The general objective of this work will be to develop tested and documented software and corresponding expository material to facilitate the application of new techniques for the analysis of survival data with random effects (""""""""frailties"""""""") to clinical and epidemiologic studies of cancer. Specific products to be developed include: (a)A package for generating simulated data from random effects models, including models with gamma,inverse Gaussian, k-point, and stable frailty distributions. (b) Programs (written in C and in some cases interfacing with """"""""S-plus' """"""""BMDP"""""""" or """"""""mathematica"""""""") for fitting random effects models to survival data in parametric and semiparametric models and providing appropriate diagnostics to assess goodness-of-fit. (c) Tables and nomograms for assessing the influence of unobserved random effects (including that due to """"""""errors in variables"""""""") on power calculations for epidemiologic studies. (d) Expository material in the form of a monograph or lecture notes explaining the use of the techniques, oriented toward a clinical and epidemiologic audience. The products will be tested on data from a large cohort study of lung cancer mortality among asbestos workers, and other sets of survival data in the cancer literature.
| Oakes, D; Jeong, J H (1998) Frailty models and rank tests. Lifetime Data Anal 4:209-28 |
