Augmenting Autism Support: Optimizing the Use of Large Language Models Through Effective Prompt Chaining to Enhance Communication of Therapeutic Guidelines

Poster Number: P191
Presentation Time: 05:00 PM - 06:30 PM

Abstract Keywords: Delivering Health Information and Knowledge to the Public, Large Language Models (LLMs), Information Extraction, Clinical Guidelines, Information Retrieval, Documentation Burden
Primary Track: Applications
Programmatic Theme: Clinical Informatics

The complexity of autism, coupled with clinicians' limited time to provide personalized advice for each child, complicates responding to caregivers' frequent inquiries through traditional communication methods. To streamline the process of responding to caregivers' questions, we introduce a Large Language Model (LLM) powered Retrieval Augmented Generation (RAG) with a three-stage prompt chaining process for the generation of autism treatment guidelines. Initial results indicated a preference among clinicians for our RAG model over a general LLM.

Speaker(s):
Deshan Wattegama, MS
University of Missouri

Author(s):
Deshan Wattegama, MS - University of Missouri; Benjamin Black, MD - University of Missouri; Elly Ranum, MD - University of Missouri; Chi-Ren Shyu, PhD, FACMI, FAMIA - University of Missouri-Columbia;