I have a dataframe of pre- and post-treatment serum lipid levels from 9 patients that I am trying to analyze in R. I want to see if a patient's baseline lipid level can predict the response to a class of medication, however, I am running into an issue where some patients received two different drugs and showed different responses to each drug. Below is an example of the dataset:
id prepost lipid conc treatment_x treatment_y treatment_x_y_combo
1 1 pre lipid x 36097 NR R NA
2 2 pre lipid x 17667 NR NR NA
3 3 pre lipid x 17949 NR R NA
4 4 pre lipid x 18839 NR R NA
5 5 pre lipid x 10965 NR NA NA
6 6 pre lipid x 40371 NA R NA
7 7 pre lipid x 38086 NA R NA
8 8 pre lipid x 118650 NA R NA
9 9 pre lipid x 58279 NA NA NR
For context, treatment x and treatment y are both in the same treatment class.
To answer my question, I was planning to run a t-test between responders (R) and non-responders (NR), I'm afraid I can't simply label each patient as an R or NR because some patients received two different drugs and showed two different responses to each drug. For example, id 1 received treatment x and y. They didn't respond to treatment x but responded to treatment y.
I tried to focus on just one treatment (i.e. treatment x) rather than the class of treatment, but it appears that all patients that received treatment x did not respond and all patients that received drug y responded, except for one.
I'm relatively new to R and data analysis in general, so I apologize if I am not explaining any of this correctly. I was hoping someone had any suggestions as to what I should do next. Is there a way to transform the data to correct the issue of different responses to different drugs? Or is there a test I could run that would take the multiple observations into account?