Dear all,
We are working on time-varying effects in survival data using the Cox model. Together with Patrick Royston (London), we have developed an approach based on fractional polynomials (FPs) to derive a Cox model including time-varying covariate effects (i.e. time-varying betas, beta(t), for time-fixed covariates). In this extended model, the beta(t)'s are modelled by using FP functions of time, t. However, getting predicted survival probabilities at the individual level is problematic with such models, due to the need to integrate the hazard function and the fact that analytic solutions to such integrals are not in general available. Therefore, numerical integration is presumably needed. Is anyone else working on this topic? Does anybody have any procedures to compute such integrals in the time-varying Cox model?
Thanks for your help.
Willi Sauerbrei and Anika Buchholz
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
