Is Narrative Reformulation During Cognitive Analytic Therapy Helpful?

Rationale Despite extensive outcome research validating psychotherapy in general as an effective treatment, the outcome literature has been slower to identify the effective, active ingredients of individual therapies. It has been proposed that different active ingredients within therapy relate to outcomes; and definitive technical features are hypothesised to gain importance as psychopathology increases. Therefore research is necessary which compares components of overall treatments to assess their contribution to outcome.

Cognitive analytic therapy (CAT) is a brief therapy (16 or 24 session intervention) drawing upon components of cognitive, and psychodynamic psychotherapies. CAT proposes the use of specific interventions at identified time points within therapy, such as the reformulation letter after a period of assessment. This narrative account reformulates distress, explicitly linking problematic repetitive behavioural patterns with their developmental origins. The client's current life situation is then described alongside target problems and their underlying maladaptive procedures. The reformulation letter is considered a 'crucial therapeutic task' and 'powerful agent of change', but its ability to act in such a manner is largely assumed as the evidence base for these claims has been poorly researched.

Reformulation Despite numerous claims regarding the benefits of cognitive case formulation, there is a paucity of methodologically sound research. This is magnified when looking specifically at the CAT reformulation process. Existing research focuses on accuracy, and whilst some research does exist regarding the impact of reformulation, these are narrative case studies and cannot be generalised. Research utilising the single case experimental design methodology has identified evidence of sudden gains at the point of reformulation.

Depression The introduction of the Increasing Access to Psychological Therapies (IAPT) initiative has focussed on the use of brief, evidence based therapies for this client group and particularly CBT at step 3 of a stepped care model. However, CAT is increasingly used in primary care settings where a naturalistic evidence base is growing regarding its effectiveness for treating depression. The contract can be shortened where 'the threshold to consultation is low' and patients symptomology suggests mild disturbance. Working with predetermined time limits acts to heighten the therapeutic process whilst protecting against over-dependence; hence time-limited interventions can be as clinically effective as lengthy therapy for most people. Brief CAT interventions may therefore be of use for mild to moderate depression.

Aims

The proposed study aims to:

1. Investigate the impact of reformulation letters with mild-moderately depressed clients in terms of the therapeutic alliance and the helpfulness of therapy
2. Compare the outcomes (symptom amelioration) for depressed clients receiving a reformulation letter as part of therapy and those who don't receive the reformulation letter over time including follow-up.
3. Provide initial practice-based evidence of the effectiveness of 8-session CAT in Primary Care with depressed clients.

Hypotheses

Clients receiving a reformulation letter will:

1. Display enhanced therapeutic alliances in comparison to clients receiving CAT without the letter.
2. Perceive therapy as more helpful in comparison to clients receiving CAT without the letter.
3. Achieve better outcomes (greater symptom amelioration) in comparison to clients receiving CAT without the letter.

Theoretical and Clinical Implications The proposed study will add to the nascent evidence base for CAT, with regard to the impact of the reformulation letter on therapeutic processes, outcome and effectiveness. Comparing the impact of receiving a reformulation letter versus excluding this specific event allows its effects to be isolated. This contributes to theoretical understandings of the active ingredients of CAT, and more particularly the role of this specific tool. In addition this may inform the debate surrounding the role of specific versus non-specific contributions in therapy. Identifying tools that augment treatment effectiveness offers the potential to improve therapeutic outcomes.

Improved understanding of how specific tools augment therapeutic processes and affect outcomes has clinical utility. Producing and negotiating reformulation letters takes valuable therapist, and in-session time. Their use, impact and perceived helpfulness are therefore important to investigate thoroughly. This is perhaps of greater importance in brief therapies, where time is a valuable resource. Furthermore, exploring the potential utility of a brief 8-session CAT for primary care mental health problems provides a test of whether CAT could be further adapted and shortened to match the service demands of Primary Care.

Design Utilising a randomised deconstruction methodology, participants will be randomly allocated into one of two active study arms (e.g. 'reformulation letter CAT' or 'no reformulation letter CAT'). Outcome measures will be matched and taken on a session-by-session basis, prior to, during, and following CAT treatment in both arms to enable the comparison of relative efficacy.

Method Participants Participants will be recruited via Barnsley PCT's 's Improving Access to Psychological Therapies (IAPT) service. They will be working age adults (18-65), referred to primary care mental health services with mild to moderate depression.

The study aims to recruit 24 participants in total.

For supervisory purposes, fidelity to the CAT model will be assessed using:

A measure of CAT competence (CCAT; Bennett \& Parry, 2004). The CCAT measures therapist competence over ten domains related to practice and model characteristics. Validity, reliability and internal consistency have been established as sufficient (Bennett \& Parry, 2004).

Procedure A Consultant Cognitive Behavioural Psychotherapist will identify clients meeting the inclusion criteria for the study. Suitable participants will be provided with an information sheet explaining the research and inviting them to participate. Clients agreeing to participate will be asked to sign a consent form. The consent form clarifies that participants have read the information sheet and had opportunities to ask questions. The voluntary nature of participation, confidentiality, and right to withdraw at any time is stated. Consent will be sought to audiotape therapy sessions for supervision purposes. Once consent to participate has been established the identified clients details will be passed to the research team for randomisation.

Participants will be randomly allocated into one of two arms ('letter' or 'no letter') using the GraphPad (2005) computer randomisation software package. Randomised participants will be offered an 8-session course of CAT. The CAT will be delivered by one of three trainee clinical psychologists in their final year of training, one of which is the researcher. All of the trainees have completed a two-day introductory CAT workshop, in addition to standard clinical psychology training to date. To minimise therapist effects participants in each arm will be equally distributed among the team.

All participants will receive eight sessions of CAT with a follow-up appointment eight weeks post-therapy. Assessment will take place over the initial two sessions followed by reformulation at session three. At the reformulation stage 12 patients ('letter' arm) will receive a reformulation letter and provisional sequential diagrammatic reformulation in line with standard CAT practice. 12 patients ('no letter' arm) will receive the SDR without a reformulation letter at session 3. The remaining intervention and termination procedures will follow standard CAT procedures. Patients in both arms will receive, and be invited to produce, a goodbye letter at termination. All participants will be invited to attend a follow-up session eight weeks post-therapy.

To ensure model adherence all therapy sessions will be taped for supervision purposes and reviewed by the CAT qualified, and supervisor-trained supervisor. A minimum of one full taped session from each therapist will be assessed for model adherence by the supervisor using CCAT. In addition, taped excerpts from sessions will be routinely reviewed in supervision; therapists will be invited to employ CCAT as a self-reflective tool in preparation for this. Issues of non-adherence will be considered a training implication, representative of areas for improvement and addressed via standard supervisory procedures. Non-adherence will not preclude data inclusion for the purposes of this research; rather it will be addressed as a limitation of the study related to therapist experience.

Data Analysis To allow comparisons between the two active study arms the primary outcome variables relevant to each hypothesis will be subject to analysis using ANOVA. Trend analyses will be additionally conducted on session-by-session data to highlight and compare patterns of change within the study arms over time. On the pre-post data assessments of clinical and reliable change in each arm will be calculated and compared using the Reliable Change Index (RCI; Jacobson \& Truax, 1991). Outcome data from the PHQ9 will be used to calculate the reliable change index (RCI; Jacobson \& Truax, 1991). This identifies reliable and clinically significant changes at small (z > 1.96, p<.05), medium (z > 2.58; p<.01), and large (z > 3.29, p<.001) effect sizes. Power analysis suggests sufficient power for a Between x Within Subjects ANOVA, paying particular attention to the Groups x Time-Points interaction effect. Assuming a 'medium' effect size of f = .25, a significance level of alpha = .05, and 2 groups of participants providing data at four time points (pre, reformulation, post, and follow up) the proposed total sample size of 24 gives 80% power. Power analysis also suggests sufficient power to conduct a trend analysis form of ANOVA to investigate patterns of change over time (session-by-session data). Assuming a 'medium' effect size of f = .25, a significance level of alpha = .05, and 2 groups of participants providing data at eight time points (session-by-session data excluding follow up) the proposed sample size of 24 will give 95% power.

Hypotheses

Clients receiving a reformulation letter will:

1. Display enhanced therapeutic alliances in comparison to clients receiving CAT without the letter.
2. Perceive therapy as more helpful in comparison to clients receiving CAT without the letter.
3. Achieve better outcomes (greater symptom amelioration) in comparison to clients receiving CAT without the letter.