1 Greater Manchester West Mental Health NHS Foundation Trust

EXPERIENCES OF TEAMS CBT

What are people’s experiences of a novel cognitive behavioural therapy for bipolar disorders? A qualitative investigation with participants on the TEAMS trial.

Emmeline Joyce1, 2*, Sara Tai2, Piersanti Gebbia2 and Warren Mansell2

1 Greater Manchester West Mental Health NHS Foundation Trust

2 The University of Manchester

*Requests for reprints should be addressed to Emmeline Joyce, The Psychosis Research Unit, Rico House, Greater Manchester West Mental Health NHS Foundation Trust, Bury New Road, Manchester, M25 3BL, United Kingdom, e-mail: , telephone: +44 (0)161 358 1395

Key words

Bipolar disorder; cognitive behavioural therapy; TEAMS; qualitative

Abstract

Background

Psychological interventions for bipolar disorders typically produce mixed outcomes and modest effects. The need for a more effective intervention prompted the development of a new cognitive behavioural therapy, based on an integrative cognitive model (TEAMS therapy). Unlike previous interventions, TEAMS addresses current symptoms and comorbidities, and helps clients achieve long-term goals. A pilot randomised controlled trial (the TEAMS trial) of the therapy has recently concluded. This study explored participants’experiences of TEAMS, recommendations for improvement, and experiences of useful changes post-therapy.

Methods

Fourteen TEAMS therapy participants took part in semi-structured interviews. Their accounts were analysed using interpretative thematic analysis. Two researchers coded the dataset independently. Member checks were conducted of the preliminary themes.

Results

Two overarching themes; “useful elements of therapy”and “changes from therapy”encompassed 12 emerging subthemes. Participants appreciated having opportunities to talk and described the therapy as person-centred and delivered by caring, approachable and skilled therapists. Some recommended more sessions than the 16 provided. Helpful therapeutic techniques were reported to be, normalisation about moods, methods to increase understanding of moods, relapse-prevention, reappraisal techniques and metaphors. However, some did not find therapeutic techniques helpful. Post-therapy, many reported changes in managing mood swings more effectively and in their thinking (although some participants reported changes in neither). Many described increased acceptance of themselves and of having bipolar disorder, increased productivity, and reduced anxiety in social situations.

Conclusions

The present study evaluates participants’therapy experiences in detail, including aspects of therapy viewed as helpful, and meaningful post-therapy outcomes.

Key Practitioner Message

•This is the first paper to qualitatively explore people’s experiences of individual psychotherapy for bipolar disorders. It highlights elements of psychotherapy described as particularly helpful or unhelpful and the clinical changes viewed as most impactful.

•Participants reported benefitting in a number of ways from TEAMS therapy. They valued learning to reappraise and problem-solve situations and manage moods.

•Participants identified TEAMS techniques as helpful, such as exploring advantages and disadvantages of moods, and building healthy self-states.

Background

Effective treatment of bipolar disorders is considered a significant challenge (Geddes & Miklowitz, 2013). Existing psychological interventions such as cognitive behavioural therapy (CBT), family-focused therapy and psychoeducation have common components including preventing relapse, stabilising mood and increasing functioning (Da Costa et al., 2011; Geddes & Miklowitz, 2013). However, such interventions have produced mixed outcomes and modest effects in domains including depressive and manic symptoms, relapses, and hospital admissions (Oud et al., 2016; Reinares, Sánchez-Moreno, & Fountoulakis,2013).

A large majority of people with bipolar disorders experience inter-episode subclinical depressive symptoms (Vieta et al., 2010), and comorbid diagnoses including anxiety disorders (Pavlova, Perlis, Alda, & Uher, 2015), substance use disorders (Merikangas et al., 2007) and eating disorders (Seixas et al., 2012). Subclinical symptoms and comorbid diagnoses are associated with a range of negative outcomes and are significantly detrimental to prognosis (De Dios et al., 2012; McElroy et al., 2011; Swann, 2010). However, existing interventions tend not to target current subclinical symptoms and comorbid diagnoses, in favour of focusing on future-oriented outcomes such as relapse prevention and improving long-term social and occupational functioning (Geddes & Miklowitz, 2013). This may partially explain the mixed findings of therapy trials.

There is also evidence that available psychotherapies for bipolar disorders may be missing opportunities to provide optimal treatment, particularly for those with comorbid bipolar and anxiety disorders. Some people with comorbid bipolar and anxiety disorders may require more intensive psychotherapy to recover from acute depression, compared to those without comorbid anxiety (Deckersbach et al., 2014). However, people with comorbid bipolar and anxiety disorders report receiving inadequate treatment for their needs (Hawke, Provencher, Parikh, & Zagorski, 2013). Additionally, given the variation in presentation, course and sociodemographic background in people with bipolar disorders (e.g., Greenwood et al., 2013) it is important to evaluate the effectiveness of new therapies across a broad range of people who vary in clinical and sociodemographic characteristics.

A cognitive model of bipolar disorders and mood swings was developed to address the above issues (Mansell, Morrison, Reid, Lowens, & Tai, 2007). The model proposes that mood swings are maintained by extreme, conflicting, numerous, personal beliefs about internal state changes (Mansell, 2010a). For example, feeling excitable may be believed to signify both an imminent breakdown and having unlimited creative energy (Mansell et al., 2007). Consequently, individuals attempt to control internal state changes in contrasting, extreme and counterproductive ways. They may try to suppress or enhance internal states at different times using specific behavioural strategies (Mansell, 2010a). Extreme, conflicting appraisals of internal states have been found to predict symptoms of hypomania and depression (Dodd, Mansell, Morrison, & Tai, 2011) and distinguish people with bipolar disorders from non-clinical controls (Mansell, 2006; Mansell & Jones, 2006; Mansell et al., 2011) and people with remitted unipolar depression (Alatiq, Crane, Williams, & Goodwin, 2010; Mansell et al., 2011).

Reports from people experiencing bipolar disorders are also consistent with the cognitive model. In studies exploring recovery from bipolar disorders, participants described conflicts between striving for wellness whilst wanting to retain the positive influences of mania (Mansell, Powell, Pedley, Thomas, & Jones,2010). Participants also reported conflicting views of mania as having intensely positive qualities whilst representing a dangerous loss of control (Veseth, Binder, Borg, & Davidson, 2012). These findings support the idea that clients have contrasting views of internal states and how to control them (Mansell et al., 2010). In contrast, another study interviewed people with only hypomanic experiences about the factors preventing them from developing bipolar disorder. No participants described extreme negative beliefsabout hypomania, or described hypomania as a highly desirable state that they strived for often (Seal, Mansell, & Mannion, 2008).

The Mansell et al. (2007) model informed a new cognitive behavioural therapy (CBT) for bipolar disorders and mood swings, called “Think Effectively About Mood Swings”(TEAMS). TEAMS aims to help clients increase awareness of extreme appraisals of internal states and conflicted attempts at controlling them, and develop more helpful ways of responding to them (Searson, Mansell, Lowens, & Tai, 2012). TEAMS incorporates metacognitive techniques such as worry and rumination monitoring, alongside cognitive behavioural techniques (Searson et al., 2012). The therapy is client-led and is present-moment-focused by addressing current symptoms rather than preventing future episodes (Mansell, 2010a). TEAMS also aims to treat comorbid diagnoses and subclinical symptoms (Mansell, 2010a), acknowledging their impact on prognosis. TEAMS has the potential to optimally support recovery in bipolar disorders as the TEAMS model was informed by clients’views. For example, clients have reported that living purposeful lives is more important to them than avoiding relapse (Mansell et al., 2010). TEAMS therefore helps clients to deal with longstanding problems and achieve important, client-specific life goals which may range from coping more effectively with low mood to improving relationships with family members or confronting longstanding social fears.

A case series has demonstrated TEAMS therapy to be preliminarily effective and acceptable (Searson et al., 2012). A pilot randomised controlled trial testing the feasibility and acceptability of TEAMS therapy has recently concluded.It was predicted that TEAMS therapy would reduce depression compared to treatment-as-usual. The primary outcome for the trial was Beck Depression Inventory (BDI) score (Beck et al., 1961) at six months post-randomisation (see Mansell et al., 2014 for trial protocol). Depression was chosen as the key outcome, rather than mania or hypomania, as depressive symptoms are the largest contributor to impaired functioning in people with bipolar disorders (Gitlin, Swendsen, Heller, & Hammen, 1995;Morriss et al., 2007).

It is valuable to consider clients’views of receiving TEAMS to understand if it works as intended and learn about any changes that could improve its acceptability and helpfulness. Qualitative methods can exploreareas that are inaccessible to quantitative methodologies: insight into subjective views and beliefs; how people react to and represent experiences (Kazdin, 2008) and are ideal for exploring change processes (Willig, 2009). Qualitatively evaluating interventions is important to more fully understand acceptability and effectiveness (Thompson & Harper, 2012). However, published literature on people’s experiences of face-to-face psychological interventions for bipolar disorders is limited, with only three studies available (Chadwick, Kaur, Swelam, Ross, & Ellett, 2011; O’Connor, Gordon, Graham, Kelly, & O’Grady Walshe, 2008; Poole, Smith, & Simpson, 2015).

After reviewing thepublished qualitative studies, we concluded there is currently insufficient information to provide insight into experiences of individual, face-to-face psychotherapy. Participants’experiences were often strongly affected by the psychotherapy being delivered within a group modality (O’Connor et al., 2008; Poole et al., 2015). Additionally, two studies reported on psychoeducation which emphasised relapse prevention (O’Connor et al., 2008; Poole et al., 2015), which is not the emphasis of TEAMS owing to its focus on current problems, quality of life and sense of purpose.

The aims of the current study were therefore to explore participants’experiences of TEAMS therapy and mechanisms of effect within the therapy and clients, and elicit recommendations for improvement. The study highlighted views about useful and less useful aspects of TEAMS, changes participants attribute to TEAMS and how they perceived these changes to have occurred. These findings would inform the future delivery of TEAMS therapy. They supplemented existing literature exploring what people with bipolar disorders need from psychological interventions, and the most impactful clinical changes.

Method

Participants

The TEAMS trial was approved by the London Queens-Square Research Ethics Committee (reference: 11/LO/1326). Eighty-two TEAMS trial participants were randomized to either Treatment As Usual (TAU) or TEAMS therapy. Interviews were conducted with 14 of the 41 participants randomised to the TEAMS therapy condition of the trial. Thirty participants were contacted in a random order using a randomisation function on Microsoft Excel. Participants who did not consent to their therapy sessions being recorded, who had withdrawn from the trial, or whose therapists advised of risk factors contraindicating interviews were not contacted (N = 11). Six participants did not respond to two contact attempts. Six declined interviews, reporting reluctance to be interviewed about their therapy experiences (N = 3), experiencing a mood episode at time of contact (N = 2), or no reason (N = 1). Of the 18 participants who preliminarily agreed, four subsequently disengaged from contact. This resulted in a sample of 14 participants[1]. The average time between participants attending their TEAMS baseline assessments and post-therapy interviews was 18.7 months (range: 8-35 months, median: 19 months).Participants were purposively sampled, due to the researcher actively recruiting only people who shared the same, unique characteristic of having received TEAMS therapy on the TEAMS trial (see Table 1 for demographic details).

The TEAMS trial sample included participants with diagnoses of Bipolar I, II and Not Otherwise Specified (NOS). All participants were help-seeking at baseline,with depressive symptoms corresponding to a scoreover 15 on the BDI(Beck et al., 1961)over at least two non-consecutive weeks. Participants with comorbid diagnoses were not excluded, other than those with substance use disorders and psychotic disorders. For further information about the TEAMS trial inclusion and exclusion criteria, please refer to the trial protocol (Mansell et al., 2014).

Interested individuals were screened with the BDI (Beck et al., 1961)and an interview based on the Brief Screening Interview. Eligible participants received a baseline assessment, comprising a battery of measures including the Structured Clinical Interview for DSM IV: Axes I and II (SCID) (First et al., 1997; First, Spitzer, Gibbon, & Williams, 2002). The SCID assessment includes screening for anxiety disorders, including an index of traumatic experiences reported by participants. Those meeting criteria for a bipolar disorder on the SCID were randomly allocated to receive either TEAMS therapy plus assessments at three, six, twelve and eighteen months post-randomisation, or TAU (only assessments at six, twelve and eighteen months post-randomisation).

(Please insert Table 1 here)

Interviews

Semi-structured interviews were digitally audio-recorded with participants’verbal permission and transcribed verbatim. Semi-structured interviews are flexible in allowing participants to raise subjects important to them (Byrne, 2012), thus reducing the possibility of missing significant topics (Langdridge & Hagger-Johnson, 2009).

The first author conducted all interviews excepting the first two (see Table 1), which were conducted with a co-interviewer. Interviews utilised a topic guide (see Table 2) developed by the TEAMS research team, including an experienced service user-researcher. Questions were refined during discussions and role-plays between three of the authors and the co-interviewer. Questions focused on participants’experiences of therapy, reasons for joining the TEAMS trial, recommendations for improving TEAMS therapy and for services and interventions for bipolar disorders. Questions were open-ended to encourage participants to provide detailed, rich answers (Byrne, 2012). Probing follow-up questions were used to elicit further detail following interesting statements (Millar & Tracey, 2009). Interviews ranged between 31 and 106 minutes long (71 minutes long on average).

(Please insert Table 2 here)

Analysis

Interview transcripts were analysed thematically in accordance with Braun and Clarke’s (2006) guidelines. Thematic analysis allows the identification of frequently occurring patterns across datasets. This produces themes which describe the dataset in rich detail and emphasise features of the data most significant to the research question (Braun & Clarke, 2006; Joffe, 2012). Analysis was conducted within a realist framework due to this study’s implicit aims; to obtain information about participants’experiences (Willig, 2009) and to ensure the findings were relevant and accessible to clinical practice (Braun, Clarke, & Terry, 2015). An interpretative approach was used to identify assumptions and conceptual ideas underlying the overt data (Braun et al., 2015). Analysis proceeded inductively, with preliminary themes arising from the data rather than according to a theoretical framework (Boyatzis, 1998). Coding proceeded alongside data production and transcription. The analytic procedure is described below.

• Two authors coded the dataset independently, and without reference to coding software. Both carefully read all transcripts at least three times before coding to ensure familiarisation.
• All data related to, participants’experiences of TEAMS therapy and previous therapies; reasons for participating in the TEAMS trial; and recommendations for improving TEAMS therapy and services/interventions for bipolar disorders was coded.
• The authors coded iteratively by re-coding transcripts at least once after initial coding, ensuring inclusion of all relevant data.
• The authors maintained an ongoing list of codes and their data extracts. After analysing each transcript, the authors searched for relationships between the listed codes. This allowed the identification of preliminary emerging themes.
• Later coding and re-coding was informed by the preliminary emerging themes that had been developed through earlier coding and that had arisen purely from the data.
• The authors held five meetings throughout analysis to discuss codes and preliminary themes, and maintained email contact. The discussions allowed authors to determine consistency between their coding, clarify and suggest ideas about codes and themes with each other, and aided preliminary theme development. Jointly discussing and agreeing upon emerging themes increased the reliability of the analysis (Mays & Pope, 1995).
• Inconsistencies in coding were resolved via discussion. Consequently some codes were merged, removed or renamed. For example, “more able to manage bipolar disorder” was merged with “more able to manage moods”.
• After the dataset had been coded and re-coded at least once, the preliminary themes and codes composing them were listed on a word processor. Three authors discussed the preliminary themes to ensure consensus about the significance of the themes.
• The first author refined the themes by drawing thematic maps; arranging the codes composing each preliminary theme into tables; and re-reading all coded data extracts at least five times. This ensured the themes encapsulated coherent, distinct data patterns and clarified the relationships between codes and themes (Braun & Clarke, 2006). Theme refinement resulted in some themes being discarded or merged, and codes being added to or removed from themes. Lastly, the first author re-read all transcripts once to ensure the final themes accurately reflected the dataset and that no relevant data had been omitted.
• Whilst in the final stages of theme refinement, the first author wrote to all 14 participants inviting their participation in member checks. Four participants agreed. All participants were female, aged 32, 40, 56 and 61; three were diagnosed with bipolar II, and one with bipolar I; two were unemployed and two employed (part-time and full-time, respectively).
• The author explained the final themes using a thematic map, and asked for verbal feedback. The author also highlighted aspects of the analysis that individual participants’ interviews had contributed to, and asked for feedback and clarification of these interpretations. This was to acknowledge that data from different participants contributed to different themes (e.g., Cutliffe & Mckenna, 1999). Participants reported that the themes accurately reflected their therapy experiences, and suggested no significant omissions. This helped to ensure the validity of the analysis (Koelsch, 2013; Lincoln & Guba, 1985).

Results