Related Experiment Video
Updated: May 5, 2026
09:52
Setting Up a Stroke Team Algorithm and Conducting Simulation-based Training in the Emergency Department - A Practical Guide
Published on: January 15, 2017
17.2K
Geographical validation of the Smart Triage Model by age group
Cherri Zhang1, Matthew O Wiens1,2,3, Dustin Dunsmuir1,3
1Institute for Global Health, BC Children's and Women's Hospitals, Vancouver, British Columbia, Canada.
PLOS Digital Health
|July 1, 2024
Summary
A clinical prediction model, Smart Triage, was validated for detecting critically ill children. The updated model shows improved predictive ability across age groups, aiding in reducing child mortality in low-resource settings.
Area of Science:
- Global Health
- Pediatrics
- Medical Informatics
Background:
- Infectious diseases cause significant under-five mortality in low- and middle-income countries.
- Clinical prediction models can improve early detection of critically ill children.
- Existing models have limited external validation, especially for neonatal mortality.
Purpose of the Study:
- To externally validate the Smart Triage clinical prediction model.
- To assess the model's performance in a combined prospective cohort from Uganda and Kenya.
- To evaluate and potentially revise the model for improved accuracy across different pediatric age groups.
Main Methods:
- External validation of the Smart Triage model using a prospective cohort of 11,595 children under five.
- Evaluation of model discrimination using area under the receiver-operator curve (AUROC).
- Calibration plots analyzed across age subsets; model coefficients and thresholds re-estimated for neonates due to performance issues.
Main Results:
- The model showed good discrimination for children under five (AUROC 0.81) but poor discrimination for neonates (AUROC 0.62).
- Sensitivity was 85% for under-five and 68% for neonates at low-risk thresholds.
- After revision, the neonatal model achieved an AUROC of 0.83 with revised low- (13%) and high-risk (41%) thresholds.
Conclusions:
- The updated Smart Triage model demonstrates good predictive performance across pediatric age groups.
- The model can be integrated into local healthcare facility triage guidelines to optimize care.
- Further external validation, particularly for the neonatal component, is recommended.
