Table 2 Comparison of the discriminative ability for machine learning (ML) based versus conventional Cox proportional-hazard models with baseline 0, 24, 48, and 72 h, respectively, after ICU admission.

From: Discrete-time survival analysis in the critically ill: a deep learning approach using heterogeneous data

  

Prediction window (days)

Method

Baseline

1

7

14

30

90

365

ML

0

0.72 (0.71–0.72)

0.73 (0.72–0.74)

0.73 (0.72–0.73)

0.73 (0.72–0.73)

0.73 (0.72–0.73)

0.73 (0.72–0.73)

ML

24

0.71 (0.71–0.72)

0.71 (0.71–0.72)

0.71 (0.70–0.72)

0.71 (0.71–0.72)

0.72 (0.71–0.72)

0.72 (0.71–0.73)

ML

48

0.71 (0.70–0.72)

0.71 (0.70–0.72)

0.71 (0.70–0.72)

0.71 (0.70–0.72)

0.71 (0.70–0.72)

0.71 (0.70–0.72)

ML

72

0.70 (0.69–0.71)

0.69 (0.68–0.70)

0.69 (0.68–0.70)

0.69 (0.68–0.70)

0.69 (0.68–0.70)

0.70 (0.69–0.70)

Cox

0

0.66 (0.66–0.66)

Cox

24

0.66 (0.66–0.66)

Cox

48

0.66 (0.66–0.66)

Cox

72

0.66 (0.66–0.66)

  1. The concordance index is shown for each prediction window with 95% CIs in parentheses.
Back to article page