Icon Legend

This session is not in your schedule.

This session is in your schedule. Click again to remove it.

Presentation Icons

Additional registration fee required

Faculty have requested this content not be shared outside of the session

CME Credit Offered

Abstract Award

Recording available 2/16-5/2

AIUM Credit

Foundation Awardee

Poster Icons

Abstract Award

Foundation Awardee

Poster Session 2

Category: Digital Health Technologies (DHT)

Poster Session 2

(535) Utilization of GPT for Enhanced Safety and Quality Review in Healthcare

Wednesday, February 11, 2026

3:30 PM - 5:00 PM

Submitting Author and Presenting Author(s)

Julia Kim, MD

Fellow, Maternal Fetal Medicine
NYU Grossman Long Island School of Medicine
Mineola, New York, United States

Coauthor(s)

IH

Isabella Harnick

NYU Langone Hospital - Long Island
Mineola, New York, United States
KW

Karyn Wat, MD

NYU Langone Hospital - Long Island
Mineola, New York, United States
EB

Erika Banks, MD

NYU Grossman Long Island School of Medicine
Mineola, New York, United States
GS

Genevieve Sicuranza, MD

NYU Grossman Long Island School of Medicine
Mineola, New York, United States
HH

Hye J. Heo, MD

Associate Professor, Department of Obstetrics and Gynecology
NYU Grossman Long Island School of Medicine
Mineola, New York, United States

Objective:
The purpose of root cause analysis (RCA) in healthcare is to systematically investigate adverse events or near misses and identify root causes. This allows for targeted interventions to prevent future occurrences, improving patient safety and overall care quality. Yet individual review of cases may miss opportunities to identify common themes to help prioritize and direct intervention. The advent of artificial intelligence (AI) technologies, such as Generative Pre-trained Transformers (GPT), offers a promising avenue. In 2023, NYU Langone launched a licensed, HIPAA compliant GPT-4. This study evaluated the feasibility and utility of GPT as a tool to enhance safety and quality review processes within healthcare organizations.

Study Design:
A retrospective analysis evaluated event summaries of RCA cases with delay in timely delivery of category II tracings from 2020-2025 using the NYU GPT ecosystem. The model was tasked with identifying common themes and suggesting risk reduction strategies. The AI-based analysis was compared with traditional review methods. Fisher's exact test was used with p< 0.05.

Results:
GPT-4 was given 11 case summaries and 8 prompts, including adjustments to prioritize themes or provide interventions. Approximately 30 minutes was used to prompt and review responses from GPT while >8 hours were spent on manual review. GPT recognized that 1 of 11 cases did not pertain to category II tracings as it was incorrectly included in the initial query. It identified 7 main themes that contributed to delay in delivery in the setting of category II tracings. Comparison of GPT to manual review revealed similarities in outcomes.

Conclusion:

GPT accurately evaluated a small cohort of RCA cases, identifying recurrent themes and objective areas of improvement, in significantly less time than manual review. Integration of GPT into safety and quality review processes can be a feasible tool to identify and prioritize interventions that can be applied to more common RCA themes within larger datasets and across multiple specialties.