Weibull regression in censored survival analysis;
A comparison with BUGS



ADMB Files
Code: kidney.tpl
Data: kidney.dat
Initial values: kidney.pin
All required files (DOS): kidney.zip
All required files (linux): kidney.tar.gz
Results: kidney.par

Running ADMB-executables
In a DOS window
Under linux

Results: Computation times
ADMB-RE: 9 seconds.
BUGS: 7 seconds,
which was the time taken by WinBUGS 1.4 to perform 1,000 iterations of two parallel chains.

Navigation
ADMB-RE home
Otter Research

Model description

The collection of examples following WinBUGS (Volume 1, WinBUGS 1.4) contains an application of Cox's proportional hazard model with random effects. A Weibull baseline hazard is used:

h0(t) = r*tr-1, r>0.

The "kidney data" contain information on time-to-infection for 38 patients, and there are several explanatory variables (see table below). A full description of the model can be found here: kidney.pdf.

Comparison with BUGS

Parameter estimates for the kidney data are shown in the following table.
b0 bage b1 b2 b3 bsex r s
ADMB-RE -4.344 0.003 0.1208 0.6058 -1.1423 -1.8767 1.1624 0.5617
Std. dev. 0.872 0.0137 0.5008 0.5011 0.7729 0.4754 0.1626 0.297
WinBUGS -4.6 0.003 0.1329 0.6444 -1.168 -1.938 1.215 0.6374
Std. dev. 0.8962 0.0148 0.5393 0.5301 0.8335 0.4854 0.1623 0.357