I am planning a longitudinal follow-up of a human population. In the
past when I needed a sample size estimate for a survival study I have
used a parametric approach to the sample size estimation, i.e. I used an
exponential survival function to estimate the sample size. I recognize
that the exponential function is not a good estimate for human survival
in that the exponential distribution is memory less. In order to obtain
a more realistic sample size estimate, I would like to simulate the
survival distribution. It is know that mortality doubles approximately
every eight years in adult human populations and that mortality is
approximately 1% at age 50. Unfortunately I have no Idea how to set up
the simulation. I hope some kind soul will be able, and willing to help.
Thanks
John Sorkin
John Sorkin MD, PhD
Chief, Biostatistics and Informatics
Baltimore VA Medical Center GRECC and
University of Maryland School of Medicine Claude Pepper OAIC
410-605-7119
john@grecc.umaryland.edu
|