Using Electronic Health Record Data to Predict Deterioriation in Hospitalized Children

About this resource

In this talk from the Machine Learning for Medical Imaging (ML4MI) community, Anoop Mayampurath (PhD) discusses the use of electronic health record (EHR) data and machine learning to predict clinical deterioration in hospitalized children. The presentation explores how traditional methods like the Pediatric Early Warning System (PEWS) fall short and introduces a novel model, pCART, which significantly improves outcomes by enabling early and accurate detection of at-risk patients. pCART (Pediatric Cardiac Arrest Risk Tool) is a gradient boosted tree model designed to identify clinical deterioration in hospitalized children, particularly those at risk of requiring ICU transfers. Unlike traditional methods like the Pediatric Early Warning System (PEWS), which rely on static, age-dependent cutoffs and subjective assessments, pCART utilizes advanced analytics and continuous tracking to provide a more accurate and actionable risk assessment.

A netID is required to view ML4MI videos: View 2024-10-14 recording.

See also

• ML4MI: Explore other talks from the ML4MI group at UW-Madison.