Online Program Home
My Program

Abstract Details

Activity Number: 359 - Contributed Poster Presentations: Section on Medical Devices and Diagnostics
Type: Contributed
Date/Time: Tuesday, July 30, 2019 : 10:30 AM to 12:20 PM
Sponsor: Section on Medical Devices and Diagnostics
Abstract #306861
Title: Prediction of Pediatric Emergency Department X-Ray and CT Utilization in the United States
Author(s): Xingyu Zhang* and Sheng Yang and Pau Medrano-Gracia and Konrad Werys and Prashant Mahajan
Companies: University of Michigan and University of Michigan and University of Auckland and University of Oxford and University of Michigan
Keywords: natural language processing; predicative model; medical imaging utilization; pediatric emergency department

The aim of the study was to estimate the association between medical imaging utilization, and socioeconomic demographic and clinical factors among pediatric patients visiting the Emergency department (ED), allowing for the development of predictive models. Secondary data analysis was conducted to predict the use of medical imaging in pediatric patients visiting the ED. Multivariate logistic regression models were applied to structured, unstructured and combined data, incorporating natural language processing. Of the 27,665 visits included in the study, 8,394 (30.3%) obtained an imaging diagnosis. The c-statistic was 0.71 for any imaging use, 0.69 for X-ray, and 0.77 for CT, in the predictive model including only structured variables. Models including only unstructured information obtained c-statistics of 0.81, 0.82 and 0.85, respectively. When both structured variables and free text variables were included, the c-statistics reached 0.82 for any imaging use, 0.83 for X-ray, and 0.87 for CT. We present several predictive models for the use of medical imaging in pediatric patients visiting the ED. The inclusion of unstructured data provided significant improvement in accuracy.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2019 program