Skip to main content

Table 2 Comparison of the results obtained with the non-parametric estimator and the regression model on pseudo values. Sixteen pseudo values are used in each setting. QIC was used to select the degrees of freedom of the splines (from a minimum of 4 to a maximum of 12). The different scenarios are described in Fig. 1. Two different sample size are considered (200 and 400 per group with 20% censoring)

From: Analyzing differences between restricted mean survival time curves using pseudo-values

  

Non Parametric

Pseudo-Values

 

Scenario

Bias

Coverage

Length

Bias

Coverage

Length

200

1

0.095

0.941

0.643

0.096

0.941

0.612

 

2

0.719

0.931

4.297

0.717

0.943

4.425

 

3

0.702

0.937

4.634

0.704

0.952

4.524

 

4

0.172

0.928

1.140

0.172

0.938

1.115

 

5

1.434

0.973

9.886

1.446

0.954

9.499

400

1

0.069

0.946

0.464

0.069

0.928

0.437

 

2

0.502

0.939

3.138

0.503

0.937

3.225

 

3

0.502

0.948

3.539

0.506

0.929

3.398

 

4

0.125

0.944

0.866

0.125

0.945

0.839

 

5

1.033

0.965

7.143

1.038

0.942

6.931