Skip to main content

Table 4 Derivation and validation of logistic regression models of the fourfold cross-validation process

From: Development of a severity score based on the International Classification of Disease-10 for general patients visiting emergency centers

 

Fold 1

Fold 2

Fold 3

Fold 4

Derivation set size (n)

10,272,021

10,269,926

10,271,579

10,272,824

Validation set size (n)

3,423,429

3,425,524

3,423,871

3,422,626

Regression coefficients

    

 Intercept

-2.593

-2.532

-2.565

-2.541

 Age score

0.307

0.308

0.308

0.308

 MEWSa

0.608

0.606

0.608

0.609

 ICESSb

-6.385

-6.447

-6.413

-6.461

Youden’s threshold

-5.900

-5.874

-5.869

-5.916

AUROCc

0.975 (0.974—0.976)

0.974 (0.973—0.976)

0.975 (0.973—0.976)

0.973 (0.972—0.975)

AUPRCd

0.741 (0.739—0.753)

0.738 (0.734—0.749)

0.746 (0.742—0.757)

0.735 (0.732—0.746)

True positive (n)

11,351

11,237

11,294

11,551

False positive (n)

201,366

192,619

192,770

200,229

True negative (n)

3,209,396

3,220,223

3,218,434

3,209,427

False negative (n)

1316

1445

1373

1419

Accuracy

0.941 (0.940—0.944)

0.943 (0.940—0.945)

0.943 (0.941—0.944)

0.941 (0.940—0.943)

Sensitivity

0.896 (0.890—0.090)

0.886059 (0.882—0.893)

0.8916081 (0.887—0.899)

0.8905937 (0.885—0.896)

Specificity

0.942 (0.941—0.944)

0.944 (0.941—0.945)

0.943 (0.941—0.944)

0.941 (0.940–0.944)

  1. The numbers in parentheses represent 95% confidence intervals
  2. aMEWS: Medical early warning score; bICESS: International Classification of Diseases-based emergency severity score; cAUROC: Area under the receiver operating characteristic curve; dAUPRC: Area under the precision‒recall curve