Skip to main content

Table 3 FEAST Paediatric Emergency Triage (PET) score and the FEAST Paediatric Emergency Triage and Laboratory (PETaL) score

From: Predicting mortality in sick African children: the FEAST Paediatric Emergency Triage (PET) Score

Factor Coefficient (95 % CI) from univariable model Coefficient (95 % CI) from multivariable model a P value from multivariable model Score value given if present Maximum possible value for PET score Maximum possible value for PETaL score
Axillary temperature: ≤37 °C 1.08 (0.86–1.31) 0.63 (0.38–0.87) <0.001 1 1 1
Heart rate: <80 bpm (bradycardia) 2.46 (2.00–2.93) 1.34 (0.92–1.77) <0.001 2 2 2
≥80 to <105 bpm 1.12 (0.57–1.68) 0.70 (0.11–1.30) 0.02 1
≥220 bpm (severe tachycardia) 1.44 (0.73–2.14) 1.34 (0.92–1.77) <0.001 2
Capillary refill time: 2 or more seconds 0.93 (0.63–1.23) 0.53 (0.21–0.85) 0.001 1 1 1
Conscious level: prostrate 1.01 (0.58–1.46) 0.68 (0.23–1.13) 0.003 1 2 2
– coma 2.24 (1.80–2.69) 1.53 (1.06–2.00) <0.001 2
Respiratory distress 0.93 (0.52–1.34) 0.55 (0.07–1.02) 0.02 1 1 1
Lung crepitations 0.77 (0.55–1.01) 0.60 (0.36–0.85) <0.001 1 1 1
Severe pallor 0.90 (0.66–1.14) 0.49 (0.22–0.76) <0.001 1 1 1
Weak pulse 1.45 (1.24–1.68) 0.73 (0.48–0.97) <0.001 1 1 1
Weight: <6 kg 0.52 (0.02–1.03) 0.41 (–0.05–0.88) 0.08  
6–8 kg 0.30 (–0.01–0.61) 0.21 (–0.03–0.45) 0.09    
Deep breathing 1.17 (0.86–1.49) 0.42 (0.06–0.77) 0.02  
Total      10  
Additional laboratory values to be added if measured       
Lactate >5 mmol/L 1.82 (1.54–2.09) 1.12 (0.79–1.46) <0.001 2   2
pH <7.2 1.80 (1.51–2.08) 0.97 (0.69–1.25) <0.001 1   1
Blood urea nitrogen >20 mg/dL 1.23 (0.95–1.52) 0.58 (0.26–0.90) <0.001 1   1
Total (laboratory score)       14
  1. a Coefficient from linear predictor of multivariable cox regression model on complete cases. First section includes clinical factors only. Second section (laboratory values) adjusted for all clinical factors
  2. Note: weight and deep breathing were the least predictive factors and were therefore excluded from the score. Univariable models and multivariable model also adjusted for randomisation arm which was not included in the score