Poster #70: Alexis Schulz

Automating ELSO Registry Data Collection Through Event-Centered Data Management and Large Language Model-Assisted Abstraction

Alexis V. Schulz BS, Elizabeth M. Malick MEd, RRT-NPS, CES-P, Andrew F. Lenart MPH, Michael P. Regan BSN, RN, Nimisha Khaitan MS, Deepak Kumar Nandagopal, Erika A. Pleskunas BS, Danielle M. Traynor MSN, RN, CCRN-K, CPHQ, Adam S. Himebauch MD, MSCE, Natalie Rintoul MD, Holly L. Hedrick MD, FACS, Garrett Keim MD, MSCE, Tiffany S. Ko PhD, Thomas A. Reynolds MFA, MBA; All authors affiliated with Children’s Hospital of Philadelphia, Philadelphia, PA

The Extracorporeal Life Support Organization (ELSO) Registry consists of 274 data elements across >40 tables and is internationally accepted as the central extracorporeal membrane oxygenation (ECMO) outcomes registry. ELSO requires Registry submission following a patient’s cessation of support, but current methods of abstraction are labor-intensive and time-consuming. Required elements are stored in structured, semi-structured, and unstructured formats within the electronic health record (EHR). Queries of structured EHR data alone is insufficient making automation challenging. The Clinical Outcomes Data Archive (CODA) is a software platform designed with a patient-centered, longitudinal approach to data management and can capture previously inaccessible, clinically nuanced data expeditiously. record (EMR) fields. II.