Skip to main content

Table 5 Median NRI values according to different aspects of design and analysis

From: A methodology review on the incremental prognostic value of computed tomography biomarkers in addition to Framingham risk score in predicting cardiovascular disease: the use of association, discrimination and reclassification

 

No.

NRI (median)

IQR

P value

1. Reporting of censor handling

 Yes

13

0.18

0.14–0.43

 

 No

33

0.26

0.19–0.35

0.4869

2. No extrapolation

 Yes

20

0.28

0.14–0.49

 

 No

26

0.23

0.16–0.34

0.4186

3. Category NRI referencea

 Quoted

38

0.25

0.18–0.39

 

 Not quoted

8

0.18

0.00–0.38

0.1922

4. Justification of NRI categoriesa

 Yes

40

0.25

0.15–0.41

 

 No

6

0.21

0.18–0.29

0.7691

5. Reporting NRI components

 Yes

22

0.24

0.13–0.34

 

 No

24

0.28

0.19–0.48

0.1907

6. Reclassification table showing the number of events and non-events

 Yes

32

0.27

0.18–0.37

 

 No

14

0.21

0.14–0.43

0.3396

7. The availability of reclassification table with sufficient information to enable the calculation of event and non-event NRI

 Yes

30

0.25

0.18–0.35

 

 No

16

0.23

0.14–0.48

0.9265

8. Describing the combined NRI as sum not a percentage

 Yes

25

0.28

0.14–0.47

 

 No

21

0.22

0.18–0.30

0.256

9. Provide any indication of the proportion of correctly reclassified

 Yes

15

0.35

0.14–0.53

 

 No

31

0.23

0.16–0.34

0.0916

10. Strong conclusion based on the reporting of NRI

 No

20

0.29

0.19–0.48

No

 Yes

26

0.23

0.13–0.34

0.08

11. NRI adequate documented

 Adequate

39

0.25

0.14–0.39

 

 Inadequate

7

0.24

0.20–0.49

0.7949

12. Types of incremental value

 > 3

18

0.28

0.2–0.44

 

 < 3

28

0.23

0.13–0.37

0.2464

  1. Abbreviations: IQR interquartile range, NRI net reclassification index
  2. P values generated using Wilcoxon ranksum test
  3. aOne group has less then 10 studies and therefore the comparison is limited