Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Comparing coefficients in different models

    I want to compare hazard ratios from cox models, across different survival models. Specifically I am looking at different measures of obesity (eg bmi and waist circumference) and I want to determine which has a stronger association with risk of cancer. I am also looking at these obesity variables per standard deviation - to make them more comparable.

    I have set my data up as follows,
    stset dateofexit fail(cancer==1) origin(age) enter(baseline)

    and then run the following two models:
    stcox bmi_sd i.sex i.education

    stcox wc_sd i.sex i.education

    I have HR and 95%CI for both models. But I want to now do a formal statistical test to say which of these variables/models (bmi or wc) has a stronger association with cancer risk (magnitude of effect).

    Any help would be much appreciated.

    Jess

  • #2
    Jessica,
    welcome to the list (as per FAQ, you are requested to register with your surname, too. Just click on the Contact us button at the bottom right of the screen).
    As far as your question on models is concerned:
    - nested models: you can use the likelihood ratio test or the Wald test to choose between them;
    - non-nested models: Akaike information criterion (Also known as AIC) can help you out.

    By the way, I would point you to the excellent: Mario Cleves, William Gould, Roberto G. Gutierrez, and Yulia V. Marchenko. An Introduction to Survival Analysis Using Stata, Third Edition. College Station, TX: Stata Press, 2010.

    Kind regards,
    Carlo
    Kind regards,
    Carlo
    (Stata 18.0 SE)

    Comment


    • #3
      Thanks Carlo

      Comment

      Working...
      X