Chandna et al report a model for predicting the risk of death for a child in an ICU with limited resources. This can help to identify a patient who may require more aggressive management. The authors are from Angkor Hospital for Children in Cambodia and the University of Oxford.
Patient selection: child in pediatric ICU in a resource-limited setting, from 1 month to 16 years
Parameters:
(1) age in months
(2) comorbidity
(3) z-score for weight for age
(4) estimated travel time in minutes
(5) intra-hospital transfer
(6) heart rate in beats per minute
(7) prolonged capillary refill time (CRT; > 2 seconds)
(8) respiratory rate in breaths per minute
(9) oxygen saturation in percent
(10) supplemental oxygen therapy
(11) abnormal mental status (Glasgow coma scale < 15)
|
Parameter
|
Finding
|
Points
|
|
age in months
|
|
-0.0013 * (age)
|
|
comorbidity
|
no
|
0
|
|
|
yes
|
0.2201
|
|
z score for weight for age
|
|
-0.1108 * (z score)
|
|
estimated travel time
|
|
0.0013 * (time)
|
|
intra-hospital transfer
|
no
|
0
|
|
|
yes
|
0.4626
|
|
heart rate
|
|
-0.0002 * (heart rate)
|
|
prolonged capillary refill
|
no
|
0
|
|
|
yes
|
1.0244
|
|
respiratory rate
|
|
0.009 * (respiratory rate)
|
|
oxygen saturation
|
|
-0.0253 * (percent)
|
|
supplemental oxygen
|
no
|
0
|
|
|
yes
|
0.1808
|
|
abnormal mental status
|
no
|
0
|
|
|
yes
|
1.0313
|
|
Combination Terms
|
Points
|
|
age and heart rate
|
0.0000086514 * (age) * (heart rate)
|
|
age and respiratory rate
|
0.000048939 * (age) * (resp rate)
|
value of X =
= SUM(points for all of the parameters) - 1.8525
probability of death =
= 1 / ( 1 + EXP((-10 * X))
Performance:
• The area under the ROC curve is 0.84.
