1
$\begingroup$

I have read that the following formula is used to estimate the cumulative baseline hazard function:

Cumulative Baseline Hazard Approximation

I also know that the basehaz command in R will also calculate an estimate for the cumulative baseline hazard.

When i build a 1 variable coxPH model in R using the BRESLOW method for ties, the output from the basehaz command matches up with my manual calculations made using the formula given above

However, when i build the same coxPH model using the EFRON method for ties, my manual calculations do not match up with the basehaz output from this coxPH model (see findings below)

Manual Calculations vs Basehaz

My question is, how is the basehaz calculated when EFRON ties are used as it simply cannot be done using the formula given in the first picture.

Improve this question
$\endgroup$
2
  • $\begingroup$ Your $t$ is $2,3,4,5,6,7,8$ it seems there are no ties at all. $\endgroup$
    – Deep North
    Commented Oct 17, 2017 at 11:55
  • $\begingroup$ I have summarized the data in the tables. For example, at time t=2 there were 373 deaths, at time t=3 there were 431 deaths and so on. I can provide a more detailed spreadsheet that shows exactly how each of these things were calculated if you would like? $\endgroup$
    – Alex
    Commented Oct 17, 2017 at 14:40

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.