Paired longitudinal analysis

Hello all,

I need your advice.

I am examining an new eye drops. In order to examine, an experiment was performed, in which a subject was enrolled, and one eye was randomized to get the drops, while the other was the control. Before giving the drops, some quantitative measure was taken on each eye, and D = intervention eye - control eye was calculate for each subject. After taking the baseline measurements, the drops were given. And the subjects were measured again in 3 more follow-up visits. Now I want to find a way of saying if the drops work, and if they don't, when do they stop working.

If at time 0, before giving the drops, the mean difference is 0 with a standard deviation of 1.5, how would you determine if at time point 2 or 3 it is still working or not ? Should I take a mean difference of more than 1.5 ? Or double that size ?

Is there a point in doing a paired t-test at every time point ?

Thanks !