Related Experiment Video
Updated: May 27, 2025
A Modified Sonographic Algorithm for Image Acquisition in Life-Threatening Emergencies in the Critically Ill Newborn
Published on: April 7, 2023
Dynamic mortality prediction in critically Ill children during interhospital transports to PICUs using explainable AI
Zhiqiang Huo1,2,3, John Booth4, Thomas Monks5
1Institute of Health Informatics, University College London, London, UK.
Insights
This study introduces PROMPT, a machine learning tool predicting 30-day mortality risk for critically ill children during inter-hospital transfers. PROMPT offers real-time, data-driven insights to improve patient care and outcomes.
Area of Science:
- Pediatric Intensive Care
- Medical Informatics
- Machine Learning in Healthcare
Background:
- Critically ill children transferred between hospitals face higher mortality risks.
- Current transport assessments lack real-time mortality risk prediction.
- There is a need for data-driven tools to monitor patients during transport.
Purpose of the Study:
- To introduce the Patient-centred Real-time Outcome monitoring and Mortality PredicTion (PROMPT) pipeline.
- To develop an explainable machine learning model for forecasting 30-day mortality risk in pediatric patients during inter-hospital transfers.
- To provide real-time, individualized mortality risk assessment during transport.
Main Methods:
- Developed an end-to-end machine learning pipeline (PROMPT).
- Integrated continuous time-series vital signs, medical records, and transport data.
- Utilized random forest and logistic regression models for mortality prediction.
Main Results:
- PROMPT demonstrated proof-of-principle in predicting mortality risk.
- Random forest model achieved an AUROC of 0.83 (95% CI: 0.79-0.86).
- Logistic regression model achieved an AUROC of 0.81 (95% CI: 0.76-0.85).
Conclusions:
- PROMPT provides real-time mortality risk prediction for transported critically ill children.
- The model offers individual-level interpretability during inter-hospital transports.
- This tool addresses the gap in data-driven assessment for pediatric transport.
Abstract:
Critically ill children who require inter-hospital transfers to paediatric intensive care units are sicker than other admissions and have higher mortality rates. Current transport practice primarily relies on early clinical assessments within the initial hours of transport. Real-time mortality risk during transport is lacking due to the absence of data-driven assessment tools. Addressing this gap, our research introduces the PROMPT (Patient-centred Real-time Outcome monitoring and Mortality PredicTion), an explainable end-to-end machine learning pipeline to forecast 30-day mortality risks. The PROMPT integrates continuous time-series vital signs and medical records with episode-specific transport data to provide real-time mortality prediction. The results demonstrated that with PROMPT, both the random forest and logistic regression models achieved the best performance with AUROC 0.83 (95% CI: 0.79-0.86) and 0.81 (95% CI: 0.76-0.85), respectively. The proposed model has demonstrated proof-of-principle in predicting mortality risk in transported children and providing individual-level model interpretability during inter-hospital transports.
