I run univariate and multivariate COX proportional analysis to find the factors that were related with time to an event (time to having a certain type of disease). However, since I have a pretty high number of patients in whom bilateral eyes of the same subjects were included, I'd like to ask what is the best way for controlling 2 eyes of same subjects in SPSS? Is there any link or an example showing how to run such a model?
Use multilevel or mixed models. You can regard the eyes as nesting at level 1 within a subject at level 2 and this allows for correlated responses - the design is a like twin study that is individuals at level 1 in a twin pair at level 2.
see in general
http://www.bristol.ac.uk/cmm/software/mlwin/mlwin-resources.html#multipleresp
more specifically
http://www.bristol.ac.uk/cmm/software/support/workshops/materials/eha.html
and in R
https://cran.r-project.org/web/packages/coxme/vignettes/coxme.pdf
for twins
http://www.stat-d.si/mz/mz1.1/xu.pdf
and in Stata
http://www.stata.com/new-in-stata/multilevel-survival-models/
In these models there would be an additional random effect for the individual that potentially applies to both eyes
You can apply a so-called frailty model which assumes similar frailty for matched clusters, or, as in your case, the same individual. Therein the model call is similar compared to a Cox model, but you have to add a random effect hth
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS