Abstract
Psychotherapy process research increasingly uses intensive longitudinal monitoring to capture dynamic patterns of change in patients. In this study, 283 psychiatric inpatients diagnosed with depression (n = 70), PTSD (n = 148), dissociative disorders (n = 26) or personality disorders (n = 39) completed the Therapy Process Questionnaire each evening over an average of 81.5 days per patient, yielding a total of 23,074 assessment days. We computed eight dynamic process characteristics, including variability, autocorrelation, instability and complexity, for each TPQ scale, both on average per patient and in a moving-window approach. Then, the diagnostic groups were compared via one-way ANOVAs with Bonferroni-Holm correction. No significant differences emerged across diagnoses in any average process characteristics or their change over time, indicating that idiosyncratic within-person process dynamics overshadow diagnostic distinctions. Further, patients' clinical improvement was associated with rising mean levels and declining variability in positive emotions, mindfulness, insight and motivation. Again, these predictions of improvement were not moderated by patients' diagnoses. Our results support a transdiagnostic approach to measurement-based care that leverages individual process characteristics to guide interventions, rather than relying on diagnostic categories to predict therapy trajectories.