Latin Square design time normalization

I have a latin square design with
  • 3 treatment levels ( control, treatment 1&2) between groups
  • 4 Dose levels (drug X) which was administered to all the subjects
The total duration of the study was about 28 days where each subject was treated everyday with either of the treatments and was dosed with drug X in a latin square manner for final 2 weeks .
Is there any way to normalize the time scale (since drug X was randomly administered after every 3 days to the different treatments) so that I can visualize the response vs time for drug X at each dose level for the different treatments.
I was thinking about using the value of response 1 day prior as my baseline and calculate the change from baseline for the next 3 days i.e pre-drug administration, drug effect, post-drug effect. So, this would give give me time on X axis and response i.e change from baseline on Y axis as a continuous dependent variable and time as the independent variable. Would doing this change data and induce any bias?

Thank you,