Related Experiment Video
Updated: Aug 4, 2025
Experimental Model to Evaluate Resolution of Pneumonia
Published on: February 17, 2023
Machine learning models to predict and benchmark PICU length of stay with application to children with critical
Colin M Rogerson1, Julia A Heneghan2, Joseph G Kohne3,4
1Division of Pediatric Critical Care, Indiana University School of Medicine, Indianapolis, Indiana, USA.
Insights
Machine learning models can predict and benchmark pediatric intensive care unit (PICU) length of stay (LOS) for critical bronchiolitis patients. These models, using administrative data, offer valuable insights for healthcare resource management and patient care strategies.
Area of Science:
- Pediatric critical care medicine
- Health informatics
- Machine learning applications in healthcare
Background:
- Bronchiolitis is a common cause of pediatric intensive care unit (PICU) admission.
- Accurate prediction and benchmarking of PICU length of stay (LOS) are crucial for resource allocation and quality improvement.
- Existing methods for LOS prediction may not fully leverage the potential of administrative databases.
Purpose of the Study:
- To develop and evaluate machine learning models for predicting and benchmarking PICU LOS in pediatric patients with critical bronchiolitis.
- To assess the accuracy of models using different data availability (all hospitalization data vs. admission data only).
Main Methods:
- Retrospective cohort study utilizing the Pediatric Health Information Systems (PHIS) database (2016-2019).
- Development of two random forest models: Model 1 for benchmarking (all data) and Model 2 for prediction (admission data only).
- Model evaluation using R-squared, Mean Squared Error (MSE), and Observed to Expected (O/E) ratio.
Main Results:
- Model 1 demonstrated superior performance with higher R-squared (0.51) and lower MSE (0.21) compared to Model 2 (R-squared: 0.10, MSE: 0.37).
- Both models showed similar O/E ratios (1.18 for Model 1, 1.20 for Model 2), indicating their ability to benchmark LOS.
- Significant institutional variability in O/E ratios (median 1.01, IQR 0.90-1.09) was observed.
Conclusions:
- Machine learning models utilizing administrative data can effectively predict and benchmark PICU length of stay for critical bronchiolitis.
- These models provide a valuable tool for improving the management of pediatric critical care resources.
- Further research can refine these models for enhanced clinical decision support.
Objective:
To create models for prediction and benchmarking of pediatric intensive care unit (PICU) length of stay (LOS) for patients with critical bronchiolitis.
Hypothesis:
We hypothesize that machine learning models applied to an administrative database will be able to accurately predict and benchmark the PICU LOS for critical bronchiolitis.
Design:
Retrospective cohort study.
Patients:
All patients less than 24-month-old admitted to the PICU with a diagnosis of bronchiolitis in the Pediatric Health Information Systems (PHIS) Database from 2016 to 2019.
Methodology:
Two random forest models were developed to predict the PICU LOS. Model 1 was developed for benchmarking using all data available in the PHIS database for the hospitalization. Model 2 was developed for prediction using only data available on hospital admission. Models were evaluated using R2 values, mean standard error (MSE), and the observed to expected ratio (O/E), which is the total observed LOS divided by the total predicted LOS from the model.
Results:
The models were trained on 13,838 patients admitted from 2016 to 2018 and validated on 5254 patients admitted in 2019. While Model 1 had superior R2 (0.51 vs. 0.10) and (MSE) (0.21 vs. 0.37) values compared to Model 2, the O/E ratios were similar (1.18 vs. 1.20). Institutional median O/E (LOS) ratio was 1.01 (IQR 0.90-1.09) with wide variability present between institutions.
Conclusions:
Machine learning models developed using an administrative database were able to predict and benchmark the length of PICU stay for patients with critical bronchiolitis.
