Inaccurate weight measures can cause critical safety events on pediatrics due to frequent use of weight-based dosing. This study focused on analyzing pediatric growth charts and automating weight error detection. A machine learning-based approach was developed to predict abnormal weight values based on variables representing weight characteristics and short-term growth dynamics. The automated algorithms showed good capacity for detecting weight errors, which offers the potential to significantly reduce medication safety events among pediatric patients.

Learning Objective: After participating in this session, the learner should be better able to:
1. Leverage machine learning algorithms to predict abnormal weight values.
2. Understand the weight characteristics and growth dynamics that contribute to the risk of weight errors.

Authors:

Lei Liu (Presenter)
University of Cincinnati

PJ Van Camp, University of Cincinnati
C. Monifa Mahdi, Cincinnati Children’s Hospital Medical Center
Stephen Spooner, Cincinnati Children’s Hospital Medical Center
Danny T.Y. Wu, Cincinnati Children’s Hospital Medical Center
Yizhao Ni, Cincinnati Children’s Hospital Medical Center

Presentation Materials:

• Abstract/Manuscript - 1