I am working on a study where patients with a chronic disease received a treatment. The goal is to determine the effect of the treatment on various outcome variables. I have pre and post data (before and after the treatment) for the sick patients and also for a control group of patients who are healthy. It would have been considered unethical to withhold treatment from sick patients, hence there was no control group of sick patients who did not receive the treatment.

I plan to analyze the data using linear mixed effects models. I’ve seen examples of how to analyze this type of data when you have pre/post data with a control group, for example including variables in the model such as the time point, whether the individual is in the treatment or control group, and the interaction between the two. However, I’m wondering about the appropriateness of this approach when the control group is healthy patients. Does it really make sense to make comparisons to a healthy group in this way when the goal of the study is to determine what effect the treatment had? By making comparisons to a group of healthy patients, is this answering a different question?

Appreciate any comments. thank you!


This makes no sense to do because the average outcome of the healthy patients is meant to stand in for the average outcome of the sick patients had they not received treatment (which is not observed due to the fundamental problem of causal inference/unbearable lightness of being).

Without some additional strong assumptions, there is not much else you can do with this design. The difference-in-differences approach you outline above assumes that the sick when untreated share a trend with the healthy (though there may be a constant gap between them), which seems unlikely to hold.

| cite | improve this answer | |

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.