Skip to main content

Table 3 Patient’s parameter values (% yes) for the most discriminative [5] predictors obtained for patients grouped using hierarchical clustering into 2 and 5 clusters

From: Machine learning profiles of cardiovascular risk in patients with diabetes mellitus: the Silesia Diabetes-Heart Project

Parameter

Two-group clustering

Five-group clustering

Cluster 1 (n = 70)

Cluster 2 (n = 168)

p-value

Cluster 1 (n = 28)

Cluster 2 (n = 56)

Cluster 3 (n = 42)

Cluster 4 (n = 39)

Cluster 5 (n = 73)

p-value

Beta blocker

8 (11.43%)

75 (44.64%)

 < 0.0001

9 (32.15%)

24 (42.86%)

3 (7.14%)

7 (17.95%)

2 (2.74)

 < 0.0001

Age

28.79 ± 7.57

61.89 ± 8.70

 < 0.0001

51.75 ± 2.15

60.05 ± 2.28

23.26 ± 3.04

39.31 ± 4.80

69.72 ± 4.60

 < 0.0001

Current left foot ulceration

1 (1.43%)

3 (1.79%)

0.845

0 (0.00%)

1 (1.79%)

0 (0.00%)

1 (2.56%)

2 (2.74%)

0.759

ACEi

3 (4.28%)

77 (45.83%)

 < 0.0001

14 (50.00%)

22 (39.39%)

0 (0.00%)

9 (23.08%)

35 (47.95%)

 < 0.0001

Healed foot ulceration

4 (5.71%)

3 (1.78%)

0.102

1 (3.57%)

1 (1.78%)

1 (2.38%)

4 (10.26%)

0 (0.00%)

0.043

CV event

0 (0.00%)

53 (31.54%)

 < 0.0001

3 (10.71%)

20 (35.71%)

0 (0.00%)

2 (5.13%)

28 (38.36%)

 < 0.0001

  1. The p-values were calculated using Mann–Whitney U-test, χ2 test, or Kruskal–Wallis tests, where appropriate