Statistical Analysis of fixed sequence 3-period 2-treatment (ABB) crossover design

#1
I want to test the efficacy of a medicine for the treatment of blood pressure and take one measurement while on placebo (A) treatment for one month, second measurement after 3 months of treatment with medicine B and 12 months after treatment with the same medicine B. The trial will be a double-blind study till the 3rd month's data collection and open-label after that.This fixed sequence 3-period 2-treatment (ABB) crossover design study is to be done on about 150 patients.
Please guide me regarding the best statistical analysis method
 
Last edited:
#2
Since nobody guided me for the correct statistical procedure, I read about the possible steps. A result of Shapiro-Wilk test for Normal distribution at three-month treatment results --accept normality. So, I am thinking of Paired samples t-test for the three-month treatment result.
Shapiro-Wilk test for Normal distribution for one-year treatment results –reject normality. So, I am thinking of Wilcoxon test for paired for the one-year treatment result.
Please advise me whether this is correct. Also please let me know, is it necessary to do logarithmic transformation always for Wilcoxon test?