This page was translated automatically from Turkish. Read the original

This page is an AI-assisted Turkish summary of the study presented at the 26th World Congress of Psychiatry. The study belongs to the presenting authors; please refer to the original source when citing it.

EV0444

A machine learning–based clinical decision support system in internet-delivered cognitive behavioral therapy for depression and anxiety: a triple-blind randomized clinical trial

A MACHINE LEARNING BASED CLINICAL DECISION SUPPORT SYSTEM FOR INTERNET-DELIVERED COGNITIVE BEHAVIORAL THERAPY FOR DEPRESSION AND ANXIETY – A TRIPLE BLIND RANDOMIZED CLINICAL TRIAL

An artificial intelligence–based clinical decision support system showed no clear effect on treatment outcomes in patients receiving internet-delivered cognitive behavioral therapy.

The aim was to test the clinical utility of an artificial intelligence–based clinical decision support system (CDSS) based on routine outcome monitoring (ROM), and its impact on therapists and patients. The machine learning model underlying the system had been trained on data from more than 6,500 patients with depression and anxiety and generated automated predictions of treatment outcomes. A total of 399 patients with depression, panic disorder, or social anxiety and 70 inexperienced internet-delivered cognitive behavioral therapy (ICBT) therapists were randomized to standard ICBT or ICBT augmented with an artificial intelligence–based CDSS in a triple-blind design. The CDSS summarized weekly measurements using four colors and provided therapists with recommendations for adapting treatment: green (success likely), yellow (uncertain for guidance), light red (failure likely), and dark red (failure very likely). In the comparison between groups involving red-coded patients, the between-group difference was not significant (p = 0.09); a significant difference was found after controlling for the time point at which the patient was coded red. The presenters discussed the weak effect in relation to therapists’ inexperience, the possibility that patients with the greatest need had not been adequately identified, the fact that ICBT already includes a modest ROM effect, and the greater complexity of mood and anxiety disorders in terms of their suitability for treatment adaptation. In conclusion, this controlled study did not demonstrate a clear benefit of the CDSS on treatment outcomes.

All WCP 2026 summaries