A Safety-Aware Role-Orchestrated Multi-Agent LLM Framework for Behavioral Health Communication Simulation
This work addresses the problem of simulating supportive behavioral health dialogue for researchers in behavioral health informatics and decision-support, though it is incremental as it builds on existing multi-agent and safety concepts.
The authors tackled the challenge of single-agent LLM systems struggling with diverse conversational functions and safety in behavioral health communication by proposing a safety-aware, role-orchestrated multi-agent framework, which demonstrated clear role differentiation, coherent coordination, and predictable trade-offs in modular orchestration, safety oversight, and response latency compared to a baseline.
Single-agent large language model (LLM) systems struggle to simultaneously support diverse conversational functions and maintain safety in behavioral health communication. We propose a safety-aware, role-orchestrated multi-agent LLM framework designed to simulate supportive behavioral health dialogue through coordinated, role-differentiated agents. Conversational responsibilities are decomposed across specialized agents, including empathy-focused, action-oriented, and supervisory roles, while a prompt-based controller dynamically activates relevant agents and enforces continuous safety auditing. Using semi-structured interview transcripts from the DAIC-WOZ corpus, we evaluate the framework with scalable proxy metrics capturing structural quality, functional diversity, and computational characteristics. Results illustrate clear role differentiation, coherent inter-agent coordination, and predictable trade-offs between modular orchestration, safety oversight, and response latency when compared to a single-agent baseline. This work emphasizes system design, interpretability, and safety, positioning the framework as a simulation and analysis tool for behavioral health informatics and decision-support research rather than a clinical intervention.