Smart Triage Model Validation for Pediatric Patients in Uganda and Kenya

Zhang, Cherri externally validated the Smart Triage clinical prediction model for critically ill children. The analysis included 11,595 participants under the age of five from three hospitals in Uganda and two in Kenya, evaluating model performance by age group. The dataset was last updated on October 13, 2025.

Use Cases

• Validating clinical prediction models based on age-stratified pediatric data.
• Assessing model discrimination and calibration based on AUROC and sensitivity/specificity metrics.
• Revising model coefficients and thresholds for specific age groups like neonates.
• Conducting geographical validation of triage models in East African healthcare contexts.

Strengths

• Includes 11,595 patient records for children under five.
• Provides specific AUROC performance metrics with confidence intervals for different age groups.
• Reports sensitivity and specificity values at defined risk thresholds.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect geographic bias inherent to the five hospitals in Uganda and Kenya.

Provenance

Source

Borealis Harvested Dataverse

Collection Method

Combined prospective baseline cohort from three hospitals in Uganda and two in Kenya.

Freshness

Last updated 2025-10-13 00:58:34; freshness should be verified.

Geography

Uganda and Kenya