LLM Questionnaire Completion for Automatic Psychiatric Assessment
This provides a novel framework for mental health assessment by bridging narrative-driven and data-driven approaches, potentially aiding clinicians in psychiatric diagnosis.
The researchers tackled the problem of converting unstructured psychological interviews into structured psychiatric questionnaires using a Large Language Model, achieving enhanced diagnostic accuracy for depression and PTSD compared to baselines.
We employ a Large Language Model (LLM) to convert unstructured psychological interviews into structured questionnaires spanning various psychiatric and personality domains. The LLM is prompted to answer these questionnaires by impersonating the interviewee. The obtained answers are coded as features, which are used to predict standardized psychiatric measures of depression (PHQ-8) and PTSD (PCL-C), using a Random Forest regressor. Our approach is shown to enhance diagnostic accuracy compared to multiple baselines. It thus establishes a novel framework for interpreting unstructured psychological interviews, bridging the gap between narrative-driven and data-driven approaches for mental health assessment.