Discussion: View Thread

  • 1.  Current status data and survival

    Posted 07-03-2012 16:04
    To whom it may concern:

    Given an opportunity to present a result of a study completed by another statistician. The statistician used the term "current status data" . I am a little bit confused by this term compared to what I have known about suvival analysis.

    Here is the statment on the slide said

    "The admissible ages-at-deaths data was plugged into the Kaplan-Meier survivor function estimate for comparison with the NPMLE survivor function estimate from current status data"

    You guidance will be greatly appreicate that !

    Thank you !
    Cindy Weng


  • 2.  RE:Current status data and survival

    Posted 07-03-2012 16:44
    I agree with David, it's probably something simple.  I wonder if it just means the current status of the patients in this study.  The status is always chaning but at a particular point (i.e. the current status) you perform the analysis with the data you have.

    Just a thought.

    -------------------------------------------
    Rocco Brunelle
    Senior Statistician
    Bowsher Brunelle Smith LLC
    -------------------------------------------






    Original Message


  • 3.  RE:Current status data and survival

    Posted 07-03-2012 16:57
    I think David Bristol's guess is probably right.  it certaiinly makes a lot of sense.  My question would be what is the acronym NPMLE survivor function and why is it used?  My guess is that it stands for nonparametric maximum likelihood estimte of the survivor function.  But that is just the Kaplan-Meier estimate.  So why not just say that you are updating the Kaplan-Meier estimate based on admissible ages at death if that is what is going on?

    -------------------------------------------
    Michael Chernick
    Director of Biostatistical Services
    Lankenau Institute for Medical Research
    -------------------------------------------






    Original Message


  • 4.  RE:Current status data and survival

    Posted 07-03-2012 17:06

    Current status refers to the fact that an individual's status is available only at one time point during follow-up. At that time, the individual could be still alive or may have died. In this case, time of death will not be available. So, the current status data structure is conceptually different from the right -censored data (where, if failed, an individual's time to failure is fully observed). Another important aspects of current status data is that the NPMLE does not have a closed-form, as does the Kaplan-Meier estimator for right-censored data problems. Therefore, the current status NPMLE is obtained numerically. A useful reference is "The Statistical Analysis of Interval-censored Failure Time Data" by Jianguo (Tony) Sun.

     

    Thanks!


    -------------------------------------------
    Adin-Cristian Andrei
    Research Associate Professor
    Northwestern University
    -------------------------------------------






    Original Message


  • 5.  RE:Current status data and survival

    Posted 07-05-2012 18:40
    That is interesting, and considerably different from what I was expecting.  So if the subject's "current status" is equal to "dead", would we say that the subject is interval-censored?  with the interval's boundaries being the start date and the follow-up date?

    -------------------------------------------
    Eric Siegel
    Biostatistician
    Univ of Arkansas for Medical Sciences
    -------------------------------------------


    Original Message