Related Experiment Video
Updated: May 3, 2026
A Method for Quantifying Upper Limb Performance in Daily Life Using Accelerometers
Published on: April 21, 2017
Pilot study using machine learning to improve estimation of physical abuse prevalence
Farah W Brink1, Charmaine B Lo2, Steven W Rust3
1Nationwide Children's Hospital, 700 Childrens Drive, Columbus, OH 43205, United States; The Ohio State University College of Medicine, 370 West Ninth Avenue, Columbus, OH 43210, United States.
Insights
Machine learning models can improve the accuracy of identifying child abuse using International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes. This approach enhances abuse detection beyond traditional coding methods.
Area of Science:
- Medical Informatics
- Pediatric Emergency Medicine
- Machine Learning in Healthcare
Background:
- International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes often underestimate physical abuse prevalence.
- Machine learning (ML) offers potential for more accurate abuse estimation by processing diverse data.
Purpose of the Study:
- To demonstrate the feasibility of using ML to identify ICD-10-CM codes associated with child abuse.
- To develop a proof-of-concept model for abuse detection.
Main Methods:
- A LASSO logistic regression model was developed using ICD-10-CM codes and patient age.
- Data from children under 5 years old, evaluated by a child protection team (CPT) between 2016-2020, were analyzed.
- Model performance was assessed using cross-validation (CV) and Receiver Operator Characteristic (ROC) curves.
Main Results:
- The ML model achieved a mean CV AUC of 0.87 in identifying confirmed physical abuse (PA) using diagnosis codes and age.
- Model performance was slightly lower (mean CV AUC = 0.81) for cases lacking specific abuse ICD-10-CM codes.
Conclusions:
- A model utilizing ICD-10-CM codes and age can enhance the accuracy of distinguishing abusive from non-abusive injuries.
- This pilot study represents a foundational step toward improving population-level abuse estimates.
Background:
International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes have been shown to underestimate physical abuse prevalence. Machine learning models are capable of efficiently processing a wide variety of data and may provide better estimates of abuse.
Objective:
To achieve proof of concept applying machine learning to identify codes associated with abuse.
Participants And Setting:
Children <5 years, presenting to the emergency department with an injury or abuse-specific ICD-10-CM code and evaluated by the child protection team (CPT) from 2016 to 2020 at a large Midwestern children's hospital.
Methods:
The Pediatric Health Information System (PHIS) and the CPT administrative databases were used to identify the study sample and injury and abuse-specific ICD-10-CM codes. Subjects were divided into abused and non-abused groups based on the CPT's evaluation. A LASSO logistic regression model was constructed using ICD-10-CM codes and patient age to identify children likely to be diagnosed by the CPT as abused. Performance was evaluated using repeated cross-validation (CV) and Reciever Operator Characteristic curve.
Results:
We identified 2028 patients evaluated by the CPT with 512 diagnosed as abused. Using diagnosis codes and patient age, our model was able to accurately identify patients with confirmed PA (mean CV AUC = 0.87). Performance was still weaker for patients without existing ICD codes for abuse (mean CV AUC = 0.81).
Conclusions:
We built a model that employs injury ICD-10-CM codes and age to improve accuracy of distinguishing abusive from non-abusive injuries. This pilot modelling endeavor is a steppingstone towards improving population-level estimates of abuse.
More Related Videos
05:59Visualization of Intensity Levels to Reduce the Gap Between Self-Reported and Directly Measured Physical Activity
Published on: March 7, 2019
07:31Implementation of a Real-Time Psychosis Risk Detection and Alerting System Based on Electronic Health Records using CogStack
Published on: May 15, 2020
