Dear statalisters,
I have some panel data that include four binary clinical outcomes (0=absent, 1=present) at two different timepoints t1 and t2, and a number of x vars.
I am interested in exploring the trajectories for each outcome, i.e. healthy (t1=0, t2=0), incident (t1=0, t2=1) remittent(t1=1, t2=0) and persistent (t1=1, t2=1), and explore how each xvar affects the trajectory of the conditions.
I guess the easiest way of doing this could simply running an mlogit model, where yvar is a categorical var with the four levels described above. would this approach be correct?
or should I explore this research question using mixed effect models?
thanks for your time,
RR
I have some panel data that include four binary clinical outcomes (0=absent, 1=present) at two different timepoints t1 and t2, and a number of x vars.
I am interested in exploring the trajectories for each outcome, i.e. healthy (t1=0, t2=0), incident (t1=0, t2=1) remittent(t1=1, t2=0) and persistent (t1=1, t2=1), and explore how each xvar affects the trajectory of the conditions.
I guess the easiest way of doing this could simply running an mlogit model, where yvar is a categorical var with the four levels described above. would this approach be correct?
or should I explore this research question using mixed effect models?
thanks for your time,
RR
