Statistic analyses with multiple values per patient

rstudio

#1

Dear R community.

Im just learning about R and i got stuck analysing my data.
I’m not sure that this is what this community is for but it cant hurt to try.

So Im measuring biochemical response to treatment with a new type of medicine.
a good deal of the patients develops inhibitors that seem to reduce treatment efficacy.
What I would like to measure is the correlation between the inhibitors and the biochemical marker (the response). However each patient has been measured between 1 and 10 times. total amount of patients is 25. I obviously cant use spearman or pearson because that would assume all data are independent and thats not the case. Ive been told a mixed model cant be used although i did not really understand why not. So now I was thinking of using a GEE model (after transforming the inhibitors to log10). Can anyone advise whether or not that is the right way to go? And give me some pointers on how to proceed in R? Ive added a scatterplot of the data to give you an idea of how it looks. The different collors are high and low dose of treatment.


#2

I noticed you never got a response to this question :disappointed:. Have you tried posting at Cross Validated?

To understand how that community works and to maximize your chances of having a good experience, take a look at these pages before posting over there: