I have a design with patients treated with a certain drug. We have patient data before and after taking drug. Within in these patients there is are two types of people, responders and nonresponders in matched pairs (one set responds to the treatment and one does not). Responders and nonresponders are matched by age, BMI and sex. Dependent variable is amount of certain Y measured before and after treatment.
I am not sure if I need to use two separate tests for within and between effects. I tried repeated measures Anova with subject id as random effect. For between subjects my thought is I can calculate % change in Y before and after treatment for each person and go for a paired t.test between responder and non respondermatched pair.
Is there a way to answer all the questions in single model/test?