Relating emotional processes to outcome in experiential psychotherapy of depression

This study examined the relationship between emotional processes and outcome in 32 depressed clients who received brief experiential psychotherapy treating depression. The Emotion Episode (EE) Method (Greenberg & Korman, 1993; Korman, 1991) was used to demarcate emotion segments from therapy transcripts in the early, middle, and late phases of therapy. The Client Emotional Arousal Scale III (Warwar & Greenberg, 1999) was used to assess clients' in-session emotional arousal. The Client Experiencing Scale (Klein, Mathieu, Gendlin, & Kiesler, 1969) was used to evaluate in-session experiencing. Each EE segment was given a modal (average) and a peak (highest) rating on emotional arousal (EA) and experiencing (EXP); these ratings were averaged for each phase of therapy. Final outcome was assessed using residual gain scores on four self-report measures that were administered to clients at pre- and post-treatment: the Beck Depression Inventory (BDI) (Beck, Ward, Mendelson, Mock, & Edbaugh, 1961; Beck, 1972); the Symptom Checklist-Revised (SCL-90-R) (Derogatis & Melisaratos, 1983); the Rosenberg Self-Esteem Scale (RSE) (Rosenberg, 1965); and the Inventory of Interpersonal Problems (IIP) (Horowitz, Rosenberg, Baer, Ureno & Villasernor, 1988). This study hypothesized that although mid-therapy EA will predict outcome after controlling for early EXP, when EXP in the middle and late phases of therapy are added to EA, together, these variables will account for more outcome variance. As well, it was predicted that these emotional processes will predict outcome over and above early the working alliance. Finally, it was hypothesized that all variables in the model will be independent predictors of outcome. Hierarchical regression analyses showed that both early modal and peak EXP were significant predictors of outcome on the SCL-90-R and the BDI. In the final model, peak emotional process variables contributed 58% of the variance explaining SCL-90-R and, mid-therapy EA and late EXP were independent predictors of SCL-90-R. By comparison, although peak emotional processes contributed 47% to the explanation of BDI, none of the emotional process variables were unique predictors of outcome in the final model. For self-esteem, mid-therapy peak EA was the only process variable that predicted outcome. No relationships were found between IIP and emotional process variables. Findings of this study showed that combining EA midtherapy with EXP predicted outcome on SCL-90-R and BDI better than either of these variables alone. This suggests that the activation of emotion in therapy is more therapeutic if it is in the service of facilitating deeper emotional experiencing. Therapeutic and theoretical implications of these findings are discussed.