May not have understood, but when you say "my outcome is binary (at the last measurement)" it sounds like you have only one occasion of the outcome measured. If that is the case then it is not data for a longitudinal/ repeated measures model. You need the outcome to be repeated measures for a longitudinal model.

If I have misunderstood, and you do have outcomes measured at repeated time points, I would look at a mixed effects (not mixed methods as you called it - i am sure that was just a typo).

If you have the outcome at one time point only then it's not a longitudinal model - but you can still incorporate repeated measures of explanatory/ independent variables.

I would suggest you determine whether you have repeated outcome measures or not, then take your model development from there.

Thank you for your answer Brian Kelly . Yes, you did not misunderstand me, the outcome is measured as pain only at the last measurement.

So, have I understood you right that even though my independent variables are measured repeatedly over a longer time I only need to take in consideration of the repeated measures and not the time?

Yes, that would be correct.

As your outcome/ dependent measure is only at one time point you would not have to consider time in relation to the outcome, so not a longitudinal model (no variation over time to model).

That is not to say that time may or may not be important in your research question. If trends/ differences/ averages in the repeated measures of independent variables are important in relation to your outcome then you can find ways to incorporate these things into your modelling strategy (in the way that you choose to use your repeated independent measures - being guided by research questions).