Symposia
Telehealth/m-Health
Brian Schwartz, Ph.D.
University of Trier, Germany
Trier, Rheinland-Pfalz, Germany
Wolfgang Lutz, PhD
Professor
University of Trier
Trier, Rheinland-Pfalz, Germany
Background: The Trier Treatment Navigator (TTN) is a data-informed and algorithm-based clinical decision-support system. It includes treatment recommendations before treatment onset, such as recommending the most promising therapeutic strategy, and treatment adaptations for patients at risk for treatment failure during the course of treatment, such as providing therapists with clinical support tools to improve treatment outcome. The presentation will focus on the implementation and prospective evaluation of the TTN in a German outpatient clinic. Additionally, its application during the COVID-19 pandemic, necessary adaptations and further developments to continue providing psychotherapy under crisis conditions will be discussed.
Methods: The development and validation sample consisted of 1234 patients treated for at least 10 sessions with cognitive-behavioral therapy at the University of Trier’s outpatient center. Predictions were based on machine learning-algorithms (e.g., LASSO and nearest neighbors) and inferential statistics. Adaptive predictions to identify patients who are not on track were updated at each session, triggering clinical support tools. The prospective randomized controlled trial (RCT) evaluating the TTN consisted of 538 outpatients, randomized to either having access to the TTN (n = 335) or not (n = 203). In addition, to adapt to crisis conditions, an initial version of a tool was developed that predicts whether a switch to videotherapy will be beneficial or detrimental to individual patients.
Results: The TTN and its implementation into clinical practice will be presented. In the prospective RCT, a differential effect size of about d = 0.3 was found for therapists following the recommended treatment strategy in the first ten sessions. Linear mixed models showed that therapists’ attitudes towards and confidence in using decision-support systems, and their symptom awareness predicted treatment outcome. No main effects of using the TTN were found, however, for therapists who reported a higher usefulness of feedback having access to the TTN resulted in better treatment outcomes. Moreover, patients treated in the recommended setting (face-to-face vs. video therapy) showed better outcomes.
Conclusion: The findings indicate that feedback and decision-support systems can improve treatment outcome. They highlight the importance of prospective studies to evaluate such systems. Implementation issues and adaptations due to the COVID-19 pandemic will be discussed and further developments presented.