Elsevier

Drug and Alcohol Dependence

Volume 152, 1 July 2015, Pages 170-176
Drug and Alcohol Dependence

Therapist–client relationships in a psychological therapy trial for psychosis and substance misuse

https://doi.org/10.1016/j.drugalcdep.2015.04.006Get rights and content

Highlights

  • Therapists’ perceptions of improvement were associated substance use outcomes.

  • Therapists’ perceptions of improvements were not associated with symptom outcomes.

  • Therapists’ perceptions of improvement were related to alliance.

  • Therapists’ perceptions of alliance were unrelated to outcomes.

  • Insecure attachment predicted outcomes for symptoms but not substance misuse.

Abstract

Background

This study aimed to explore factors associated with outcomes in a randomised controlled trial of integrated motivational interviewing and cognitive behavioural therapy for psychosis and substance misuse.

Method

Clients and therapists completed self-report measures of alliance and clients completed a self-report measure of adult attachment. Trial therapists were also asked to identify challenges in therapy, client strengths and reasons for client making and not making changes in relation to substance misuse.

Results

Neither therapist-rated nor client-rated alliance was significantly related to objective outcomes. Client insecure attachment avoidance was associated with poorer symptoms and functioning at 12 and 24 months; although not changes in substance misuse. Therapists’ perceptions of therapeutic processes (e.g., challenges to therapy, client strengths, client reasons for change and alliance) were consistent with previous literature. Therapists’ perceptions of client improvement were associated with reductions in substance use at the end of treatment and their ratings of therapeutic alliance.

Conclusion

Insecure adult attachment styles may be a potentially important predictor of symptom outcomes for people with psychosis and substance misuse. Trial therapists may also provide an important source of information about therapeutic processes and factors associated with outcome.

Introduction

Often people with a dual diagnosis of psychosis and substance misuse have low motivation to reduce their use and there are obstacles to delivering therapy to this group (Barrowclough et al., 2007). A recent Cochrane paper reviewed 32 trials comparing psychosocial interventions for substance misuse with standard care in people with serious mental illness and concluded that there was no evidence that any existing treatments were effective (Hunt et al., 2013). Randomised controlled trials (RCTs) of complex interventions have been criticised as it is often difficult to know why interventions worked or not (Oakley et al., 2006). The current study explores a number of different contextual factors which influence outcomes for individual patients in the largest trial for people with a dual diagnosis of psychosis and substance misuse to date (Barrowclough et al., 2010a). This RCT involved 327 participants and compared MI integrated with cognitive behavioural therapy (CBT) and treatment as usual only. The authors found a significant effect of therapy on amount of substance used per substance using day which was maintained over two year follow-ups; although the therapy had no effect on hospital admissions or symptoms. Identifying factors which contribute to these outcomes has important treatment implications as it can help future researchers and therapists to modify therapies to maximise benefit (Webb et al., 2010).

One process factor that is frequently investigated to help explain outcomes is therapeutic alliance. Good alliance has been shown to predict outcomes across a range of different therapeutic modalities and client groups (Martin et al., 2000), including people with substance misuse (Meier et al., 2005) and in CBT for psychosis (Dunn and Bentall, 2007). The strength of the association between alliance and outcome has been questioned due to the relatively small nature of its magnitude and the possibility of third variables confounding the relationship (Crits-Christoph et al., 2006). However, there is some evidence from research evaluating interpersonal psychotherapy for depression that alliance remains significantly associated with changes in symptoms even after controlling for a wide range of patient-related confounding variables including prior improvement (Klein et al., 2003). This suggests that alliance is still a potentially important variable to consider in therapy process-outcome research.

Given the potentially important role of alliance in predicting outcomes it is beneficial to identify factors that are associated with alliance. Identifying such factors may inform interventions to improve alliance and, ultimately, outcomes. Barrowclough and colleagues (2010b) previously reported baseline predictors of therapist-rated and client-rated alliance from the trial. Client baseline variables that were associated with therapist perceptions of better alliance included: White race, living with others, positive attitude towards medication, greater insight, lower levels of depression and lower levels of dysphoria. Poorer insight at baseline was related to poorer client-rated alliance.

Other studies investigating predictors of alliance in therapy for people with psychosis have identified a range of predictors of therapist-rated alliance, but therapist and client perceptions of alliance are not highly correlated and there is relatively little consensus regarding determinants of client-rated alliance (Couture et al., 2006, Svensson and Hansson, 1999). One consistent finding in the general psychotherapy literature is that the quality of clients’ current or past attachment relationships predicts the quality of therapeutic relationships. A number of studies have found that clients who have secure attachment styles and are comfortable with close emotional relationships develop better alliances (Diener and Monroe, 2011, Smith et al., 2010). Attachment styles may be related to improved therapy outcomes as a result of their associations with therapeutic alliance; a recent study involving people with early psychosis found associations between insecure attachment and poorer recovery from symptoms 12 months later (Gumley et al., 2014). It is also possible that interpersonal factors, such as attachment style, act as a ‘third’ variable which explains previous associations between alliance and outcomes (Crits-Christoph et al., 2006). Client attachment style is associated with both therapist- and client-rated alliance in the context of relationships between people with psychosis and psychiatric nurses (Berry et al., 2008), but to our knowledge there are no studies investigating associations between attachment and alliance and attachment and outcomes in psychological therapy for people with psychosis.

One important perspective that has the potential to offer insight into factors influencing alliance and therapy outcome, but which has been neglected by previous research, is that of the trial therapists. Trial therapists have direct experience in therapy delivery and, thus, provide a useful source of information about therapeutic processes influencing outcomes, such as barriers to treatment engagement, client strengths and reasons for change/no change (McGowan et al., 2005). They have insight into factors that facilitated or impeded change, which may not be captured by outcome measures. A number of studies involving people with a diagnosis of psychosis have found that both case managers and therapists report more difficulties in engaging those with insecure attachment patterns (Berry et al., 2007). However, to our knowledge, no previous studies have investigated how therapists’ perceptions of a range of different aspects of client presentations relate to alliance or actual outcomes. Indeed, we do not even know if trial therapists are able to provide accurate inferences about clients’ objective level of improvement.

The first aim of this study was to investigate associations between alliance, attachment and outcomes. We hypothesised positive associations between attachment and alliance and between both attachment and alliance and outcomes. Our secondary aim was to describe trial therapists’ perceptions of challenges and obstacles in delivering MI and CBT for clients with dual diagnoses and therapists’ perceptions of clients’ strengths and resources that facilitated change. In addition, we aimed to describe therapists’ perceptions of reasons for clients making and not making changes in relation to their substance misuse. Finally, we aimed to investigate therapists’ perceptions of how much each client improved as a result of therapy and whether therapists’ perceptions were related to actual outcomes and alliance. We hypothesised positive associations between therapist perceptions of outcomes and actual outcomes and alliance.

Section snippets

Participants and procedure

The sample comprised participants in the treatment arm of the MIDAS trial (Motivational Interventions for Drug and Alcohol misuse in Schizophrenia or psychosis) (Barrowclough et al., 2010a) and the five therapists. Patients were randomised into the intervention arm of the trial, or the monitoring and assessment arm, and followed up at 12 months (end of treatment for those in the treatment arm of the trial) and 24 months.

End of therapy forms

At the end of treatment sessions, trial therapists completed forms about

Sample characteristics

The MIDAS sample comprised 327 patients and 164 were allocated to treatment. Of these 164 patients, 135 were followed up at 12 months and 129 were followed up at 24 months. Participants were predominantly male, living alone and in their late 30s. The majority had long histories of psychosis and substance misuse and nearly all met DSM-IV criteria for non-affective as opposed to affective psychosis. A more detailed summary of sample characteristics and descriptive statistics for measures is

Overview

This study aimed to improve understanding of factors associated with outcomes in an RCT of MI and CBT for psychosis and substance misuse. Neither therapist nor client alliance were related to client attachment nor outcomes. Insecure avoidant attachment was associated with more severe symptoms at 12 and 24 months; although not changes in substance misuse at 12 and 24 months. Therapists were able to identify a number of challenges and obstacles to therapy, client strengths and resources that

Role of funding source

This research was supported by the Medical Research Council (MIDAS trial grant number – G0200471).

Contributors

KB – conception and design, analysis and interpretation of data and drafting article. RH, LG, GH, MF and CB – design, interpretation of data and revising article for important intellectual content.

Conflict of interest

No conflict declared.

Acknowledgements

We thank all participants who took part in the studies from the NHS Trusts and all the NHS staff who helped with recruitment. We also thank Jasmine Elwheshi for assistance with data coding.

References (33)

  • K.B. Carey et al.

    Decisional balance regarding substance use among persons with schizophrenia

    Community Ment. Health J.

    (1999)
  • S.M. Couture et al.

    Do baseline client characteristics predict the therapeutic alliance in the treatment of schizophrenia?

    J. Nerv. Ment. Dis.

    (2006)
  • P. Crits-Christoph et al.

    Does the alliance cause good outcome? Recommendations for future research on the alliance

    Psychotherapy

    (2006)
  • M.J. Diener et al.

    The relationship between adult attachment style and therapeutic alliance in individual psychotherapy: a meta-analytic review

    Psychotherapy

    (2011)
  • G. Dunn et al.

    Modelling treatment-effect heterogeneity in randomized controlled trials of complex interventions (psychological treatments)

    Stat. Med.

    (2007)
  • P.A. Garety et al.

    Cognitive-behavioural therapy and family intervention for relapse prevention and symptom reduction in psychosis: randomised controlled trial

    Br. J. Psychiatry

    (2008)
  • Cited by (0)

    View full text