How to analyze continuous Y and impact of time of event

Retrospective data should be used to analyze a continuous Y (namely estimated glomerular filtration rate, eGFR) and the impact of infections on eGFR after kidney transplantation. It is plausible for early infections to decrease (long term) eGFR more than later occurring infections. It is also hypothesized that a higher number of infections also decreases eGFR.
In RMS 17.6.4 approaches to incorporate time (and/or severity) of an event in survival analysis are mentioned.

Are there similar approaches for continuous Y?
Would an analysis with time-varying variables be appropriate?
Should we model eGFR or its slope?

Any hints are appreciated!

Edit: added after kidney transplantation (and a typo…)

Early and late infections with respect to what? Do you have a well-defined baseline or index date for the people in your study?

Ah, I forgot (and edited the orirginal post): it’s about the course of eGFR after kidney transplantation. So early and late refer to the time after kidney transplantation. However, “early” and “late” are ill-defined, we could apply some arbitrary thresholds but would prefer avoiding this categorization.

1 Like

Thanks for the update! The approach proposed in this paper might be of interest to you:

Edit: link to Pubmed page: Modeling of cumulative effects of time-varying drug exposures on within-subject changes in a continuous outcome - PubMed

1 Like

Thank you, this was an excellent starting point!

1 Like