend for more HAd than LAd subjects to be prescribed lithium (p = .07). 3.1. Utility of the Necessity-Concerns Framework in explaining adherence to medication As shown in Fig. 1, compared with the HAd group (Necessity M = 3.82 (SD = .73); Concerns M = 2.76 (SD = .77)), the LAd participants (Necessity M = 3.52 (SD = .72); Concerns M = 3.17 (SD = .70)), had statistically signiﬁcantly lower perceived need for treatment (t (209) = 2.7, p b .01) and greater concerns about treatment (t (209) = 3.7, p b .01). Logistic regression analysis also revealed that patients' beliefs about treatment were associated with reported levels of medication adherence (see Table 2). Low adherence was associated with higher concerns about treatment (OR = 2.00; 95% CI: 1.20–3.34; p b .01), and lower perceived personal need for treatment (OR = .50; 95% CI: .31–.82; p b .01). Table 2 Logistic regression analysis of predictors of low as compared to high adherence.
Demographics Clinical variables
Fig. 1. Differences in treatment perceptions between high and low adherence groups.
Age Gender Age of diagnosis No. drugs prescribed BDI score ASRM score BMQ Necessity BMQ Concerns
.98 1.29 1.02 .87 1.03 1.04 .50 2.00
.94 .62 .99 .66 .99 .95 .31 1.19
1.08 2.72 1.07 1.15 1.06 1.14 .82 3.34
.29 .49 .24 .33 .14 .39 .006 .008
J. Clatworthy et al. / Journal of Affective Disorders 116 (2009) 51–55
Fig. 2. Proportion of the total sample allocated to each attitudinal group deﬁned by perceptions of medications⁎. ⁎Skeptical (low necessity, high concerns), Ambivalent (high necessity, high concerns), Indifferent (low necessity, low concerns) and Accepting (high necessity, low concerns).
3.2. Attitudinal analysis When participants were categorized into Accepting (n = 92, 43%), Ambivalent (n = 89, 42%), Skeptical (n = 19, 8%) and Indifferent (n = 15, 7%) subgroups it was found that 85% subjects were located in the Accepting or Ambivalent groups, indicating high perceived need for treatment (see Fig. 2). A chi-squared analysis demonstrated statistically signiﬁcant differences in the proportion of LAd individuals in the attitudinal sub-groups (X2 = 21.2, df 3, p b .001) with only 14% of individuals in the Accepting group classiﬁed as LAd, whilst 40% of the Indifferent, 41% of the Ambivalent and 53% of the Skeptical sub-groups were classiﬁed as LAd. 4. Discussion This study is the ﬁrst to assess perceived need for treatment and concerns about treatment in individuals with bipolar disorders and to quantify the association between these medication beliefs and adherence. The ﬁnding that nonadherence was related to doubts about personal need for treatment and concerns about potential adverse effects is consistent with studies of other disorders where long-term medication is recommended (Aikens et al., 2005; Brown et al., 2005; Horne et al., 2004; Horne and Weinman, 1999) and provides support for the utility of the Necessity-Concerns Framework. This sample of individuals who accessed the Manic Depression Fellowship newsletter and volunteered for our study may not be representative of patients being treated for bipolar disorder in the UK and there was no independent assessment of the diagnosis. However, the clinical character-
istics, rates of high and low adherence and the pattern of medications prescribed are remarkably similar to a number of studies conducted in the USA and Europe (Colom et al., 2005; Ghaemi et al., 2006; Nolen et al., 2004; Scott et al., 2006; Scott and Pope, 2002). The cross-sectional design of the study means it is only possible to conclude that beliefs about medication are associated with adherence, we cannot infer causality (i.e. that patients' beliefs cause non-adherence). A prospective longitudinal design will be needed to demonstrate whether certain beliefs actually lead to lower adherence with prescribed treatments. Repeated assessments will also be needed to shed light on the consistency of beliefs over time, or to determine if there are changes in attitudes during periods of low disease activity. Whilst it is important to acknowledge these limitations, they do not undermine the key purpose of this study which was to evaluate the relationship between medication beliefs and levels of adherence. Having established a link between these variables, further studies are now justiﬁed to examine the generalizability of these ﬁndings. High rates of adherence were associated with an Accepting attitude to medication (high necessity, low concerns); views that are likely to be consistent with the medical rationale of most of the prescribers (i.e. treatment beneﬁts outweigh risk). However, the majority of patients' did not show concordance with this view. 42% of participants were ambivalent towards medication (high necessity with high concerns). For this group, lower levels of adherence, often characterized by taking some but not all the medication, may represent a logical ‘common sense’ response to a necessityconcerns dilemma (Horne, 2006). 15% of participants did not perceive treatment to be necessary and were categorized as
J. Clatworthy et al. / Journal of Affective Disorders 116 (2009) 51–55
Skeptical (low necessity, high concerns) or Indifferent (low necessity, low concerns). These groups reported signiﬁcantly lower adherence than Accepting patients. Prescribing is unlikely to be associated with adherence unless it incorporates a process of eliciting and responding to individuals' personal beliefs about the treatment. The potential importance of the Necessity-Concerns Framework is that it could provide clinicians in day-to-day practice with a meaningful and readily applicable approach to identifying common barriers to adherence with proposed treatments and offers a template for them to make targeted interventions (Horne, 2006). Role of funding source The research was funded by an unrestricted educational grant from Astra Zeneca, who had no input into the study design, the data collection/analysis/ interpretation, the writing of the report or the decision to submit the paper for publication. Conﬂict of interest Jane Clatworthy, Rob Horne, Tim Rank and Rhian Parham have no other competing interests to report. Jan Scott has received unrestricted educational grants from Astra Zeneca and Jansen Cilag and has received fees for CME talks or attendance at advisory boards from Astra Zeneca, BSM Otsuka, EIi Lilly, GSK, Jansen Cilag, and Sanoﬁ Aventis. Richard Bowskill has received fees for CME talks or attendance at advisory boards from Astra Zeneca, Eli Lilly, BSM Otsuka and GSK.
Acknowledgements We are grateful to the Manic Depression Fellowship: The Bipolar Organisation who allowed us to advertise our research study in Pendulum magazine and to the participants who gave up their time to complete our questionnaire. References Aikens, J.E., Nease Jr., D.E., Nau, D.P., Klinkman, M.S., Schwenk, T.L., 2005. Adherence to maintenance-phase antidepressant medication as a function of patient beliefs about medication. Ann. Fam. Med. 3, 23–30. Altman, E.G., Hedeker, D., Peterson, J.L., Davis, J.M., 1997. The Altman SelfRating Mania Scale. Biol. Psychiatry 42, 948–955. Beck, A.T., Ward, C.H., Mendelson, M., Mock, J., Erbaugh, J., 1961. An inventory for measuring depression. Arch. Gen. Psychiatry 4, 561–571. Bowskill, R., Clatworthy, J., Parham, R., Rank, T., Horne, R., 2007. Patients' perceptions of information received about medication prescribed for bipolar disorder: implications for informed choice. J. Affect. Disord. 100, 253–257. Brown, C., Battista, D.R., Bruehlman, R., Sereika, S.S., Thase, M.E., Dunbar-Jacob, J., 2005. Beliefs about antidepressant medications in primary care patients: relationship to self-reported adherence. Med. Care 43, 1203–1207. Clatworthy, J., Bowskill, R., Rank, T., Parham, R., Horne, R., 2007. Adherence to medication in bipolar disorder: a qualitative study exploring the role of patients' beliefs about the condition and its treatment. Bipolar Disord. 9, 656–664. Cohen, J.L., Mann, D.M., Wisnivesky, J.P., Leventhal, H., Musumeci, T., Halm, E.A., 2008. Assessing the validity of the Medication Adherence Reporting Scale (MARS). Inner City Asthmatic Adults. 10th International Congress for Behavioural Medicine, Tokyo, Japan. Colom, F., Vieta, E., Tacchi, M.J., Sanchez-Moreno, J., Scott, J., 2005. Identifying and improving non-adherence in bipolar disorders. Bipolar Disord. 7 (Suppl 5), 24–31. Durrenberger, S., Rogers, T., Walker, R., de Leon, J., 1999. Economic grand rounds: the high costs of care for four patients with mania who were not compliant with treatment. Psychiatr. Serv. 50, 1539–1542. Garber, M.C., Nau, D.P., Erickson, S.R., Aikens, J.E., Lawrence, J.B., 2004. The concordance of self-report with other measures of medication adherence: a summary of the literature. Med. Care 42, 649–652.
George, J., Kong, D.C., Thoman, R., Stewart, K., 2005. Factors associated with medication nonadherence in patients with COPD. Chest 128, 3198–3204. Ghaemi, S.N., Hsu, D.J., Thase, M.E., Wisniewski, S.R., Nierenberg, A.A., Miyahara, S., Sachs, G., 2006. Pharmacological treatment patterns at study entry for the ﬁrst 500 STEP-BD participants. Psychiatr. Serv. 57, 660–665. Haynes, R.B., Taylor, D.W., Sackett, D.L., Gibson, E.S., Bernholtz, C.D., Mukherjee, J., 1980. Can simple clinical measures detect patient noncompliance? Hypertension 2, 757–764. Horne, R., 1997. The nature, determinants and effects of medication beliefs in chronic illness. PhD Thesis. University of London. Horne, R., 2006. Beliefs and adherence to treatment: the challenge for research and clinical practice. In: Halligan, P.W., Aylward, M. (Eds.), The Power of Belief: Psychosocial Inﬂuence on Illness, Disability and Medicine. Oxford University Press, Oxford, pp. 115–136. Horne, R., 2007. Adherence to treatment, In: Ayers, S., Baum, A., McManus, C., Newman, S., Wallston, K., Weinman, J., West, R. (Eds.), Cambridge Handbook of Psychology, Health & Medicine, 2nd ed. Cambridge University Press, Cambridge, pp. 417–421. Horne, R., Weinman, J., 1999. Patients' beliefs about prescribed medicines and their role in adherence to treatment in chronic physical illness. J. Psychosom. Res. 47, 555–567. Horne, R., Weinman, J., 2002. Self-regulation and self-management in asthma: exploring the role of illness perceptions and treatment beliefs in explaining non-adherence to preventer medication. Psychol. Health 17, 17–32. Horne, R., Weinman, J., Hankins, M., 1999. The beliefs about medicines questionnaire: the development and evaluation of a new method for assessing the cognitive representation of medication. Psychol. Health 14, 1–24. Horne, R., Buick, D., Fisher, M., Leake, H., Cooper, V., Weinman, J., 2004. Doubts about necessity and concerns about adverse effects: identifying the types of beliefs that are associated with non-adherence to HAART. Int. J. STD AIDS 15, 38–44. Keck Jr., P.E., McElroy, S.L., Strakowski, S.M., West, S.A., Sax, K.W., Hawkins, J.M., Bourne, M.L., Haggard, P., 1998. 12-month outcome of patients with bipolar disorder following hospitalization for a manic or mixed episode. Am. J. Psychiatry 155, 646–652. Knapp, M., King, D., Pugner, K., Lapuerta, P., 2004. Non-adherence to antipsychotic medication regimens: associations with resource use and costs. Br. J. Psychiatry 184, 509–516. Lingam, R., Scott, J., 2002. Treatment non-adherence in affective disorders. Acta Psychiatr. Scand. 105, 164–172. Maarbjerg, K., Aagaard, J., Vestergaard, P., 1988. Adherence to lithium prophylaxis: I. Clinical predictors and patient's reasons for nonadherence. Pharmacopsychiatry 21, 121–125. Mardby, A.C., Akerlind, I., Jorgensen, T., 2007. Beliefs about medicines and self-reported adherence among pharmacy clients. Patient Educ. Couns. 69, 158–164. Morisky, D.E., Green, L.W., Levine, D.M., 1986. Concurrent predictive validity of a self-reported measure of medication adherence. Med. Care 24, 67–74. Morselli, P.L., Elgie, R., 2003. GAMIAN-Europe/BEAM survey I—global analysis of a patient questionnaire circulated to 3450 members of 12 European advocacy groups operating in the ﬁeld of mood disorders. Bipolar Disord. 5, 265–278. Nolen, W.A., Luckenbaugh, D.A., Altshuler, L.L., Suppes, T., McElroy, S.L., Frye, M.A., Kupka, R.W., Keck Jr., P.E., Leverich, G.S., Post, R.M., 2004. Correlates of 1-year prospective outcome in bipolar disorder: results from the Stanley Foundation Bipolar Network. Am. J. Psychiatry 161, 1447–1454. Scott, J., 2002. Using health belief models to understand the efﬁcacyeffectiveness gap for mood stabilizer treatments. Neuropsychobiology 46, 13–15. Scott, J., Pope, M., 2002. Self-reported adherence to treatment with mood stabilizers, plasma levels, and psychiatric hospitalization. Am. J. Psychiatry 159, 1927–1929. Scott, J., Paykel, E., Morriss, R., Bentall, R., Kinderman, P., Johnson, T., Abbott, R., Hayhurst, H., 2006. Cognitive-behavioural therapy for severe and recurrent bipolar disorders: randomised controlled trial. Br. J. Psychiatry 188, 313–320. SPSS, 2006. SPSS for Windows, Rel. 15.0.0. SPSS Inc., Chicago. Tacchi, M.J., Scott, J., 2005. Improving Medication Adherence in Schizophrenia and Bipolar Disorders. John Wiley & Sons, Chichester. World Health Organisation, 2003. In: Sabatee, E. (Ed.), Adherence to Long-term Therapies: Evidence for Action. World Health Organisation.