Mixed Effects Cox Models

Like many other projects, this software library started with a data set and a problem. From this came statistical ideas for a solution, followed by some initial programming — which more than anything else helped to define the real computational and statistical issues — and then a more ambitious programming solution. The problem turned out to be harder than I thought; the first release-worthy code took over 3 years in gestation and resulted in the kinship library. This in turn led to application to a larger set of problems, and a complete re-design of the underlying code. It was then split into separate libraries: coxme contains the central code for mixed effects Cox and linear models, kinship2 contains code for the construction and manipulation of pedigree data, and bdsmatrix contains underlying routines for block-diagonal sparse matrices.

[1]  T M Therneau,et al.  Fifty-year follow-up of cancer incidence in a historical cohort of Minnesota breast cancer families. , 1999, Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology.

[2]  J. Palmgren,et al.  Estimation of Multivariate Frailty Models Using Penalized Partial Likelihood , 2000, Biometrics.

[3]  I. S. Goldenberg Variables Related to Human Breast Cancer , 1958, The Yale Journal of Biology and Medicine.

[4]  D. Grabrick,et al.  Evaluation of familial clustering of breast and prostate cancer in the Minnesota Breast Cancer Family Study. , 2003, Cancer detection and prevention.

[5]  R. Gray Modeling Survival Data: Extending the Cox Model , 2002 .