Skip to main content

Table 2 Logistic regression results for models fit on the over/under-sampled cohorts for each outcome

From: Using a cohort study of diabetes and peripheral artery disease to compare logistic regression and machine learning via random forest modeling

Covariate

Amputation

Death

Reintervention

OR

t-stat

p-value

OR

t-stat

p-value

OR

t-stat

p-value

Foot Exam

1.16

10.49

< 0.001

1.01

0.74

0.457

0.73

-22.84

< 0.001

HbA1c

1.36

21.23

< 0.001

0.49

-47.56

< 0.001

1.29

17.90

< 0.001

Vascular Image

0.90

-7.66

< 0.001

1.54

28.89

< 0.001

1.18

11.55

< 0.001

Urban

0.59

-31.33

< 0.001

1.01

0.62

0.534

0.84

-10.43

< 0.001

Comorbidity Count

1.09

19.27

< 0.001

1.25

54.12

< 0.001

1.13

30.49

< 0.001

Dual Eligible

1.09

4.80

< 0.001

1.76

33.05

< 0.001

0.86

-9.05

< 0.001

White

0.84

-9.54

< 0.001

1.26

11.18

< 0.001

0.62

-25.99

< 0.001

Female

0.56

-40.52

< 0.001

0.80

-14.69

< 0.001

0.71

-24.59

< 0.001

Age in years

 65–69

(reference)

(reference)

(reference)

 70–74

0.62

-24.85

< 0.001

1.23

8.42

< 0.001

0.85

-8.60

< 0.001

 75–79

0.52

-31.36

< 0.001

1.63

19.67

< 0.001

0.77

-12.36

< 0.001

 80–84

0.44

-35.87

< 0.001

2.45

36.31

< 0.001

0.71

-15.20

< 0.001

 85 + 

0.29

-53.17

< 0.001

4.42

65.84

< 0.001

0.46

-34.40

< 0.001