August 9, 2025

The Best of All Worlds: A Hybrid Approach to Cohort Identification with Rules, Small and Large Language Models.

Key Points

MAIN FINDING: The hybrid strategy effectively combines curated rules and language models for cohort identification.
KEY EVIDENCE: Performance tests using clinical notes from the US VA system revealed improved efficiency and validity.
APPROACH: Integrates small and large language models with tailored rules for more accurate cohort identification.
SIGNIFICANCE: This method addresses the challenges of implementing natural language processing in healthcare settings.

Abstract

Balancing operational feasibility with the performance of natural language processing (NLP) systems is a significant challenge. This study presents a hybrid strategy to integrate manually curated rules, small language model (SLM), and large language model (LLM) for cohort identification tasks. This approach demonstrates superior performance in terms of both computational efficiency and NLP validity, as shown here in two separate tasks using large number of clinical notes from the US Department of Veteran Affairs (VA) Healthcare system.

AI से पूछें

Bookmark