I have longitudinal data following patients before and after a dichotomous treatment. Exploratory analyses strongly suggest that the effect of the treatment depends on the initial burden of the patient: If a patient's burden is relatively low the treatment won't help the patient much, but if the initial burden of the patient is high, the treatment will help a lot.

How can I analyze this using inferential statistics? How can I estimate the treatment effect depending on the initial burden of the patient, and obtain its standard error?

I initially thought about seperating the patients in groups based on the intial burden and then to run a simple moderator analysis. This procedure seems however very crude to me. Isn't there some more appropriate technique?


Sounds like a case for mixed linear models to me. Basically you would compare patients before and after treatment, accounting for both each patient, and each patient's burden. It would be weird to post the same answer so I'm linking a (slightly different) question, but I believe the same answer applies: How to account for participants in a study design?

  • $\begingroup$ Yes. This sounds like pretty standard case of mixed models or GEE. $\endgroup$ – StatsStudent Oct 27 '15 at 15:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.