In a clinical study, 100 patients are evenly divided into two treatment group, trtA and trtB. For each patient, a biomarker is measured at 5 different visit timepoints. Y is the measured biomarker level.

I'm using linear mix model to evaluate treatment effect along the time course. I can build two models with Y or log(Y) as below.

# model1
Y = lmer(Y~Visit+treatment+(1|subject)), data=data )
# model2 (log-transformed Y)
log(Y) = lmer(Y~Visit+treatment+(1|subject), data=data)

My question is: say treatment effect is significant, how should I interpret model result (significance of treatment effect), using original Y vs. log-transformed Y? (model1 vs model2). I feel somehow they should be different....

There is a statistically significant difference between 
trtA and trtB along the time course...?

One additional question: How should I interpret model result, when Visit is treated as factor variable vs continuous variable (number of days), say the treatment term is significant.

Thanks a lot in advance.


Your Answer

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

Browse other questions tagged or ask your own question.