Skip to main content

Table 3 Comparison of simulated responder analysis results when 50% responses are missing using non-response imputation, impute-before-dichotomizing and dichotomize-then-impute multiple imputation1

From: Imputation strategies when a continuous outcome is to be dichotomized for responder analysis: a simulation study

Dropout model

Imputation method

% Responders Trt A

% Responders Trt B

Difference in proportions (95% CI)

% Bias

Coverage of the 95% CI

Power

1: Lack of efficacy

NRI

12.8

4.8

8.0 (0.3, 15.7)

− 46.8

55.6

0.52

DTI MI

27.5

11.2

16.3 (5.7, 26.9)

8.8

91.5

0.72

DTI MI with CV

24.2

9.7

14.5 (4.4, 24.6)

−3.2

90.6

0.66

IBD MI

25.8

11.1

14.8 (4.3, 25.2)

−1.5

94.1

0.59

IBD MI with CV

23.3

9.7

13.6 (3.5, 23.6)

−9.6

92.9

0.56

2: Differing mechanism

NRI

12.8

6.3

6.6 (−1.4, 14.6)

− 56.2

45.5

0.37

DTI MI

27.5

10.9

16.6 (6.1, 27.2)

11.0

86.9

0.71

DTI MI with CV

24.6

9.8

14.8 (4.7, 25.0)

− 1.2

88.7

0.65

IBD MI

25.9

11.1

14.8 (4.4, 25.3)

−1.1

92.9

0.58

IBD MI with CV

23.6

9.9

13.7 (3.6, 23.8)

−8.6

92.2

0.56

3: Differing mechanism, reversed

NRI

13.9

4.8

9.0 (1.1, 16.9)

− 39.8

66.5

0.62

DTI MI

26.7

11.0

15.7 (5.1, 26.2)

4.5

85.4

0.64

DTI MI with CV

23.9

9.7

14.2 (4.1, 24.2)

−5.4

86.1

0.61

IBD MI

26.1

11.1

15.1 (4.6, 25.6)

0.7

92.0

0.58

IBD MI with CV

23.6

9.8

13.8 (3.7, 23.8)

−8.1

91.1

0.55

4: Differential dropout rates

NRI

18.3

1.8

16.5 (8.5, 24.4)

10.0

93.9

0.99

DTI MI

26.2

14.5

11.7 (0.9, 22.5)

−21.8

77.5

0.48

DTI MI with CV

24.0

11.1

12.9 (2.7, 23.1)

−13.8

84.1

0.58

IBD MI

25.7

11.8

13.9 (3.4, 24.5)

−6.9

92.8

0.49

IBD MI with CV

23.8

9.4

14.4 (4.4, 24.4)

−3.7

94.4

0.60

5: Lack of efficacy, sensitivity of mechanism

NRI

13.7

5.6

8.1 (0.1, 16.1)

− 45.9

59.9

0.53

DTI MI

26.9

10.7

16.2 (5.7, 26.7)

8.1

91.9

0.72

DTI MI with CV

24.2

9.5

14.7 (4.6, 24.8)

−2.1

92.6

0.67

IBD MI

25.9

10.9

15.0 (4.5, 25.4)

−0.2

94.1

0.62

IBD MI with CV

23.5

9.8

13.8 (3.7, 23.8)

−8.1

93.6

0.60

6: Lack of efficacy and tolerability

NRI

13.2

4.9

8.3 (0.5, 16.1)

− 44.6

59.5

0.57

DTI MI

26.9

11.0

15.9 (5.4, 26.5)

6.1

91.3

0.68

DTI MI with CV

24.1

9.7

14.4 (4.3, 24.5)

−4.0

91.2

0.65

DTI MI with AE

27.4

11.9

15.5 (4.8, 26.2)

3.4

90.1

0.63

IBD MI

25.8

11.1

14.7 (4.2, 25.2)

−2.0

93.3

0.59

IBD MI with CV

23.5

9.8

13.6 (3.6, 23.7)

−9.0

92.5

0.57

IBD MI with AE

26.0

11.4

14.6 (4.1, 25.1)

−2.7

93.6

0.57

  1. NRI: Non-response imputation; DTI MI: Dichotomize then impute multiple imputation; IBD MI: Impute before dichotomizing multiple imputation
  2. 1Results are from a linear response profile with 50% data missing at random, N = 200. In fully observed data, % responders in Treatment A and B was 25.6 and 10.6, respectively for a difference of 15.0 and power = 0.80