Related Experiment Video
Updated: Sep 29, 2025
Assessment of Child Anthropometry in a Large Epidemiologic Study
Published on: February 2, 2017
Development and Evaluation of an Automated Approach to Detect Weight Abnormalities in Pediatric Weight Charts
Lei Liu1,2, Danny T Y Wu1,3, S Andrew Spooner2,3
1Department of Biomedical Informatics, College of Medicine, University of Cincinnati, Cincinnati, OH.
Insights
This study developed an automated weight abnormality detector (AWAD) using machine learning to accurately analyze pediatric weight data. The AWAD significantly improves detection of abnormal weight values, enhancing clinical research and patient safety.
Area of Science:
- Biomedical Informatics
- Machine Learning in Healthcare
- Pediatric Health Data Analysis
Background:
- Inaccurate body weight measurements pose risks in clinical settings and impede research.
- Pediatric weight charts require robust analysis for accurate growth monitoring.
Purpose of the Study:
- To develop a machine learning-based automated weight abnormality detector (AWAD).
- To analyze pediatric weight chart growth dynamics and identify abnormal weight values.
- To improve the reliability of weight data for clinical research and healthcare.
Main Methods:
- Developed an automated weight abnormality detector (AWAD) utilizing machine learning algorithms.
- Employed deep learning models, specifically bi-directional long short-term memory networks.
- Evaluated model performance on real-world clinical data using reference-standard benchmarks.
Main Results:
- Machine learning models demonstrated significant capacity in detecting weight abnormalities.
- The bi-directional long short-term memory network achieved high predictive performance (AUCs ≥0.989).
- The system showed potential for over 98% reduction in screening effort for weight abnormalities.
Conclusions:
- The developed AWAD effectively detects abnormal weight values in pediatric data.
- AWAD significantly outperforms existing literature methods for weight abnormality detection.
- The system holds substantial potential to advance clinical research and healthcare delivery reliant on accurate weight measures.
Abstract:
Inaccurate body weight measures can cause critical safety events in clinical settings as well as hindering utilization of clinical data for retrospective research. This study focused on developing a machine learning-based automated weight abnormality detector (AWAD) to analyze growth dynamics in pediatric weight charts and detect abnormal weight values. In two reference-standard based evaluation of real-world clinical data, the machine learning models showed good capacity for detecting weight abnormalities and they significantly outperformed the methods proposed in literature (p-value<0.05). A deep learning model with bi-directional long short-term memory networks achieved the best predictive performance, with AUCs ≥0.989 across the two datasets. The positive predictive value and sensitivity achieved by the system suggested more than 98% screening effort reduction potential in weight abnormality detection. Consequently, we hypothesize that the AWAD, when fully deployed, holds great potential to facilitate clinical research and healthcare delivery that rely on accurate and reliable weight measures.
