Table 2 Final 9 predictors selected for the logistic regression model using stepwise regression with Bayesian Information Criterion

From: Interpretable machine learning models for stroke risk prediction in patients with newly diagnosed atrial fibrillation

 

Univariate analysis

Multivariable analysis

Wald’s test

LR test

 

Odds Ratio (95%CI)

Odds Ratio (95%CI)

P value

P value

OHA

1.97 (1.68–2.31)

1.20 (0.84–1.72)

0.324

0.324

Warfarin

2.33 (1.98–2.75)

1.75 (1.32–2.32)

<0.001

<0.001

APA

2.06 (1.82–2.33)

1.24 (1.01–1.54)

0.044

0.045

AAD

0.79 (0.66–0.94)

0.52 (0.39–0.69)

<0.001

<0.001

Age

1.03 (1.03–1.04)

1.01 (1.00–1.02)

0.003

0.003

Prior ischemic stroke

102.80 (86.43–122.27)

90.59 (75.84–108.20)

<0.001

<0.001

Chronic lung disease

0.86 (0.69–1.07)

0.78 (0.55–1.10)

0.151

0.147

Hypertension

1.90 (1.70–2.13)

1.19 (0.99–1.43)

0.058

0.058

Diabetes Mellitus

1.85 (1.63–2.11)

1.14 (0.87–1.49)

0.358

0.360

  1. AAD antiarrhythmic drugs, APA antiplatelet agents, CI confidence interval, LR Likelihood Ratio, OHA oral hypoglycemic agents.