Skip to main content

Table 4 Simulation results for the estimation of attributable risk A(.) under nonproportional hazards with regression parameter β= ln(2) and probability of exposure q=0.5

From: Comparison of methods for estimating the attributable risk in the context of survival analysis

Estimation method

  

n=1, 000

n=10, 000

 

Time

A(t)

Bias

SEE

SSD

CP

Bias

SEE

SSD

CP

KM

τ/4

0.181

0.001124

0.045053

0.045787

0.954

0.000289

0.014277

0.014126

0.949

 

τ/2

0.133

0.001330

0.037581

0.037647

0.953

−0.000029

0.011915

0.012154

0.935

 

3 τ/4

0.109

0.001211

0.036543

0.036593

0.953

−0.000301

0.011618

0.011608

0.952

 

τ

0.093

0.002743

0.043713

0.051764

0.933

−0.000888

0.016362

0.019957

0.950

WKM

τ/4

0.181

0.001138

0.045090

0.045739

0.954

0.000291

0.014274

0.014130

0.949

 

τ/2

0.133

0.001347

0.037587

0.037593

0.956

−0.000024

0.011911

0.012151

0.938

 

3 τ/4

0.109

0.001165

0.036511

0.036518

0.952

−0.000293

0.011612

0.011607

0.956

 

τ

0.093

0.001685

0.042617

0.049261

0.920

−0.000708

0.016157

0.019107

0.946

COX

τ/4

0.181

−0.018761

0.037521

0.037543

0.933

−0.019843

0.011869

0.011939

0.621

 

τ/2

0.133

0.010548

0.033500

0.033580

0.941

0.009504

0.010588

0.010676

0.847

 

3 τ/4

0.109

0.023376

0.030960

0.031017

0.879

0.022314

0.009775

0.009879

0.368

 

τ

0.093

0.030360

0.029427

0.029588

0.830

0.029168

0.009323

0.009456

0.127

PCH

τ/4

0.181

0.026479

0.048525

0.049191

0.908

−0.017516

0.011688

0.012080

0.672

 

τ/2

0.133

0.057418

0.044915

0.045594

0.738

0.011082

0.010391

0.010768

0.806

 

3 τ/4

0.109

0.070045

0.042342

0.043042

0.607

0.023478

0.009571

0.009936

0.313

 

τ

0.093

0.075924

0.040403

0.041050

0.525

0.029848

0.009011

0.009360

0.098

  1. KM nonparametric approach based on Kaplan-Meier estimation for S(t), WKM nonparametric approach based on weighted Kaplan-Meier estimation for S(t), COX semiparametric approach, PCH parametric approach using a piecewise constant hazards model, Bias sampling mean of the difference between \(\hat {A}(t)\) and A(t), SEE sampling mean of standard error estimate of A(t), SSD sampling standard deviation of \(\hat {A}(t)\), CP coverage probability of the 95% Wald confidence interval