Supervision of psychotherapists and counsellors, especially in the early years of practice, is widely accepted as being important for professional development and to ensure optimal client outcomes. The process and characteristics of clinical supervision have been extensively studied and many theoretical approaches and models of supervision have been developed. Notwithstanding the broad acceptance of supervision, there has been little empirical investigation to establish if supervision can achieve its primary goal of enhancing psychotherapy practice and client outcome. This thesis evaluates the impact of clinical supervision on psychotherapy practice and client outcome by examining the effect of supervision on the working alliance and client symptom reduction in the brief psychotherapeutic treatment of major depression.
The focus of supervision was on the development of the working alliance because the quality of alliance has been found to be a strong mediator of therapy outcome in the treatment of depression. It was hypothesised that supervision would result in stronger working alliance and consequently better treatment outcome.
This study used a hierarchical nested design with multiple intervals of measurement. The experimental variable was supervision (3 levels: process focus, skill focus, no supervision). Dependent variables were: working alliance in therapy (measurement points at first, third and eighth session of treatment), client symptom scores (measurement at intake assessment. Session 1 and Session 8), dropout before completion of a full course of treatment, and satisfaction with therapy. The clinical intervention was a standardised therapy approach (problem solving treatment [PST]) that does not vary across conditions.
Covariate measures were included to identify potential internal or external mechanisms by which supervision may influence working alliance and client outcome. These included supervision skill or process foci adherence (SFAS), supervision evaluation, supervisory working alliance (SWAI), treatment approach (PST) adherence, client psychosocial stress measure (SRLE) and measures of therapist and supervisor level of verbal and nonverbal social skill (SSI).
One hundred and twenty seven therapists were recruited and randomly assigned to a skill focus, process focus or unsupervised group. Skill and process focus supervision was provided as two different methods to develop working alliance. Thirty-eight supervisors provided skill and process foci alliance supervision. One hundred and twenty seven clients meeting diagnostic criteria for major depression were randomly assigned to therapists. All therapists allocated to supervised conditions attended a pretreatment supervision session with their supervisor that focused on techniques to enhance early alliance prior to the first client treatment session. Participating therapists were provided with manualised training in PST and were required to use this approach as the treatment protocol in the study.
The ANOVA procedure was used for the main data analysis of total working alliance between groups, 1 = unsupervised therapy vs., 2 = skill supervision vs., 3 = process supervision. Repeated measures analysis of variance was used to test within subject effects (assessment intervals 1, 3, and 8). Between subject variables were supervision condition (no supervision, skill supervision working alliance foci and process supervision working alliance foci) and dependent variables Working Alliance Inventory (WAI) and Beck Depression Inventory (BDI) scores. Client treatment satisfaction was evaluated using analysis of variance testing for between subject effects of supervision on client treatment evaluation. In each case between subject variables were time (assessment interval 8) and supervision condition (no supervision, skill supervision working alliance foci process supervision working alliance foci). The relationship between supervision condition and treatment attrition was analysed using the Chi squared equation. Relationships between study variables and covariants were analysed using Bivariate correlation.
This is the first time that the role of supervision in client outcome has been investigated using a randomised controlled treatment trial methodology. The results showed a significant effect for both skill and process supervision on working alliance in therapy, symptom reduction and treatment retention and evaluation, but no differences between the two supervision conditions. The strongest effect of supervision was on working alliance after the first session of therapy. The benefits of supervision on working alliance were maintained over the duration of therapy. Clients treated by supervised therapists were significantly more likely to stay in treatment, were significantly more satisfied with treatment and had significantly greater reduction in BDI scores when compared with clients of unsupervised therapists.
The strength of effect found in this study provides initial evidence for the importance of supervision for both working alliance and treatment outcome in the brief treatment of depression. The finding that a single session of supervision had a major impact on working alliance scores raises important questions about the timing, quantity and nature of supervision required to achieve meaningful results.