Chapter 9- page 17

Discussion in 'CT4' started by nluashok, Feb 9, 2014.

  1. nluashok

    nluashok Member

    Hi,

    I am difficult to understand below quoted lines given in chapter 9 page 17 for estimating regression parameters.

    "Unlike the Kaplan-Meier method, this partial likelihood considers observed deaths only, not the times at which the deaths occurred, nor any censoring observed between deaths."

    I know that partial likelihood don't takes times of death i.e fine. However as per my understanding from material, it takes censoring into consideration.

    Can you please guide why above lines of material says that partial likelihood don't take censoring into consideration?

    Regards,
    Ashok
     
  2. John Potter

    John Potter ActEd Tutor Staff Member

    Hi Ashok,

    Yes, I must confess, I'm also confused on this one. I'm not sure there's any difference between the treatment of censored data in the two methods. They both ignore the times of the censored data but keep track of population numbers.

    Just ignore this paragraph until you hear otherwise. I will get it changed in the notes or get back to you if I discover that you and I have both misunderstood something,

    John
     
  3. Hemant Rupani

    Hemant Rupani Senior Member

    Censoring time in Kaplan Meier Model is in between deaths commonly.
    But censoring takes place in Cox PH Model individually at the time of exit.
    So the given statement is correct.
     
  4. John Potter

    John Potter ActEd Tutor Staff Member

    Hemant,

    You just seem to have repeated the bit of the notes that Ashok was a bit unsure about?

    I think I still agree with Ashok that it is a confusing sentence in the sense that both Kaplan-Meier and the Cox regression model take into account censoring that occurs between the deaths.

    If deaths occur between deaths then they do NOT then get included in the d/n in the Kaplan-Meier method. They do NOT get included in the denominator of the contribution to the partial likelihood,

    Good luck!
     
  5. Hemant Rupani

    Hemant Rupani Senior Member

    Oops! :eek: :p
     

Share This Page