Announcement

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

  • Reporting relative risk after flexible parametric model on propensity score matched cohort

    Hello all. I have a propensity score matched cohort (treatment and control) and wish to evaluate the effect of treatment on the risk of failure (time to event) of a surgical procedure. I have established that the hazards are not proportional between the control and treatment cohorts, and so have opted to use a flexible parametric survival model with time varying covariates:

    Code:
    stpm2 treated, scale(h) df(3) tvc(treated) dftvc(2)
    I would like to report both absolute as well as relative risk. I am planning on reporting absolute risk by reporting the restricted mean survival time at x years (2 years, 5 years, for instance).

    My question is: what is the best way to report the relative risk? The instantaneous hazard ratio clearly varies over time, and reporting it at a given time point does not really reflect the relative risk until that time point from the start.

    Suggestions/advice very much welcome. Thanks in advance.
Working...
X