Audit of Factors Predicting Drop Out from Cognitive Analytic Therapy

Channer, K., and Jenaway, A., 2015. Audit of Factors Predicting Drop Out from Cognitive Analytic Therapy. Reformulation, Winter, pp.33-35.

Introduction

In the Liaison Psychiatry department of Addenbrooke’s Hospital in Cambridge, patients referred for psychological treatment are offered either Cognitive Behaviour Therapy (CBT) or Cognitive Analytic Therapy (CAT). Patients referred for CAT are more likely to have experiences of early trauma or abuse, and to have evidence of difficult relationships with family, in their friendships, or with medical staff teams. The aim for CAT for patients with physical health problems is to explore how their relationship patterns, learned in early life, can impact on their physical health and the way they manage their symptoms. Most patients have either a serious physical illness, such as diabetes, cancer, liver transplant, and are struggling to cope with them, or they are suffering from medically unexplained symptoms (MUS) The CAT approach to MUS is described by Alison Jenaway in a previous issue of Reformulation (Jenaway, 2011).

As we have taken on more trainees in the department (this includes CAT practitioner trainees, junior psychiatrists and trainee clinical psychologists), the number of patients taken on for CAT has increased so that at any one time, around 40 patients are engaged in CAT. There was some concern from consultant liaison psychiatrists that allocating such complex patients to trainee therapists may result in the patient receiving less effective therapy than those seen by trained therapists. In addition, some patients are travelling a long way for treatment because Addenbrooke’s hospital is a specialist centre for many disorders and attracts complicated patients from many outlying areas, some as much as 62 miles away. The liaison psychiatrists were interested in whether the distance patients had to travel was affecting drop-out rates from CAT, in which case it may be better to recommend that these patients are treated nearer to their home (although, in practice, CAT is often not available to patients with these kinds of medical problems). We also decided to look at age, gender and whether the patient had medically unexplained symptoms (MUS) to investigate whether those factors had any impact on rates of successful therapeutic outcome.

The current evidence base has suggested that the therapeutic alliance is the most consistent predictor of positive outcome in psychotherapy, and significant research has been conducted to better understand this relationship. For example, Martin, Garkse and Davis (2000) conducted a metaanalysis and found that the overall relation of therapeutic alliance to outcome is moderate, but consistent, regardless of many of the variables that have been posited to influence this relationship. Ryle and Golynkina (2000) found that ratings of the therapist’s quality was associated with better outcomes. Ackerman and Hilsenroth (2003) reviewed studies of therapists’ activity and attributes that this positively influences the therapeutic alliance. They identified a number of traits (e.g., exploration, reflection, support, warmth, confidence and attending to patient experience) which positively related to outcomes. However, few researchers have extended this question to look at other important factors which may contribute to outcome.

CAT research is still in its infancy and we were not aware of any other papers looking specifically at dropout from CAT therapy. In their trial of CAT for personality disorders, Clarke et all reported that 10 out of 50 patients allocated to CAT discontinued therapy before completion of 24 sessions, giving a drop out rate of 20% (Clarke S., 2013). Thus, this report aimed to identify whether age, gender, medically unexplained symptoms, distance from home to service, and who the client was seen by (i.e., trainee or qualified staff member) were significantly related to whether or not the client completed a course of CAT treatment.

Method

The Liaison Psychiatry database of CAT completers was analysed and 52 consecutive clients, who had engaged in at least the first session of therapy, were examined to see which factors contributed to completion of CAT. Of the 52 participants included in the data analyses, 38 completed treatment whereas 14 dropped out (Mean number of sessions attended by drop outs was 5).

Results

 Age

Age was categorised into 6 decades, and the number, and percentage of dropouts was calculated as illustrated in table 1. Statistical analyses revealed no significant difference between completers and non-completers in terms of age (p >0.05).

Table 1: Age

Gender

Gender was categorised, counted and percentage of drop-out was calculated as illustrated in table 2. Statistical analyses revealed no significant difference between completers and non-completers in terms of gender (p >0.05). Men and women were equally likely to drop out of treatment.

Table 2: Gender

Medically unexplained symptoms (MUS)

Referring clinician’s opinion that the patient’s symptoms were medically unexplained was categorised (i.e., no, yes and maybe) counted and their percentage of drop-out was calculated as illustrated in Table 3. Those without medically explained symptoms were more likely to drop out of treatment as compared to those with medically unexplained symptoms. However, this difference was not significant (p > 0.05). The category of “maybe” was included as it is not always clear to referring physicians whether or not the symptoms have an organic origin. Clearly this categorisation is not precise and some patients who are referred with medically unexplained symptoms get a more formal diagnosis of organic disease during the therapy or at follow up, however, this did not happen to any of this cohort of patients.

Table 3: MUS

Area

The region that the patients lived in was categorised, according to region, counted and the percentage of drop out was calculated as illustrated in table 4 . All areas outside Cambridge were amalgamated and compared with patients who lived in Cambridge (where the service is situated). It seems that those who live outside of Cambridge were more likely to drop out of treatment as compared to those who live in Cambridge as illustrated in table 5. However, this difference was not significant (p >0.05).

Table 4: In or out of Cambridge

Therapist-client allocation

Patients were categorised in terms of whether they were seen by a trainee CAT practitioner or a qualified CAT practitioner. This was categorised, counted and percentage of dropout was calculated as illustrated in table 5. Patients who were seen by trainees were more likely to drop out of treatment as compared to patients who were seen by qualified staff as illustrated in table 6. However, this difference was not significant (p >0.05). It is also worth noting that patients are carefully assessed prior to allocation to trainees and those who are likely to be less complex and easier to work with are allocated to trainees. Therefore the drop out rate is lower for patients who are seen by qualified staff, even though these are the more difficult, complex patients.

Table 5: Seen by

Discussion

This report looked at whether age, gender, MUS, area and who the client was seen by, was significantly related to treatment completion. Whilst no statistically significant relationships were found, it was possible to demonstrate trends in that participants who lived further away from the service or were seen by trainee practitioners were more likely to drop out of treatment. Future research must make efforts to identify which factors underlie these differences. It is possible that significant differences were not found because a relatively small sample size was used. Furthermore, due to the distribution of the data, non-parametric statistics were employed which reduces statistical and predictive power. It is important that subsequent research is conducted to see which factors best predict treatment completion. Such research should be approached with greater scientific rigour, use larger sample sizes, and make attempts to look other factors of importance (i.e., presenting problem, length of treatment, outcome measures). It may also be that these findings would not generalise to a more typical mental health setting, rather than a liaison psychiatry service, where many patients have physical illness and disability. This makes it generally harder for patients to attend regular appointments and also, patients are often more preoccupied with their physical symptoms than the need for therapy.. This may explain why drop out rates were slightly higher than those reported in the personality disorder trial (Clarke, 2013). The audit was presented at a team meeting and the results discussed. Since we are a teaching hospital, the results of the audit has not affected our practice in the issue of allocating patients to a trainee therapist or a qualified therapist, however, I personally am more open with patients about the process of making this decision collaboratively with them. Qualified therapists are likely to feel more confident in handling complex cases and therefore likely to appear more relaxed and more able to contain distress. However, many of these qualities may be part of the intrinsic personality of the therapist, or other life experiences, rather than related to specific therapy training. The team has decided that we will no longer offer psychological therapy to patients living a long way from Cambridge and will try and promote the development of suitable therapy in other areas.

References

Ackerman, S. J., & Hilsenroth, M. J. (2003). A review of therapist characteristics and techniques positively impacting the therapeutic alliance. Clinical Psychology Review, 23, 1–33.

Clarke, S., Thomas, P., and James, K. (2013) Cognitive Analytic Therapy for Personality Disorder: Randomised Control Trial British Journal of Psychiatry, Feb 202 (2), 129-134.

Jenaway, A., (2011). Using Cognitive Analytic Therapy for Medically Unexplained Symptoms – some theory and initial outcomes. Reformulation, Winter, pp.53-55.

Martin, D., Garkse, J. P., & Davis, M. K. (2000). Relation of the therapeutic alliance with outcome and other variables: A metaanalytic review. Journal of Consulting and Clinical Psychology, 68, 438–450.

Ryle, A., & Golynkina, K. (2000). Effectiveness of time-limited cognitive analytic therapy of borderline personality disorder: Factors associated with outcome. British Journal of Medical Psychology, 73, 197-210.