Skip to main content
. 2018 Jul 16;37(25):3661–3678. doi: 10.1002/sim.7842

Table 1.

Percentage of simulations in which the null hypotheses of proportional hazards for X 1 and X 2 were rejected using joint Wald tests35, 36 for binary and continuous covariates. For a given rejection percentage π, the Monte Carlo SE is π(100π)500×1100

Scenario 1 Scenario 2 Scenario 3 Scenario 4 Scenario 5
X 1 X 2 X 1 X 2 X 1 X 2 X 1 X 2 X 1 X 2
Binary X1,X2
Complete‐data 3 3 89 3 33 3 56 6 45 4
Complete‐case 2 3 42 3 14 2 17 2 14 3
MI‐Approx 0 0 21 0 3 0 4 0 2 0
MI‐SMC 0 0 17 0 2 0 3 0 1 0
MI‐TVE‐Approx 2 3 67 3 16 2 27 2 21 2
MI‐TVE‐SMC 3 4 68 6 24 6 34 5 27 5
Continuous X1,X2
Complete‐data 7 5 100 3 78 5 99 3 96 4
Complete‐case 5 6 94 5 42 4 78 4 60 5
MI‐Approx 0 0 70 0 7 0 40 0 28 0
MI‐SMC 0 0 82 0 9 0 48 0 30 0
MI‐TVE‐Approx 4 5 91 2 23 2 83 3 70 3
MI‐TVE‐SMC 5 7 99 4 53 5 88 3 75 5

Abbreviations: MI, multiple imputation; SMC, substantive‐model‐compatible; TVE, time‐varying effect.