.png)
Symposia
Treatment - CBT
Danilo Moggia, Ph.D. (he/him/his)
University of Trier, Germany
Trier, Rheinland-Pfalz, Germany
Brian Schwartz, Ph.D.
Postdoctoral Researcher
University of Trier, Germany
Trier, Rheinland-Pfalz, Germany
Antonia Vehlen, PhD (she/her/hers)
Postodoctoral Researcher
University of Trier
Trier, Rheinland-Pfalz, Germany
Wolfgang Lutz, PhD
Professor
University of Trier
Trier, Rheinland-Pfalz, Germany
Background: Data-informed psychological therapy involves the use of digital technologies and statistical algorithms in the context of measurement-based care to inform clinical decision making. One of these technologies is feedback systems designed to monitor treatment progress. Feedback systems have been shown to help clinicians identify patients who do not improve or deteriorate during treatment so that clinicians can address barriers to progress and prevent dropout. Nevertheless, little is known about the patterns of change that patients who are at risk of deterioration follow once they have been identified.
Method: An archived dataset from a university outpatient clinic that implemented a comprehensive feedback system was analyzed. The sample consisted of 573 patients receiving CBT who were not-on-track (NOT) during the first ten therapy sessions. Growth mixture modeling was used to identify patterns of change ten sessions after the therapist first received the NOT alert. Patient characteristics (e.g., symptom severity) were analyzed as predictors of patient class membership. The association of the classes with final treatment outcome and dropout were studied.
Results: Four classes were identified in terms of symptom severity and response 10 sessions after the NOT alert: 1) Moderate-Severe non-responders (24.6%), 2) Moderate-Severe fast responders (12.7%), 3) Mild steady responders (62.3%), 4) Severe fast responders (0.3%). Moderate-Severe and Severely impaired classes were associated with more than one NOT alert, poorer treatment outcome and higher dropout rates. However, a fast response after the NOT alert was associated with a good prognosis. Therapist reported usefulness and attitudes toward feedback moderated these effects.
Conclusion: The results show considerable variability in response among patients identified as NOT. Even though most patients tended to respond, the response was moderated by symptom severity.