Skip to main content

Table 2 Comparison of simulated responder analysis results 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

17.6

6.9

10.6 (1.7, 19.5)

−29.2

81.3

0.64

DTI MI

26.5

10.7

15.9 (5.4, 26.4)

6.0

95.2

0.77

DTI MI with CV

24.5

9.7

14.8 (4.6, 25.0)

−1.3

94.9

0.74

IBD MI

25.7

10.8

14.9 (4.5, 25.3)

−0.6

95.2

0.70

IBD MI with CV

24.1

9.9

14.1 (4.0, 24.3)

−5.7

94.3

0.69

2: Differing mechanism

NRI

17.6

7.9

9.6 (0.3, 18.7)

− 35.7

77.2

0.55

DTI MI

26.5

10.5

16.0 (5.5, 26.5)

6.7

94.8

0.77

DTI MI with CV

24.7

9.8

14.8 (4.7, 25.0)

−1

94.3

0.74

IBD MI

25.7

10.8

14.9 (4.5, 25.3)

−0.7

94.9

0.69

IBD MI with CV

24.2

10.1

14.1 (3.9, 24.3)

−5.8

94.5

0.68

3: Differing mechanism, reversed

NRI

18.3

6.9

11.4 (2.4, 20.4)

−24.1

86.3

0.69

DTI MI

26.1

10.5

15.5 (5.1, 26.0)

3.7

93.4

0.74

DTI MI with CV

24.2

9.7

14.5 (4.4, 24.6)

−3.4

93.4

0.72

IBD MI

25.8

10.8

15.0 (4.6, 25.5)

0.2

94.1

0.70

IBD MI with CV

24.2

10.0

14.2 (4.0, 24.4)

−5.4

93.4

0.68

4: Differential dropout rates

NRI

21.5

5.3

16.2 (7.2, 25.3)

8.5

93.8

0.94

DTI MI

26.0

10.8

15.2 (4.8, 25.7)

1.8

93.8

0.71

DTI MI with CV

24.8

9.1

15.6 (5.6, 25.7)

4.5

93.3

0.79

IBD MI

25.6

10.9

14.7 (4.3, 25.1)

−1.8

94.5

0.69

IBD MI with CV

24.6

9.4

15.2 (5.1, 25.3)

1.7

94.8

0.77

5: Lack of efficacy, sensitivity of mechanism

NRI

16.5

6.7

9.7 (1.1, 18.4)

−35

75.4

0.59

DTI MI

26.6

10.5

16.1 (5.6, 26.5)

7.1

93.7

0.76

DTI MI with CV

24.4

9.6

14.8 (4.6, 24.9)

−1.5

94.3

0.72

IBD MI

25.8

10.8

15.0 (4.6, 25.5)

0.4

94.3

0.68

IBD MI with CV

24.0

9.9

14.2 (4.0, 24.3)

−5.5

93.6

0.67

6: Lack of efficacy and tolerability

NRI

18.0

7.1

10.9 (2.0, 19.9)

−27.1

83.8

0.67

DTI MI

26.3

10.6

15.7 (5.2, 26.2)

4.7

93.7

0.77

DTI MI with CV

24.4

9.7

14.7 (4.6, 24.9)

−1.9

93.8

0.75

DTI MI with AE

26.5

11.0

15.5 (5.0, 26.0)

3.4

93.3

0.74

IBD MI

25.6

10.8

14.8 (4.4, 25.2)

−1.2

93.0

0.69

IBD MI with CV

24.1

10.0

14.2 (4.0, 24.3)

−5.5

93.8

0.69

IBD MI with AE

25.7

10.9

14.8 (4.3, 25.2)

−1.5

93.2

0.69

  1. NRI: Non-response imputation; DTI MI: Dichotomize then impute multiple imputation; IBD MI: Impute before dichotomizing multiple imputation; CV: Correlated variable; AE: Adverse event status
  2. 1 Results are from a linear response profile with 30% 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