My question is similar to Sample size calculation for multivariable Cox regression for treatment comparison. I have a prospective longitudinal study over three years. The purpose of the study is to see if the event variable can be predicted by a time-dependent covariate (there is a possibility that we could also include the second time-dependent covariate in the model). We have discerned time survival as there are a limited number of time points as the patient comes to the clinic every four months to do a test. We will have censor data. (there is no pilot data for this study). What kind of design could this study have and how could the sample size be calculated?
$\begingroup$
$\endgroup$
Add a comment
|