Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

RescueGPT: An Automated System for Detecting Adverse Safety Events in Prehospital Emergency Medical Service Notes

Tina Yi Jin Hsieh1,2, Carl Eriksson3, Garth Meckler4

  • 1Department of Biomedical Informatics Harvard Medical School Boston Massachusetts USA.

Learning Health Systems
|June 8, 2026
PubMed

Related Concept Videos

Pharmacovigilance01:19

Pharmacovigilance

Post-marketing surveillance is a critical component of pharmaceutical regulation, often uncovering unanticipated adverse drug reactions (ADRs) once a drug is widely used over an extended period.
This process, termed pharmacovigilance, aims to detect, evaluate, and minimize harmful effects related to medication use. The data collection for pharmacovigilance depends on spontaneous reporting systems, where healthcare professionals or patients voluntarily report suspected ADRs.
In some cases, there...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

A pilot randomized trial of ketamine for suicidal ideation in a pediatric emergency department.

CJEM·2026
Same author

The Oral Microbiome Is a Population-Scale Readout of the Exposome, Age, and Systemic Health.

bioRxiv : the preprint server for biology·2026
Same author

Standardized Workload Assessment Metric for Pediatric Emergency Departments (SWAMPED): Multicenter Derivation and Evaluation of a Task-Level Workload Measure.

Academic emergency medicine : official journal of the Society for Academic Emergency Medicine·2026
Same author

Eligibility for redirection from the pediatric emergency department: a qualitative study of professional and leader perspectives.

Pediatric research·2026
Same author

Epinephrine vs. Advanced Airway First in Pediatric Out-of-Hospital Cardiac Arrest.

Academic emergency medicine : official journal of the Society for Academic Emergency Medicine·2026
Same author

Effective Community Consultation for Prehospital EFIC Research: Lessons from a 20-Site Pediatric Trial.

Prehospital emergency care·2026
Same journal

Person-Centered Care Planning for People With Multiple Chronic Conditions: An Environmental Scan of Models and Approaches.

Learning health systems·2026
Same journal

The HALO Model: A Learning Health System Framework for Artificial Intelligence.

Learning health systems·2026
Same journal

Surveillance of Pharmaceutical Risk-Mitigation Behavior: Applying and Comparing Statistical Process Control Methods Using Real World Data.

Learning health systems·2026
Same journal

The Value and Challenges of Stakeholder Engagement in Rehabilitation Learning Health Systems: A Qualitative Pilot Study of Rehabilitation Directors.

Learning health systems·2026
Same journal

Shared Experiences and Care Improvement Priorities for Multimorbidity Management: An Experience-Based Co-Design Study.

Learning health systems·2026
Same journal

AI-powered nursing handoffs: Introducing and evaluating the patient report template.

Learning health systems·2026
See all related articles
Summary
This summary is machine-generated.

This study automated adverse safety event (ASE) detection in pediatric out-of-hospital cardiac arrest (OHCA) using a large language model (LLM) and a knowledge base. While effective for common events, accuracy for rare events depends on clear emergency medical service (EMS) documentation.

Area of Science:

  • Emergency Medicine
  • Health Informatics
  • Artificial Intelligence

Background:

  • Manual review of emergency medical service (EMS) data for adverse safety events (ASEs) is time-consuming and limits scalability.
  • Automating ASE identification is crucial for improving patient safety in high-volume pediatric out-of-hospital cardiac arrest (OHCA) data.

Purpose of the Study:

  • To explore the use of a large language model (LLM) with a knowledge base to automate ASE extraction from unstructured EMS notes for pediatric OHCA.
  • To develop and evaluate a system for identifying ASEs in pediatric prehospital care.

Main Methods:

  • Utilized a national EMS provider's pediatric OHCA records (2017-2020) and the Pediatric Prehospital Adverse Safety Event Detection System (PEDS) knowledge base.
  • Developed an ontology using the LinkML framework and employed the Structured Prompt Interrogation and Recursive Extraction of Semantics (SPIRES) method to generate schema-driven prompts for GPT-3.5.
Keywords:
adverse safety events (ASE)emergency medical services (EMS)information extractionlarge language model (LLM)pediatric out‐of‐hospital cardiac arrest (OHCA)

Related Experiment Videos

  • Mapped unstructured EMS narratives to structured concepts for entity extraction.

    • Main Results:

    • The developed framework, RescueGPT, demonstrated high accuracy in detecting common ASEs such as Patient Rhythm, Age, Weight, and Length.
    • Challenges were observed in identifying rare ASEs (e.g., Failure to Establish IV Access, Incorrect Airway Equipment Size) due to inconsistent and complex documentation.
    • Performance varied based on the completeness and clarity of EMS narratives, particularly for rare events.

    Conclusions:

    • RescueGPT shows promise for scaling automated ASE detection in pediatric OHCA.
    • The accuracy of automated ASE detection is significantly influenced by the quality and standardization of clinical documentation in EMS systems.
    • Future work should focus on improving detection of rare events, enhancing transparency, and refining text processing techniques.