LAUSR

OBJECTIVE Knowledge of patients' in-session experiences that lead to symptom change in psychotherapy is limited. This study aims to investigate the within- and between-patient relationships between three in-session processes in psychotherapy (coping skills, therapeutic relationship quality, and emotional involvement) and symptom change on a session-by-session level. METHOD Participants (n = 1550) with various disorders, including primarily depression and anxiety, were treated with CBT in a German outpatient clinic. Symptom distress was assessed before each session and patients' in-session experiences were assessed at the end of each session using session reports. Person-mean centering was applied to disaggregate within- and between-patients. Within- and between-patient process scores were tested in multilevel models as predictors of next session symptom change. RESULTS On a within-patient level, better session-specific coping skills, better therapeutic alliance, and deeper emotional involvement were followed by next session symptom improvements. In a combined model, only coping skills specifically predicted next session symptom change. Additionally, these coping skills were especially helpful when combined with a better therapeutic relationship quality. On a between-patient level, better therapeutic alliance and more coping skills were associated with lower symptom scores during treatment, while deeper emotional involvement was associated with higher symptom scores. Testing these between-patient effects in a combined model left only coping skills (the more, the greater symptom improvement) and emotional involvement (the deeper, the less symptom improvement) as significant predictors. These two also exhibited a combined effect on symptom change on the between-patient level. DISCUSSION The results highlight the importance of a thorough disaggregation of within- and between-patient variability in psychotherapy process-outcome research as well as the consideration of several potentially important time-varying covariates. While coping skills showed to be the most central for subsequent symptom change, therapeutic relationship quality only seemed to be a facilitative factor in enhancing these effects, but was not sufficiently helpful on its own.