JPCB Journal

1. Introduction

Since early 1970s treatment for the management of bipolar illness was predominantly pharmacotherapy. Although mood stabilizers have proven their effectiveness for a large proportion of patients [1], 75% relapse within 5 years and 37% continue to show disabling mood fluctuations during remission phases, leading to difficulties in carrying out daily activities, disrupting relationships and affecting self-esteem. The literature has largely documented the need for integrated treatment in bipolar disorders associating pharmacotherapy and psychotherapeutic interventions [2-5]. In the past ten years’ psychosocial approaches emerged to contribute, with medication, to the management of bipolar disorders, such as psychoeducation [6], Cognitive and Behavioral Therapy [7-10], interpersonal therapy [11] and family therapy [12]. Major indications for Cognitive and Behavioral Therapies (CBT) include pharmacological non compliance or partial compliance, risk of relapse, and severe interpersonal problems. The majority of these treatments takes place in outpatients’ settings and has been described in different guidelines [7-10]. Their common goals are to improve quality of life and social functioning, reduce the number and severity of episodes, and increase treatment adherence. The specific components of Lam et al. program [9] include psychoeducation, medication compliance, learning cognitive and behavioral techniques to manage prodromes, the establishment of routines and the identification of long term vulnerability factors. Only some of them were assessed in randomized trials [4,13-14].

CBT received much attention during the last two decades [15], and two recent meta-analyses interestingly showed that its impact is either of limited effect [15] or not significant regarding for example recurrence rate [16]. A significant proportion of patients drops out before treatment ends (between 8% and 15%). Premature drop-out from psychological intervention plus medication is a major but neglected topic of research within the context of bipolar disorder. Understanding the factors influencing attrition rate may help clinicians to identify patients at risk of poor outcome (because of premature drop-out) and to potentially adapt the therapy for this more vulnerable group of patients.

The role of personality dimensions in the drop-out rate from psychological intervention has been rarely investigated in mood disorders. Persons, Burns and Perloff [17] assessed predictors of drop-out and outcome in patients receiving cognitive therapy (CBT) for depression. Their results showed that drop-outs were associated with high initial Beck Depression Inventory score. By contrast, a study of Simons et al. [18] found no association between depression scores and attrition rate. Studying 135 depressive patients receiving CBT for depression, Oei and Kazmierczak [19] failed to found an effect of depressive symptoms on attrition rate.

Authors interested in the association between bipolar disorder and attrition rate focused on drug treatment and clinical severity.

By comparing 65 Bipolar I patients, 29 Bipolar II patients, and 37 Cyclothymic patients with or without Substance Use Disorder (SUD), Mazza et al. [20] found no difference between bipolar I patients with or without this comorbidity in drop-out rates from drug treatment. However, drop-outs were significantly associated with low treatment dose. In 1993, Strakowski et al. [21] assessed the impact of mood symptoms and personality at discharge on symptomatic recovery at 6 months in 27 bipolar patients treated for a first manic episode. They found that higher novelty seeking scores at discharge were associated with lower symptomatic recovery at 6 months.

Several studies have investigated personality dimensions as potential markers of bipolar disorder using clinical [22-25] and non-clinical [25-31] populations as control samples. The role of personality seems therefore to be a major topic of interest in the prediction of therapeutic adherence.

In the present study, we tested the role of residual symptoms (depressive and manic) and personality dimensions in predicting drop-out in a sample of bipolar I outpatients who had been consecutively admitted in a specialized unit of CBT for mood disorders. Our hypothesis was that personality dimensions might predict drop-out in an out-patient setting.

2. Method

2.1. Patients

Seventy-five outpatients (males, females), currently meeting criteria for bipolar I disorder according to DSM-IV classifications, participated in this study. All patients were consecutively admitted in a specialized unit of CBT for mood disorders in an outpatient setting. The study was approved by the local ethics committee and written informed consent was obtained from all participants (ClinicalTrials.gov ID: NCT02472483). Patients’ inclusion criteria were being between 18 and 65 years old, and on regular prophylactic medication. Patients’ exclusion criteria were currently in an acute bipolar episode and/or with a primary addictive problem according to the Mini International Neuropsychiatric Interview - MINI [32]. The final sample consisted of N=75 bipolar I outpatients.

2.2. CBT Program for Bipolar Disorder

This program, based on 20 weekly sessions within a 6-month period, includes the following steps:

- A psychoeducational phase with interventions specific to CBT (i.e., inductive and deductive questioning, reformulations), which can address interactive topics such as bipolar disorder, pharmacological treatments, personal and idiosyncratic symptoms of depression and mania. The history of bipolar disorder is reconstructed in a "life chart", which is completed throughout the duration of the therapy. All patients reconstruct their own history of bipolar illness, the benefits of drug compliance and psychological stress factors. The patient identifies that certain situations or cognitive stress promote a manic phase, while others can induce depressive relapse.

- A cognitive and behavioral phase, which allows the patient to identify his mood swings, to detect their origins (environment, personality...), to develop behaviors to address the symptoms of depression or mania, to identify prodromes and personal psychological vulnerabilities.

- A consolidation phase aimed at checking the understanding of tools which were supposed to be learned and addressing potential problems of patients.

2.3.  Definition of Attrition

All patients were classified as drop-outs when they unilaterally decided to interrupt the therapy before the end of the 20 sessions, or when they participated to less than 10 sessions (<50% of sessions) altogether.

2.4.  Measures

2.4.1. Demographic and clinical characteristics

For each patient we collected demographic information such as age and gender. A clinical interview according to the Mini International Neuropsychiatric Interview (MINI) specified the age of onset, the number of previous bipolar episodes (depressive and manic/hypomanic) and the number of hospitalizations or suicide attempts. 

2.4.2. Mood symptoms

The Hamilton Depression Rating Scale - HDRS [33] and the Mania Rating Scale -MRS [34] were used to provide severity ratings of depressive and manic symptoms.

2.4.3. Personality dimensions

The Temperament and Character Inventory-TCI [35] was used to assess personality dimensions. This tool is a self-administered questionnaire completed by the patient and includes 226 items, rated as true or false, measuring four dimensions (Novelty Seeking - NS, Harm Avoidance -HA, Reward Dependence -RD, Persistence - P) and three trait dimensions (Self-Determination - SD, Cooperation – C, and Self-Transcendence -ST).

2.4.4. Statistical analyses

The study was exploratory, based on a sample of convenience and there was no a priori calculation of power and sample size. Parameters to be studied were: 1) the proportion of premature drop-out from CBT; 2) mean differences between completers and drop-outs on clinical characteristics, baseline mood symptoms and personality dimensions and 3) the validity of personality dimensions as predictive of premature drop-out from CBT.

Continuous variables were recorded as means and standard deviations and categorical variables as percentage. Differences between groups (completers vs. drop-outs) in continuous data were analyzed using Analysis of Variance (ANOVA) while differences between groups in categorical variables were analyzed using the chi-square test.

Covariance analyses were carried out to control for a potential effect of clinical characteristics (e.g. age of onset, number of previous manic and depressive episodes, number of previous hospitalizations, number of suicide attempts) and baseline mood symptoms (e.g. HDRS score for depressive symptoms, MRS score for manic symptoms) on personality dimensions. All results were considered to be significant at the 5% critical level (p<0.05).

The validity of personality dimensions as predictive of premature drop-out from CBT was investigated by computing the sensitivity, specificity, and predictive positive and negative values.

To determine the validity of personality dimensions as predictive of premature drop-out from CBT we also performed a Receiver Operating Characteristics (ROC) curve.

3. Results 

3.1. Drop-Out Rate

On the 75 patients included at baseline, 59 (78.7%) completed therapy (completers), and 16 (21.3%) dropped out before the end of the therapy (drop-outs). All completers participated in at least 10 of the 20 weekly sessions. Drop-out patients interrupted their program at session 1 (N=1), session 3 (N=14) and session 4 (N=1). Therefore, all drop-outs occurred before the fifth session.

3.2. Completers vs. Drop-outs

3.2.1. Demographical and clinical characteristics

The proportion of women was high in the two groups (Table 1), however there was no significant difference (χ2=0.01, df=1, p=0.91). The average age of patients did not differ between the two groups (F(1,73)=1.32; p=.25).

The two groups of patients did not show significant differences in age of onset (F(1,68)= 0.10; p=.75), number of previous depressive (F(1,68)= 2.47; p=.12) or manic / hypomanic (F(1,72)= 2.96; p=.09) episodes, and number of hospitalizations (F(1,61)=1.33; p=.25) or suicide attempts (F(1,62)= 2.42; p=.12).

3.2.2. Mood symptoms

Depressive and manic symptoms were low at baseline (Table 1). However, there was a significant group effect of manic symptoms that were higher for the drop-outs group (F(1,73)= 4.67; p=.03). Such difference was not observed for HDRS scores (F(1,73)= 0.30; p=.59).

3.2.3. Personality dimensions

Completers and drop-outs were compared on baseline measures of personality. Differences between groups were significant for Novelty Seeking, with higher values for drop-outs (F(1,73)=9.35; p<0.01) and Harm Avoidance (F(1,73)= 4.04; p=.05).

As clinical and psychological variables were previously found to influence these dimensions, covariance analyses were performed to control for the potential impact of (a) clinical characteristics (e.g. age of onset, number of previous depressive and manic episodes, number of previous hospitalizations, number of suicide attempts) and (b) baseline mood symptoms (HDRS, MRS) on these personality dimensions. After controlling for these clinical and psychological variables, a reduced but still significant effect of group on Novelty Seeking was found (F(1,49)=5.57; p=.02) as on Harm Avoidance (F(1,49)=4.58; p=.04).

3.2.4. Using Novelty Seeking scores to predict drop-outs from CBT among bipolar patients

A 100% sensitivity is reached for Novelty Seeking score ≤ 13 (the related specificity being 11.86%), while the 100% specificity value is obtained for a threshold of over 29 (the corresponding sensitivity being 18.75%). According to the ROC curve, the score of 29 gives the most appropriate threshold, with a specificity of 100.00%, a sensitivity of 18.75%, a positive predictive value of 100.00% of and a negative predictive value of 81.94%.

4. Discussion

Our main finding is that higher novelty seeking scores at baseline predicts drop-out from CBT among bipolar I patients. Previous studies found that high novelty seeking score was predictive of poor outcome in bipolar patients [21] but also that high novelty seeking was a potential marker of bipolar illness [22-24,28]. This dimension could therefore constitute a predictive factor of poor treatment adherence in bipolar disorders, whatever the type of treatment. Interestingly, difference between completers and drop-outs on novelty seeking remained significant after controlling for clinical variables, such as depressive, manic and anxious symptoms. We can therefore propose that some personality dimensions predict drop-out.

Our drop-out rate was higher than those of Lam et al. studies [4,13]. This higher drop-out rate could be partially attributed to the clinical characteristics of our sample. Bipolar I patients admitted to our department had a long history of illness (around 20 years), experienced numerous previous bipolar episodes, often have comorbidities, apart from the ones that are exclusion criteria for CBT, and experienced numerous psychotherapies.

The present results might have some clinical implications. Further understanding of the factors associated with the decision to interrupt treatment may help the clinicians to identify patients at risk, and to adapt their therapeutic strategies to engage the patients in the therapy. Novelty seeking may partially explain why bipolar patients discontinued their therapy before promoting changes. The temperamental characteristics associated with high novelty seeking, such as Impulsiveness may favor the impulsive decision to interrupt the therapy. Previous studies showed that high novelty seekers were more likely to drop-out before the end of the treatment than low novelty seekers [36], and that high score on impulsivity significantly differentiated patients who dropped out from CBT and completers [37]. Therapists may help patients with high novelty seeking dimension to promote the elaboration of different strategies to reduce their emotional and behavioral dysregulation.

This study has some limitations, such as the lack of control for recurrence rate during the 6 months CBT. There was no a priori calculation of power and sample size, this study being based on a sample of convenience. Moreover, some results could partly be explained by the presence of personality disorders, rather than dimensions. Future assessments, including categorical measure of personality disorders are therefore required.

5. Conclusions

To summary, we provide preliminary evidence that high Novelty Seeking (NS) score predicts premature drop-out from CBT for bipolar I outpatients. An appropriate focus in future research is to clarify whether other clinical or psychological characteristics such as severity of the illness or the presence of personality disorders could contribute to the attrition rate from CBT for these patients.

6. Declaration of Interest

The authors declare that there is no conflict of interest relevant to the content of the article.

Table 1: Demographic and clinical characteristics of 75 Bipolar I patients referred to a CBT center, distinguishing completers and drop-outs.

1. Gitlin MJ, Swendsen J, Heller TL, Hammen C (1995) Relapse and impairment in bipolar disorder. Am J Psychiatry 152: 1635-1640.
2. Colom F, Vieta E, Martinez-Aran A, et al. (2003) A randomized trial on the efficacy of group psychoeducation in the prophylaxis of recurrences in bipolar patients whose disease is in remission. Arch Gen Psychiatry 60: 402-407.
3. Frank E, Kupfer DJ, Thase ME, et al. (2005) Two-year outcomes for interpersonal and social rhythm therapy in individuals with bipolar I disorder. Arch Gen Psychiatry 62: 996-1004.
4. Lam DH, Watkins ER, Hayward P, et al. (2003) A randomized controlled study of cognitive therapy for relapse prevention for bipolar affective disorder. Arch Gen Psychiatry 60: 145-152.
5. Miklowitz DJ, Otto MW, Frank E, et al. (2007) Intensive psychosocial intervention enhances functioning in patients with bipolar depression: results from a 9-month randomized controlled trial. Am J Psychiatry 164: 1340-1347.
6. Colom F, Vieta E (2006) Psychoeducation Manual for Bipolar Disorder. Cambridge: Cambridge University Press.
7. Basco MR, Rush AJ (1996) Cognitive-Behavioral Therapy for bipolar disorder. New York: The Guilford Press.
8. Bauer MS, McBride L (2001) Thérapie de groupe pour le trouble bipolaire : une approche structurée, le programme d’objectifs personnels. Paris: Editions Médecine & Hygiène.
9. Lam DH, Jones SH, Hayward P, Bright JA (1999) Cognitive therapy for bipolar disorder: a therapist’s guide for concepts, methods and practice. Chichester: Wiley & Sons.
10. Newman CF, Leahy RL, Beck AT, Reilly-Harrington NA, Gyulai L (2002) Bipolar disorder: a cognitive approach. Washington: American Psychological Association, USA.
11. Frank E, Hlastala S, Ritenour A, Houck P, Tu XM, et al. (1997) Inducing lifestyle regularity in recovering bipolar disorder patients: results from the Maintenance Therapies in Bipolar Disorder Protocol. Biol Psychiatry 41: 1165-1173.
12. Miklowitz DJ, Goldstein MJ (1997) Bipolar Disorder: A Family Focused Treatment Approach. New York: The Guilford Press.
13. Lam DH, Bright J, Jones S, Hayward P, Schuck N, et al. (2000) Cognitive therapy for bipolar illness: A pilot study relapse prevention. Cognitive Therapy and Research 24: 503-520.
14. Lam DH, Hayward P, Watkins ER, Wright K, Sham P (2005) Relapse prevention in patients with bipolar disorder: Cognitive therapy outcome after 2 years. Am J Psychiatry 162: 324-329.
15. Szentagotai A, David D (2010) The Efficacy of Cognitive-Behavioral Therapy in Bipolar Disorder; A Quantitative Meta-Analysis. J Clin Psychiatry 1: 66-72.
16. Lynch D, Laws KD, McKenna PJ (2010) Cognitive behavioural therapy for major psychiatric disorder: does it really work? A meta-analytical review of well-controlled trials. Psychol Med 40: 9-24.
17. Persons JB, Burns DD, Perloff JM (1988) Predictors of dropout and outcome in cognitive therapy for depression in a private practice setting. Cognitive Therapy and Research 12: 557-575.
18. Simons AD, Levine JL, Lustman PJ, Murphy GE (1984) Patient attrition in a comparative outcome study of depression. J Affect Disord 6: 163-173.
19. Oei TP, Kazmierczak T (1997) Factors associated with dropout in a group cognitive behaviour therapy for mood disorders. Behaviour Research and Therapy 35: 1025-1030.
20. Mazza M, Mandelli L, Di Nicola M, Harnic D, Catalano V, et al. (2009) Clinical features, response to treatment and functional outcome of bipolar disorder patients with and without co-occurring substance use disorder: 1-year follow-up. J Affect Disord 115: 27-35.
21. Strakowski SM, Stoll AL, Tohen M, Faedda GL, Goodwin DC, et al. (1993) The Tridimensional Personality Questionnaire as a predictor of six-month outcome in first episode mania. Psychiatry Research 48: 1-8.
22. Mula M, Pini S, Monteleone P, Preve M, Tortorella A, et al. (2008) Different temperament and character dimensions correlate with panic disorder comorbidity in bipolar disorder and unipolar depression. J Anxiety Disord 22: 1421-1426.
23. Nowakowska C, Strong CM, Santosa CM, Wang PW, Ketter TA (2005) Temperamental commonalities and differences in euthymic mood disorder patients, creative controls, and healthy controls. J Affect Disord 85: 207-215.
24. Young TL, Bagby RM, Cooke RG, Parker JDA, Levitt AJ, et al. (1995) A comparison of tridimensional personality questionnaire dimensions in bipolar disorder and unipolar depression. Psychological Research 58: 139-143.
25. Zaninotto L, Souery D, Calati R, Di Nicola M4, Montgomery S, et al. (2015) Temperament and character profiles in bipolar I, bipolar II and major depressive disorder: impact over illness course, comorbidity pattern and psychopathological features of depression. J Affect Disord 184: 51-59.
26. Engström C, Brändström S, Sigvardsson S, Cloninger R, Nylander PO (2004) Bipolar disorder I. Temperament and character. J Affect Disord 85: 131-134.
27. Loftus ST, Garno JL, Jaeger J, Malhotra AK (2008) Temperament and character dimensions in bipolar I disorder: a comparison to healthy controls. J Psychiatr Res 42: 1131-1136.
28. Nery FG, Hatch JP, Glahn DC, et al. (2008) Temperament and character traits in patients with bipolar disorder and associations with comorbid alcoholism or anxiety disorders. J Psychiatr Res 42: 569-577.
29. Osher Y, Cloninger CR, Belmaker RH (1996) TPQ in euthymic manic depressive patients. J Psychiatr Res 30: 353-357.
30. Osher Y, Lefkifer E, Kotler M (1999) Low persistence in euthymic manic-depressive patients: a replication. J Affect Disord 53: 87-90.
31. Sayin A, Kuruoglu AC, Yazici Gulec M, Aslan S (2007) Relation of temperament and character properties with clinical presentation of bipolar disorder. Compr Psychiatry 48: 446-451.
32. Sheehan DV, Lecrubier Y, Harnett Sheehan K, et al. (1997) The validity of the Mini International Neuropsychiatric Interview (MINI) according to the SCID-P and its reliability. European Psychiatry 12: 232-241.
33. Hamilton M (1960) A rating scale for depression. J Neurol Neurosurg Psychiatry 23: 56-62.
34. Bech P, Rafaelsen OJ, Kramp P, Bolwig TG (1978) The Mania Rating Scale: Scale construct and inter-observer agreement. Neuropharmacology 17: 430-431.
35. Cloninger CR (1994) Temperament and personality. Curr Opin Neurobiol 4: 266-273.
36. Helmus TC, Downey KK, Arfken CL, Henderson MJ, Schuster CR (2001) Novelty seeking as predictor of treatment retention for heroin dependent cocaine users. Drug Alcohol Depend 61: 287-295.
37. Leblond J, Ladouceur R, Blaszczynski A (2003) Which pathological gamblers will complete treatment? Br J Clin Psychol 42: 205-209.