[R] comparing SAS and R survival analysis with time-dependent covariates
goran.brostrom at gmail.com
Fri Jul 22 10:38:05 CEST 2011
2011/7/21 Abdullah OUESLATI <aboueslati at gmail.com>:
> <quote author="Göran Broström">
> I do not understand why you expect to get comparable results with SAS
> "discrete" and coxph "exact". They are two different approaches to
> handling ties (as Terry explained; of course, some comparability
> should be expected in "normal" cases).
> I didn't know precisely the specifities of each approximation method.
> I thus came back to section 3.3 of Therneau and Grambsch, Extending the Cox
> Model. I think I now see things more clearly. If I have understood
> correctly, both "discrete" option and "exact" functions assume "true"
> discrete event times in a model approximating the Cox model. Cox partial
> likelihood cannot be exactly maximized, or even written, when there are some
> ties, am I right ?
> In my sample, many of the ties (those whithin a single observation of the
> process) are due to the fact that continuous event times are grouped into
> So I think the logistic approximation may not be the best for my problem
> despite the estimate on my real data set (shown on my previous post) do give
> interessant results regarding to the context of my data set !
I would say that you should use the discrete option ("ml" or "mppl" in
coxreg). Compare that with what you get with Efron's approximation.
You should forget about the 'exact' method with large data sets. Maybe
that's what Terry meant..
> I was thinking about distributing the events uniformly in each interval.
> What do you think about this option ? Can I expect a better approximation
> than directly applying Breslow or Efron method directly with the grouped
> event data ? Finally, it becomes a model problem more than a computationnal
> or algorithmic one I guess.
> <quote author="Terry Therneau-2">
> I never use the discrete option, having found the Efron
> approximation to be sufficient in every practical situation.
> Which criteria would you use to determinate if Efron approximation is good
> or not with grouped event data ? Is it possible to compare appproximation
> models by the results and not by the model assumption ?
> Thank you very much.
More information about the R-help