Table 3 One-year outcomes of patient groups according to the clustering method (k = 3/test set).

From: Neural network-based clustering model of ischemic stroke patients with a maximally distinct distribution of 1-year vascular outcomes

 

Cluster 0

Cluster 1

Cluster 2

P-value

(A) K-prototype

N

949

168

413

 

Primary outcome

136 (14.3%)

16 (9.5%)

49 (11.9%)

0.158

Stroke

26 (2.7%)

3 (1.8%)

14 (3.4%)

0.557

MI

0 (0.0%)

0 (0.0%)

2 (0.5%)

0.067

All-cause mortality

117 (12.3%)

14 (8.3%)

37 (9.0%)

0.096

(B) SSC-Bair

N

164

952

414

 

Primary outcome

15 (9.1%)

136 (14.3%)

50 (12.1%)

0.150

Stroke

3 (1.8%)

26 (2.7%)

14 (3.4%)

0.579

MI

0 (0.0%)

0 (0.0%)

2 (0.5%)

0.067

All-cause mortality

13 (7.9%)

117 (12.3%)

38 (9.2%)

0.100

(C) DLC-MMD

N

966

527

37

 

Primary outcome

69 (7.1%)

117 (22.2%)

15 (40.5%)

 < 0.001

Stroke

20 (2.1%)

23 (4.4%)

0 (0.0%)

0.022

MI

1 (0.1%)

1 (0.2%)

0 (0.0%)

0.885

All-cause mortality

52 (5.4%)

101 (19.2%)

15 (40.5%)

 < 0.001

(D) DLC-Kuiper

N

156

677

697

 

Primary outcome

65 (41.7%)

91 (13.4%)

45 (6.5%)

 < 0.001

Stroke

8 (5.1%)

23 (3.4%)

12 (1.7%)

0.031

MI

0 (0.0%)

2 (0.3%)

0 (0.0%)

0.283

All-cause mortality

63 (40.4%)

69 (10.2%)

36 (5.2%)

 < 0.001

  1. P-value; Chi-squared test.