Skip to main content

Table 6 Performance for scenario V with true intervention effect of \({\beta }_{t}=-0.264\)

From: Comparison of statistical models for estimating intervention effects based on time-to-recurrent-event in stepped wedge cluster randomized trial using open cohort design

Event generation model

    

AG model

PWP-TT model

PWP-GT model

E

Bias

MSE

CP

Bias

MSE

CP

Bias

MSE

CP

Poisson

1.5

0.0265

0.0040

0.937

0.0014

0.0041

0.937

0.0498

0.0064

0.848

2

0.0232

0.0045

0.943

0.0015

0.0046

0.938

0.0603

0.0084

0.842

4

0.0110

0.0088

0.944

-0.0012

0.0095

0.938

0.0786

0.0152

0.856

6

0.0066

0.0119

0.954

-0.0021

0.0128

0.948

0.0878

0.0212

0.870

Mixed-Poisson

1.5

0.0289

0.0039

0.930

0.0098

0.0036

0.948

0.0660

0.0079

0.796

2

0.0246

0.0045

0.944

0.0081

0.0045

0.945

0.0740

0.0096

0.809

4

0.0144

0.0086

0.939

0.0050

0.0092

0.928

0.0869

0.0169

0.840

6

0.0082

0.0119

0.947

0.0016

0.0124

0.945

0.0944

0.0218

0.863

Weibull (constant)

1.5

0.0721

0.0084

0.768

0.0689

0.0080

0.793

0.0002

0.0034

0.957

2

0.0876

0.0115

0.724

0.0855

0.0112

0.734

0.0013

0.0042

0.945

4

0.1148

0.0203

0.713

0.1139

0.0202

0.731

-0.0044

0.0080

0.944

6

0.1274

0.0264

0.757

0.1267

0.0264

0.766

-0.0019

0.0107

0.957

Weibull (change)

1.5

0.1363

0.0201

0.122

0.0497

0.0050

0.834

-0.0005

0.0025

0.937

2

0.1391

0.0214

0.213

0.0656

0.0073

0.780

0.0009

0.0031

0.954

4

0.1437

0.0251

0.521

0.0961

0.0149

0.769

-0.0016

0.0063

0.949

6

0.1447

0.0275

0.629

0.1061

0.0193

0.788

-0.0040

0.0089

0.952

  1. AG Andersen-Gill, PWP-TT Prentice-Williams-Peterson Total-Time, PWP-GT Prentice-Williams-Peterson Gap-Time, MSE Mean square error, CP Coverage probability