Comparative Evaluation of Different Pediatric Early Warning Scores in Predicting Clinical Outcomes in Pediatric
Merve Tazegül1, Aysu Türkmen Karaağaç, Şirin Güven
1Pediatry Department, University of Health Sciences Sancaktepe İlhan Varank Research and Training Hospital, Istanbul, Turkey.
Insights
The Parshuram pediatric early warning score (PEWS) model showed the best accuracy in predicting patient outcomes in our pediatric emergency department. Tailoring PEWS to local patient characteristics is recommended for optimal use.
Area of Science:
- Pediatric Emergency Medicine
- Clinical Assessment Tools
- Patient Monitoring
Background:
- Pediatric early warning scores (PEWS) aim to predict clinical deterioration in pediatric emergency departments (PEDs).
- Existing PEWS models lack comprehensive comparative accuracy studies for predicting clinical outcomes.
- This study compares the diagnostic performance of four PEWS models in a tertiary hospital's PED.
Purpose of the Study:
- To evaluate and compare the diagnostic accuracy of four pediatric early warning score (PEWS) models.
- To determine the most effective PEWS model for predicting clinical outcomes in a pediatric emergency department setting.
Main Methods:
- A prospective observational cohort study included 193 patients over four months.
- Key outcomes measured were discharge, ward admission, intensive care unit (ICU) admission, and mortality.
- Statistical analysis included sensitivity, specificity, likelihood ratios, and area under the receiver operating characteristic curve (AUC) for each PEWS model.
Main Results:
- The Parshuram PEWS model demonstrated the highest diagnostic accuracy among the evaluated models.
- For ward admissions (threshold ≥4), the Parshuram model achieved an AUC of 0.678 (sensitivity 72.34%, specificity 63.27%).
- For ICU admissions (threshold ≥8), the Parshuram model showed an AUC of 0.816 (sensitivity 68.75%, specificity 85.71%).
Conclusions:
- Pediatric early warning scores (PEWS) are valuable for assessing outcomes in pediatric emergency departments.
- The Parshuram PEWS model exhibited superior diagnostic performance in this specific patient population.
- Optimal selection and threshold setting for PEWS should consider the unique sociodemographic and clinical characteristics of each healthcare setting.
Background:
Pediatric early warning scores (PEWS) are designed to predict clinical deterioration and enhance patient management in emergency observation units. Several PEWS models exist, but few studies have compared their accuracy in predicting clinical outcomes. This study evaluated and compared the diagnostic performance of 4 PEWS models (Monaghan, Parshuram, Tibbals, Brilli) in our tertiary hospital's pediatric emergency department (PED).
Methods:
A prospective observational cohort study was conducted from January to April 2023, involving 193 patients. The measured outcomes were discharge, ward admission, intensive care unit admission, and mortality. For each PEWS model, sensitivity, specificity, likelihood ratios, and the area under the receiver operating characteristic curve (AUC) were statistically analyzed.
Results:
The median age of the patients was 4 years, with 57% males and 43% females. Primary diagnoses included respiratory, neurological, and gastrointestinal diseases. Ward admissions were 40.41%, and ICU admissions were 8.30%. Among the evaluated PEWS models, the one developed by Parshuram and colleagues exhibited the highest diagnostic accuracy in our PED. For ward admissions at a threshold score ≥4, the AUC was 0.678, with sensitivity 72.34% and specificity 63.27%. For ICU admissions at a threshold score ≥8, the AUC was 0.816, with sensitivity of 68.75% and specificity of 85.71%.
Conclusion:
PEWS are valuable tools for outcome assessment in PEDs. The Parshuram model demonstrated superior diagnostic performance in the studied population. However, the selection of an appropriate PEWS model and determination of suitable threshold scores should be tailored to the specific sociodemographic and clinical characteristics of each health care center.
