Predicting Treatment Response Using Machine Learning: A Registered Report

Abstract

ObjectivePrevious research on psychotherapy treatment response has mainly focused on outpatients or clinical trial data which may have low ecological validity regarding naturalistic inpatient samples. To reduce treatment failures by proactively screening for patients at risk of low treatment response, gain more knowledge about risk factors, and to evaluate treatments, accurate insights about predictors of treatment response in naturalistic inpatient samples are needed. MethodsWe will compare the performance of different machine learning algorithms in predicting treatment response, operationalized as a substantial reduction in symptom severity as expressed in the Patient Health Questionnaire Anxiety and Depression Scale. To achieve this goal, we will use different sets of variables—(a) demographics, (b) medical indicators, (c) psychological indicators, and (d) treatment-related variables—in an inpatient sample to specify their joint and unique contribution to treatment success.

Machine Learning Meets Psychological Data: Challenges and Future Directions

“…For example, one method that emerged out of the replication crisis in psychology to make explicit the effect of researchers' degrees of freedom, and which can also now be applied to machine learning, is the multi-verse analysis (Bell et al, 2022). Another example is pre-registration, or registered reports, a practice increasingly common in psychology in the wake of the replication crisis, but currently, there are limited guidelines on how to pre-register studies using machine learning (Nosek et al, 2018) (but see the workshop at NeurIPS 2021 and Jankowsky et al (2022)). The exchange of ideas in this area could help research at the intersection of the two fields to establish itself as a rigorous and reputable research area -drawing on out-of-sample evaluation of models from machine learning, combined with theoretical/causal reasoning and pre-registration from psychology.…”