To SNAP or Not to SNAP: Examining Safety Net Antibiotic Prescription Patterns in Pediatric Acute Otitis Media Using a Large Language Model

Poster Number: P93
Presentation Time: 05:00 PM - 06:30 PM

Abstract Keywords: Pediatrics, Infectious Diseases and Epidemiology, Large Language Models (LLMs), Natural Language Processing, Healthcare Economics/Cost of Care, Information Extraction, Real-World Evidence Generation
Primary Track: Applications
Programmatic Theme: Clinical Research Informatics

The primary aim of this study is to describe safety net antibiotic prescriptions (SNAPs) practice patterns and pharmacy dispense data among three clinical settings in children aged 6 months to 18 years, using a retrospective, observational cohort study design. A HIPAA-secure large language model (LLM) was utilized for chart review to determine if a prescription was a SNAP in place of human readers. LLM classifications were matched with pharmacy dispense data to examine antibiotic usage.

Speaker(s):
Jessica Pourian, MD
UCSF

Author(s):
Jessica Pourian, MD - UCSF; Ben Michaels, PharmD - UCSF; Augusto Garcia Agundez Garcia, PhD - UCSF; Valerie Flaherman - UCSF;