Skip to main content

Table 2 Frequency of measurement and item sub-scores of candidate items for Bedside Paediatric Early Warning System score

From: Development and initial validation of the Bedside Paediatric Early Warning System score

All patients

Controls

Urgent ICU admission

   

Item

Proportion of times with measurements

n/N

Mean maximum subscore

Proportion of observation times with measurements

Mean maximum subscore

Proportion of observation times with measurements

Difference of means

P

AU ROC

HR

69.6%

0.87

52.9%

2.45

89.9%

1.58

<0.001

0.814

SBP

33.1%

0.78

23.0%

1.52

45.5%

0.74

<0.001

0.670

CRT

27.4%

0.50

22.3%

1.93

33.6%

1.43

<0.001

0.679

Pulse

18.7%

0.04

 

0.46

 

0.42

<0.001

0.627

RR

48.9%

0.64

32.0%

2.00

69.5%

1.36

<0.001

0.795

Respiratory effort

70.5%

0.20

81.7%

1.77

56.9%

1.56

<0.001

0.786

Saturation

65.1%

0.45

45.3%

1.18

89.2%

0.73

<0.001

0.677

Oxygen therapy

92.9%

0.40

92.3%

2.47

93.6%

2.07

<0.001

0.835

Bolus

13.1%

0.03

10.4%

0.11

12.9%

0.08

0.067

0.542

Temperature

25.2%

0.10

21.5%

0.55

0.323

0.45

<0.001

0.697

      

0.000

  

Infant behaviour scale

8.6%

0.900

5.2%

1.40

14.8%

0.500

0.075

0.563

Bromage sedation

79.5%

0.000

90.6%

0.67

60.4%

0.670

<0.001

0.583

Level of Consciousness

88.1%

0.900

94.8%

1.93

83.8%

1.033

0.004

0.629

  1. We present a description of the candidate items and their performance as discriminators between 60 case patients who were urgently admitted to the intensive care unit (ICU; without 'code blue' event) and 120 control patients who had neither urgent ICU admission nor code blue event. The first, third and fifth columns show the number of measurements expressed as a percentage of the number of times that any clinical measurement was made. For example heart rate (HR) measurements were recorded on 79.6% of occasions when any clinical measurement was made. For each candidate item the mean of the maximum sub-scores for case patients and control patients is presented. The maximum sub-score value from each patient was used in logistic regression to calculate the area under the receiver operating characteristics curve (AUROC) for each candidate item. The difference between these measurements is presented in the column titled difference of means and the P value represents the comparison between case and control patients using logistic regression. CRT = capillary refill time in seconds, RR = respiratory rate; SBP = systolic blood pressure.