Skip to main content

Table 3 Multivariate analysis for the complete case, multiple imputation and sensitivity analysis, with M= 1000 imputed data sets

From: Practical considerations for sensitivity analysis after multiple imputation applied to epidemiological studies with incomplete data

  

Complete Case (CC)

Multiple Imputation (MI)

Sensitivity Analysis (SA)

 

Missing values %

aOR 95% CI

SE

CV (%)

aOR 95% CI

SE

CV (%)

VRMI(MI vs CC)(%)

δ

aOR*95% CI

SE

CV (%)

VRSA(SA vs MI)(%)

Alcohol consumption

11.1

2.32

0.39

17

2.82

0.37

13

21.86

-0.40

2.86

0.37

13

1.29

  

[1.66,3.23]

  

[2.18,3.66]

    

[2.21,3.70]

   

Genotype3

26.3

1.51

0.24

16

1.66

0.23

14

9.70

0.15

1.60

0.21

13

3.56

  

[1.10,2.07]

  

[1.27,2.16]

    

[1.23,2.06]

   

HIV

16.3

1.56

0.41

27

1.80

0.34

19

15.52

0.70

1.91

0.36

19

6.12

serostatus

 

[0.92,2.62]

  

[1.24,2.61]

    

[1.32,2.76]

   
  1. Note. aOR, odds ratio adjusted on sex, age, duration of HCV infection, alcohol consumption and HIV serostatus; aOR*, odds ratio obtained from the MI adjusted odds ratio estimates.
  2. CI, confidence interval; CV, coefficient of variation of the aOR; VR MI , variation rate of the aOR for CC and MI analyses; VR SA , variation rate of the aOR for MI and senstitivity analyses.
  3. Covariates included in the model were: sex, age, duration of HCV infection, alcohol consumption, genotype 3 and HIV serostatus. Sensitivity analysis: the weighting process was applied to alcohol consumption, genotype 3 and HIV serostatus indepedently.