Ambalavanan et al developed CART (classification and regression tree) models for predicting the outcome in a neonate with hypoxic-ischemic encephalopathy. This can help identify an infant who requires more aggressive management. The authors are from multiple universities in the United States participating in the National Institute of Child Health and Human Development Neonatal Research Network.
Outcome scores:
(1) death or disability
(2) death
Parameters:
(1) pH in cord blood gas
(2) PaCO2 in mm Hg in cord blood gas
(3) base deficit in first postnatal arterial blood gas in mmol/L (acidosis)
(4) spontaneous activity
CART model for disability or death
pH |
spontaneous activity |
base deficit |
PaCO2 |
Probability |
<= 6.7 |
NA |
NA |
NA |
91% |
> 6.7 |
N |
NA |
NA |
76% |
> 6.7 |
Y |
> 18.5 |
NA |
55% |
> 6.7 |
Y |
<= 18.5 |
<= 87 mm Hg |
32% |
> 6.7 |
Y |
<= 18.5 |
> 87 mm Hg |
12% |
where:
• The lower probability associated with a higher PaCO2 in the final node is not intuitive.
CART model for death
base deficit |
pH |
Probability |
> 24.5 mmol/L |
NA |
72% |
<= 24.5 mmol/L |
<= 6.74 |
48% |
<= 24.5 mmol/L |
> 6.74 |
15% |
Specialty: Neurology, Pedatrics