Statistical issues in survival analysis (Part XVVVVVVI) January 1, 2024 TABLE 11. Sensitivity analyses for the estimation of ATE based on NHEFS data. IPW AIPW MIS Method EST SE p-value EST SE Naive
0.25
0.20
0.15 0.10
0.05
p
LR
0.058
0.956
0.951
0.058
0.956
0.952
rf
2.915
0.811
0.000
3.084
0.807
0.000
LR
0.030
1.193
0.980
0.044
1.185
0.970
rf
1.963
0.973
0.044
1.985
0.956
0.034
LR
0.050
1.182
0.966
0.049
1.156
0.966
rf
2.126
0.967
0.028
2.155
0.948
0.023
LR
0.053
1.167
0.964
0.053
1.148
0.956
rf
2.186
0.952
0.022
2.194
0.927
0.018
LR
0.055
1.112
0.961
0.056
1.110
0.960
rf
2.348
0.941
0.013
2.782
0.906
0.002
LR
0.069
1.008
0.945
0.066
1.004
0.948
rf
2.951
0.935
0.002
3.280
0.887
0.000
The authors were interested in the average treatment effect (ATE) which reflects how the treatment affects the potential outcome. In order to estimate ATE, propensity scores have been adapted for their estimation. such as the inverse probability weighted (IPW) or augmented IPW (AIPW) estimation methods. As they have pointed out, the key idea of the propensity score is based on the conditional probability for an individual to receive a treatment, given pre-treatment confounders. Since this estimation required precise measurements of variables but measurement error often exists, this problem would affect the ATE. In addition, in estimating the propensity score, usually parametric models are used but the authors wanted to use a model with an unknown link function, and they used machine learning methods to estimate that. The authors therefore discussed the estimation of propensity methods non-parametrically and ways to handle the measurement error in order to correct for the error. They focused on using random forests for the non-parametric estimation. In their causal inference methods section, they discussed these issues. In the framework of causal inference, the following assumptions are required when estimating ATE: (A1) Strong ignorable treatment assumption (SITA), (A2) Stable unit treatment value assumption (SUTVA), and (A3) The positivity assumption. The A1 assumption suggested that the confounder is