LLM Questionnaire Completion for Automatic Psychiatric Assessment

TL;DR

This paper introduces a novel method using Large Language Models to convert unstructured psychological interviews into structured data, improving psychiatric diagnosis accuracy through a data-driven approach.

Contribution

It presents a new framework that leverages LLMs to interpret narrative interviews and predict psychiatric measures, bridging narrative and data-driven mental health assessments.

Findings

Enhanced diagnostic accuracy over baselines

Effective conversion of interviews into structured questionnaires

Bridging narrative and data-driven mental health assessment

Abstract

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.