Why are some cases not on track? An item analysis of the assessment for signal cases (ASC) during inpatient psychotherapy

Within the Routine Outcome Monitoring System “OQ‐Analyst”, the questionnaire “Assessment for Signal Cases” (ASC) supports therapists in detecting potential reasons for not‐on‐track trajectories. Factor analysis and a machine learning algorithm (Lasso with 10‐fold cross‐validation) were applied and potential predictors of not‐on‐track classifications were tested using logistic multilevel modelling methods. The factor analysis revealed a shortened (30‐item) version of the ASC with good internal consistency (α = 0.72 – 0.89) and excellent predictive value (AUC = 0.98; +PV = 0.95; ‐PV = 0.94). Item‐level analyses showed that interpersonal problems captured by specific ASC items (not feeling able to speak about problems with family members; feeling rejected or betrayed) are the most important predictors of not‐on‐track trajectories.