Choosing variables for mutivariable survival prediction

data analysis variable selection
21

I think that’s right, without further understanding the study. A few rule:

  • For the chosen index time (diagnosis, surgery, first exposure to your health system, …) everything being analyzed needs to be fully prospectively defined
  • No exclusions of observations are allowed after the index time
  • Assessments, including pathology and ‘omics’, should be masked to patient status if the assessment is at all subjective

Frank

22

Thank you again for the great inputs, Frank.

I will now take the time to process all of the information I received in these past few days and properly plan the next steps of my research.

1 Like