BIOMETRY RESEARCH GROUP

Mathematica Packages

Discussion of Double Sampling for Survival Analysis

(written by Stuart G. Baker)

These programs are related to the discussion article "Discussion of Double Sampling for Survival Analysis" to appear in Biometrics following the manuscript by Frangakis and Rubin (FR). Most of the software is based on the model in Baker, Wax, and Patterson (1993) (BWP)

The first step is to load composite linear model software into a directory called clm.
Then download the following files.
To reproduce calculations in the manuscript, type

<<double.m
double[sym]

For a hypothetical example with time-varying administrative censoring type

<<doublev.m
comparev[mod1]

Filename	File Notes
double.m	main program for discussion (loads all others)
doublev.m	main program for extension to time-varying administrative censoring
exbwpg.m	generate hypothetical data
exbwp.m	BWP for full data
exbwpr.m	BWP for reduced data
exbwpf.m	BWP for full follow-up only
exbwpfr.m	FR method for simple example
exbwpv.m	BWP for administrative censoring at time 3 for full data
exbwpvr.m	BWP for administrative censoring at time 3 for reduced data

References

Baker SB, Wax Y, Patterson B. Regression analysis of grouped survival data: informative censoring and double sampling. Biometrics, 1993;49:379-389.

Frangakis CE, Rubin DB. A note on addressing an idiosyncrasy in estimating survival curves using double sampling in the presence of self-selected right censoring. Biometrics, 2001, In press