Enhancing Patient Intake Process in Mental Health Consultations Using RAG-Driven Chatbot

Abstract

In this paper, we develop and evaluate an empathic chatbot, called EmoBot, for a mental healthcare patient in-take scenario. EmoBot employs Retrieval-Augmented Generation (RAG) methods with Large Language Models (LLMs), to provide empathic support and advice, while ensuring the intake process is completed. We evaluate EmoBot's performance both by automated metrics using available datasets and through dialogues with a simulated patient model that mimics varying levels of depression severity. Our evaluations showed good agreement between EmoBot's categorization of depression severity levels and human raters. EmoBot's topic classification system achieved 78.48% accuracy on the Primate2022 dataset [1] without fine-tuning. EmoBot's responses showed a 0.93 similarity with patient inputs, proving contextual relevance. These findings highlight the effectiveness of RAG-based methods in guiding and providing safety guardrails for LLMs and their potential as supportive tools for health assessment and support.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 1 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.