Add vignette to document the different behaviors of coxph variants

[nanxstats]

Thanks @larry-leon for discovering the intriguing difference and kindly agree to contribute the vignette about ties.

[LittleBeannie]

Hi @nanxstats and @larry-leon, do you happen to have a timeline to get the vignette published?

[nanxstats]

Hi @nanxstats and @larry-leon, do you happen to have a timeline to get the vignette published?

Larry might have an estimate.

[larry-leon]

Sorry, I forgot about this. Can I get a draft by end of next week? Also, in simtrial, I assume there is not a way to set “timefix” to FALSE for coxph ? If not, I will just mention the issue, and discuss how the aeqSurv() function can be applied directly  Thanks,Larry Sent from Yahoo Mail for iPhone On Thursday, October 24, 2024, 11:59 AM, Nan Xiao ***@***.***> wrote: Hi @nanxstats and @larry-leon, do you happen to have a timeline to get the vignette published? Larry might have an estimate. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

[LittleBeannie]

Hi @larry-leon , next week works perfectly. Could you please provide more information of what is "timefix"?

[larry-leon]

In coxph.control, timefix=TRUE is default which implements aeqSurv to “resolve any near ties”. So in order to match how simtrial implements the logrank test, we would need to set timefix=FALSE However, if the main goal of the vignette is just to raise awareness, then we do not need to revise the coxph call. Sent from Yahoo Mail for iPhone On Thursday, October 24, 2024, 1:34 PM, Yujie Zhao ***@***.***> wrote: Hi @larry-leon , next week works perfectly. Could you please provide more information of what is "timefix"? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

[LittleBeannie]

Thanks for the clarification! I believe simtrial is timefix = FALSE, and I agree a vignette would be super useful!