Outcome trajectories and mediation in psychotherapeutic treatments of major depression

Trajectories and mediators of change were investigated in a process-outcome study. Patients were allocated at random to psychoanalytic therapy (PA) or psychodynamic therapy (PD), and later to cognitive-behavioral therapy (CBT). Measurement points were at pre-treatment, during ongoing treatment, at post-treatment, and during a three-year follow-up. Outcome trajectories were assessed using the Beck Depression Inventory (BDI; Hautzinger et al. 1994), the Symptom Checklist 90 Revised Version (SCL-90-R; Franke 1995), and the Inventory of Interpersonal Problems (IIP; Horowitz, Strauss, and Kordy 2000). Therapeutic alliance and introject were tested as mediators, assessed using the Helping Alliance Questionnaire (HAQ; Bassler, Potratz, and Krauthauser 1995) and INTREX, introject surface (Tress 1993). Multilevel modeling was applied to estimate outcome trajectories and to test for mediation. Symptoms decreased in early and ongoing treatment in all treatment groups. After the end of treatment, depressive and general psychiatric symptoms continued to decrease in significantly greater degree in the PA group than in the PD and CBT cohorts. During early treatment, interpersonal problems decreased significantly more in those allocated to PD than in the PA and CBT groups. During ongoing treatment, improvement in interpersonal problems was significantly higher in the PA group than in the others and, compared to CBT, continued to increase significantly after termination. Mediational analyses revealed that neither introject affiliation nor therapeutic alliance mediated differential treatment effects.