running A. null after the window, strong association before it ...
running B. as A, but a real post-window effect ...
running C. null after the window, weak association before it ...
running D. null after the window, sampling depends on severity ...
running E. real effect, sparse creatinine widens the interval ...
running F. real effect, outcome definition does not change ...

==============================================================================================================
SIMULATION RESULTS
==============================================================================================================
                                              scenario   method  target_log_hr  replicates  mean_estimate    bias  empirical_se  coverage_pct  rejection_rate_pct  median_concurrent_fraction_pct  post_landmark_events_with_earlier_onset_pct
A. null after the window, strong association before it    naive         0.0008         500         0.4160  0.4153        0.0496           0.0               100.0                            33.9                                         37.8
A. null after the window, strong association before it landmark         0.0008         500         0.2357  0.2349        0.0404           0.0               100.0                            33.9                                         37.8
A. null after the window, strong association before it interval         0.0008         500         0.2749  0.2741        0.0396           0.0               100.0                            33.9                                         37.8
                B. as A, but a real post-window effect    naive         0.3214         500         0.5826  0.2612        0.0525           0.0               100.0                            34.9                                         39.8
                B. as A, but a real post-window effect landmark         0.3214         500         0.3879  0.0665        0.0401          56.8               100.0                            34.9                                         39.8
                B. as A, but a real post-window effect interval         0.3214         500         0.3780  0.0566        0.0419          73.8               100.0                            34.9                                         39.8
  C. null after the window, weak association before it    naive        -0.0029         500         0.1465  0.1494        0.0471          10.4                88.6                            30.4                                         35.9
  C. null after the window, weak association before it landmark        -0.0029         500         0.0849  0.0878        0.0363          30.8                66.4                            30.4                                         35.9
  C. null after the window, weak association before it interval        -0.0029         500         0.0939  0.0968        0.0348          22.0                75.0                            30.4                                         35.9
D. null after the window, sampling depends on severity    naive         0.0044         500         0.5575  0.5531        0.0515           0.0               100.0                            37.4                                         36.4
D. null after the window, sampling depends on severity landmark         0.0044         500         0.2987  0.2943        0.0423           0.0               100.0                            37.4                                         36.4
D. null after the window, sampling depends on severity interval         0.0044         500         0.1952  0.1908        0.0387           0.0               100.0                            37.4                                         36.4
 E. real effect, sparse creatinine widens the interval    naive         0.3199         500         0.4288  0.1089        0.0491          37.8               100.0                            21.5                                         50.1
 E. real effect, sparse creatinine widens the interval landmark         0.3199         500         0.3078 -0.0121        0.0382          93.6               100.0                            21.5                                         50.1
 E. real effect, sparse creatinine widens the interval interval         0.3199         500         0.3848  0.0648        0.0475          78.8               100.0                            21.5                                         50.1
    F. real effect, outcome definition does not change    naive         0.3213         500         0.5329  0.2115        0.0563           2.6               100.0                            28.3                                         32.4
    F. real effect, outcome definition does not change landmark         0.3213         500         0.4412  0.1199        0.0344           5.0               100.0                            28.3                                         32.4
    F. real effect, outcome definition does not change interval         0.3213         500         0.3219  0.0005        0.0317          97.0               100.0                            28.3                                         32.4

written to results\manuscript_v5\af_simulation.csv
