Skip to main content

Table 4 Reclassification in predicted risk of methicillin-resistant Staphylococcus aureus carriage between models

From: Variable performance of models for predicting methicillin-resistant Staphylococcus aureus carriage in European surgical wards

Stepwise model

Best BMA model

BMA model

Simple model

Frequency (Row %)

<4%

≥4%

Total

<4%

≥4%

Total

<4%

≥4%

Total

Patients who screened MRSA positive

<4%

13 (68.4)

6 (31.6)

19

16 (84.2)

3 (15.8)

19

16 (84.2)

3 (15.8)

19

≥4%

2 (6.7)

28 (93.3)

30

2 (6.7)

28 (93.3)

30

0 (0)

30 (100)

30

Total

15

34

49

18

31

49

16

33

49

Patients who screened MRSA negative

<4%

852 (89.7)

98 (10.3)

950

917 (96.5)

33 (3.5)

950

915 (96.3)

35 (3.7)

950

≥4%

56 (12.4)

396 (87.6)

452

53 (11.7)

399 (88.3)

452

41 (9.1)

411 (90.9)

452

Total

908

494

1,402

970

432

1,402

956

446

1,402

NRI (95% CI)

5.2% (−6.3%-16.6%) P = .376

5.6% (−3.4%-14.6%) P = .222

6.6% (−0.5%-13.6%) P = .068

  1. Note. The table shows the results when a random sample of 50% of the full cohort was used as the derivation dataset with the remaining data used as the validation dataset. CI, confidence interval; MRSA, methicillin-resistant Staphylococcus aureus; NRI, net reclassification improvement of each model compared with the Stepwise model.