Skip to main content

Table 4 Comparison of longitudinal effect on survival (N = 1000)

From: A comparison of time dependent Cox regression, pooled logistic regression and cross sectional pooling with simulations and an application to the Framingham Heart Study

Scenarios

CSP_UNADJUSTED

CSP_ADJUSTED & TDCM

Event rate

γ

Estimate

SE

CP

Bias

MSE

Estimate

SE

CP

Bias

MSE

90 %

0.000

0.003

0.054

0.957

0.003

0.006

0.003

0.055

0.954

0.003

0.006

0.500

0.498

0.055

0.954

−0.002

0.006

0.498

0.056

0.952

−0.002

0.006

1.000

1.083

0.058

0.720

0.082

0.014

1.002

0.059

0.958

0.002

0.007

50 %

0.000

−0.001

0.075

0.953

−0.001

0.011

−0.002

0.076

0.944

−0.002

0.012

0.500

0.499

0.070

0.947

−0.001

0.010

0.499

0.071

0.944

−0.001

0.010

1.000

1.001

0.073

0.946

0.001

0.011

1.002

0.074

0.948

0.002

0.011

10 %

0.000

0.007

0.168

0.944

0.007

0.058

0.007

0.171

0.938

0.007

0.060

0.500

0.501

0.158

0.947

0.001

0.051

0.501

0.161

0.946

0.001

0.053

1.000

1.067

0.145

0.906

0.067

0.048

1.003

0.147

0.937

0.003

0.045

 

PLR_UNADJUSTED

PLR_ADJUSTED

Event rate

γ

Estimate

SE

CP

Bias

MSE

Estimate

SE

CP

Bias

MSE

90 %

0.000

0.003

0.066

0.957

0.003

0.008

0.003

0.067

0.957

0.003

0.009

0.500

0.599

0.069

0.709

0.099

0.019

0.601

0.070

0.711

0.101

0.020

1.000

1.342

0.080

0.005

0.342

0.130

1.255

0.082

0.111

0.255

0.078

50 %

0.000

−0.001

0.080

0.950

−0.001

0.013

−0.002

0.081

0.945

−0.002

0.013

0.500

0.545

0.077

0.909

0.045

0.014

0.545

0.079

0.912

0.045

0.014

1.000

1.109

0.084

0.746

0.109

0.026

1.109

0.086

0.744

0.109

0.027

10 %

0.000

0.007

0.170

0.945

0.007

0.059

0.007

0.173

0.937

0.007

0.061

0.500

0.510

0.161

0.947

0.010

0.053

0.509

0.164

0.941

0.009

0.055

1.000

1.091

0.150

0.888

0.091

0.055

1.027

0.152

0.926

0.027

0.049

  1. Abbreviations: SE Standard Error, CP 95 % Coverage Probability, MSE Mean Square Error, CSP_UN Unadjusted Cross Sectional Pooling, CSP_AD Adjusted Cross Sectional Pooling; PLR_UN Unadjusted Pooled Logistic Regression, PLR_AD Adjusted Pooled Logistic Regression, TDCM Time Dependent Cox Regression Modeling