Skip to main content
. Author manuscript; available in PMC: 2019 Jun 12.
Published in final edited form as: Stat Methods Med Res. 2018 May 2;28(6):1676–1688. doi: 10.1177/0962280218772592

Table 5.

Monte Carlo Simulation study: estimation of Cox regression with dependent censoring, where T Exponential T~ExponentialeIn2×-In2Z, CExponentiale-2X+0.IZ, XNormal0.I5+Z,I and δXBernoullipZ,δt,Y=II+eI.5+0.5Z-2Y.

βx = In (2) = 0.693
βZ = −ln(2) = −0.693
Method Esta SDb SEc RMSEd CRe Est SD SE RMSE CR Divf
N = 200
 FO 0.668 0.1 13 0.1 14 0.1 16 94.4 -0.690 0.318 0.324 0.318 95.8
 CC 0.700 0.168 0.170 0.168 96.1 -0.869 0.446 0.447 0.479 94.4
 PMM1 0.592 0.123 0.143 0.159 92.5 -0.621 0.329 0.346 0.337 95.2
 AIPW11 0.666 0.160 0.151 0.162 93.4 -0.687 0.346 0.373 0.346 96.5 1
 NNMI11 ,(0.8,0.2) 0.656 0.150 0.150 0.155 94.3 -0.670 0.337 0.350 0.338 95.6
 NNMI11(0.2,0.8) 0.655 0.154 0.153 0.159 94.5 -0.643 0.334 0.349 0.338 95.4
 AIPW12 0.902 0.322 0.097 0.384 60.3 -0.814 0.575 0.517 0.588 89.8 1765
 NNMI12(0.8,0.2) 0.616 0.127 0.142 0.149 92.9 -0.645 0.329 0.347 0.333 95.9
 NNMI12(0.2,0.8) 0.594 0.1 16 0.138 0.153 91.8 -0.628 0.324 0.344 0.330 95.4
 aipw21 0.667 0.162 0.151 0.164 93.4 -0.688 0.347 0.373 0.347 96.5 1
 NNMI21(0.8,0.2) 0.656 0.148 0.149 0.153 94.2 -0.667 0.337 0.349 0.338 95.7
 NNMI21(0.2,0.8) 0.655 0.154 0.154 0.159 94.1 -0.641 0.334 0.348 0.338 95.4
N = 400
 FO 0.677 0.080 0.080 0.082 94.1 -0.695 0.224 0.227 0.224 94.9
 CC 0.716 0.120 0.1 18 0.122 95.6 -0.870 0.307 0.309 0.354 91.6
 PMM1 0.602 0.088 0.099 0.127 85.5 -0.639 0.234 0.243 0.240 95.4
 AIPW11 0.676 0.106 0.106 0.107 94.6 -0.695 0.241 0.254 0.241 95.6 0
 NNMI11 (0.8,0.2) 0.673 0.105 0.106 0.107 94.7 -0.686 0.240 0.244 0.240 95.4
 NNMI11(0.2,0.8) 0.673 0.108 0.107 0.110 95.1 -0.667 0.240 0.243 0.241 95.6
 AIPW1 2 0.926 0.224 0.069 0.323 44.0 -0.829 0.371 0.424 0.395 88.7 1596
 NNMI12(0.8,0.2) 0.631 0.093 0.101 0.112 90.8 -0.663 0.233 0.243 0.235 96.3
 NNMI12(0.2,0.8) 0.609 0.084 0.098 0.119 87.2 -0.643 0.228 0.242 0.233 95.6
 aipw21 0.676 0.107 0.106 0.108 94.6 -0.695 0.241 0.254 0.241 95.6 0
 NNMI21(0.8,0.2) 0.672 0.105 0.105 0.107 94.9 -0.685 0.240 0.244 0.240 95.7
 NNMI21(0.2,0.8) 0.672 0.107 0.107 0.109 94.9 -0.667 0.239 0.243 0.240 95.3

Note: Censoring rate: 0.33; Missing rate: 0.47.

a

Average of 1000 point estimates.

b

Empirical standard deviation.

c

Average estimated standard error.

d

Root mean square error: square root of bias2+SD2

e

Coverage rate of 1000 95% confidence intervals.

f

Number of disconvergences for AIPW