Processes in psychotherapy: A scoping review with LLM-assisted clustering

Clinical psychological science has shown limited progress in improving treatment efficacy, refining intervention models, and identifying processes of change, that are traditionally associated with common factors (such as therapeutic alliance and empathy).

To examine research trends on this topic, we systematically surveyed the literature for studies that examine processes of change in the context of psychological interventions.

A total of 778 studies reported on 684 processes of therapeutic change since 2007. Using an iterative, AI-assisted human-in-the-loop clustering procedure, these processes were subsequently organized into 32 process clusters using OpenAI's GPT-5 nano model.

The largest process cluster identified concerned common factors (i.e., therapeutic alliance and collaborative processes, and interpersonal functioning), accounting for 20.6% of all investigated processes of change. The remaining processes were primarily associated with specific factors related to cognitive behavioral therapy, such as cognitive appraisal and belief change processes. The research focus and number of studies on therapeutic processes have not changed substantially over the years.

Despite urgent calls to improve our understanding of therapeutic processes, the focus and volume of research have remained unchanged, with the primary focus remaining on common factor processes.