PSYCHE: A Multi-faceted Patient Simulation Framework for Evaluation of Psychiatric Assessment Conversational Agents

TL;DR

PSYCHE is a comprehensive framework for evaluating psychiatric assessment conversational agents by simulating patient interactions based on detailed psychiatric profiles, enabling standardized, safe, and cost-effective benchmarking.

Contribution

It introduces a novel multi-faceted simulation framework for assessing PACAs, addressing the lack of standardized evaluation methods in psychiatric conversational AI.

Findings

Validated with 10 psychiatrists' assessments

Demonstrated effective simulation of patient profiles

Provided quantitative metrics for PACA evaluation

Abstract

Recent advances in large language models (LLMs) have accelerated the development of conversational agents capable of generating human-like responses. Since psychiatric assessments typically involve complex conversational interactions between psychiatrists and patients, there is growing interest in developing LLM-based psychiatric assessment conversational agents (PACAs) that aim to simulate the role of psychiatrists in clinical evaluations. However, standardized methods for benchmarking the clinical appropriateness of PACAs' interaction with patients still remain underexplored. Here, we propose PSYCHE, a novel framework designed to enable the 1) clinically relevant, 2) ethically safe, 3) cost-efficient, and 4) quantitative evaluation of PACAs. This is achieved by simulating psychiatric patients based on a multi-faceted psychiatric construct that defines the simulated patients' profiles, histories, and behaviors, which PACAs are expected to assess. We validate the effectiveness of PSYCHE through a study with 10 board-certified psychiatrists, supported by an in-depth analysis of the simulated patient utterances.