Regression Analysis for Outcome-Dependent Sampling Design under the Covariate-Adjusted Additive Hazards Model
Table 2
Simulation results when .
Cutpoints
Method
Bias
SD
SE
CP
Bias
SD
SE
CP
(0.3, 0.7)
Oracle
0.0774
0.0761
0.935
0.0071
0.0560
0.0544
0.945
Naive
0.0770
0.0759
0.943
0.1131
0.0391
0.0369
0.155
Proposed
0.0774
0.0760
0.932
0.0035
0.0591
0.0571
0.941
(0.2, 0.8)
Oracle
0.0670
0.0645
0.931
0.0059
0.0480
0.0476
0.945
Naive
0.0666
0.0645
0.928
0.1123
0.0344
0.0319
0.090
Proposed
0.0670
0.0645
0.932
0.0028
0.0503
0.0499
0.947
(0.3, 0.7)
Oracle
0.0694
0.0706
0.958
0.0110
0.0524
0.0527
0.950
Naive
0.0696
0.0705
0.958
0.1142
0.0352
0.0357
0.118
Proposed
0.0692
0.0706
0.961
0.0084
0.0550
0.0550
0.955
(0.2, 0.8)
Oracle
0.0617
0.0640
0.962
0.0106
0.0478
0.0480
0.958
Naive
0.0621
0.0640
0.959
0.1156
0.0339
0.0323
0.067
Proposed
0.0618
0.0640
0.959
0.0074
0.0502
0.0501
0.958
(0.3, 0.7)
Oracle
0.0862
0.0851
0.945
0.0169
0.0664
0.0655
0.951
Naive
0.0864
0.0851
0.940
0.1177
0.0461
0.0443
0.224
Proposed
0.0863
0.0851
0.946
0.0146
0.0695
0.0683
0.949
(0.2, 0.8)
Oracle
0.0843
0.0837
0.935
0.0180
0.0663
0.0636
0.934
Naive
0.0844
0.0837
0.939
0.1168
0.0453
0.0434
0.213
Proposed
0.0842
0.0837
0.936
0.0160
0.0689
0.0664
0.935
“Oracle” denotes the oracle estimator calculated based on the true covariate. “Naive” denotes the naive estimator calculated based on the unadjusted covariate. “Proposed” denotes the proposed estimator calculated based on the adjusted covariate.