LAUSR

381 ple based on machine learning approaches applied to large datasets in order to recommend the optimal treatment, treatment strategy, or therapist for an individual patient. During treatment, therapists are made aware of patients at risk for treatment failure, dropout or self-harm by adaptive decision tools. Additionally, therapists are provided with feedback and clinical problem-solving tools to support treatment for these patients. Currently, the implementation and prospective evaluation of such systems are rare. However, such studies and new developments are already on their way. For example, more than a decade of our department’s research activity has resulted in the development of a digital decision support and navigation system called the Trier Treatment Navigator (TTN). The system combines outcome tracking, prediction, and prescription tools, providing continuous feedback to clinicians and supporting them to apply targeted clinical strategies at the onset of and during treatment. The online navigation system includes two components of patient-specific treatment recommendations: a) a pre-treatment clinical strategy recommendation and b) adaptive recommendations and support tools for patients at risk for treatment failure. The prospective evaluation on 538 patients showed an advantage in outcomes, with an effect size of about 0.3, when patients were treated with the recommended strategy during the first ten sessions. Furthermore, therapist symptom awareness, attitude, and confidence using the system were found to be significant predictors of outcome, while therapist-rated usefulness of such feedback moderated the feedback-outcome association. A similar approach, the Leeds Risk Index (LRI), was developed based on a sample of 1,347 patients and prognostically tested on 282 patients in the Improving Access to Psychological Therapies (IAPT) programme, to recommend either low or high intensity treatments. Results indicated that such stratified care improves efficiency by generating comparable outcomes with less treatment sessions. The goal of these developments is the time ly translation of research into clinical practice. Of course, many more prospective studies are necessary. However, in the future, the field might be better able to operationalize change processes, regarding both how patients experience them and how therapists induce them. These developments could be the basis of a trans-theoretical, process-based, personalized and data-informed psychological treatment approach, which includes both an idiographic (e.g., intensive longitudinal assessments on single cases) and a nomothetic (e.g., large databases of patients and therapists) perspective. Such advancements could finally make a difference for patients previously not profiting from psychological interventions.