Mohamed Yehia Junior Egypt, 20170815 15:33 Posting: # 17687 Views: 1,126 

Hi All, Hope all of you are doing well. I have 2 questions regarding 2 treatment 3 period 3 sequence BE study: 1) How to calculate CVintra and CVinter for Cmax using proc GLM; (According to my knowledge; in 2*2*2 we use proc mixed and we use Residual to get CVintra and Subject(Sequence) to get CVinter using the following formulas): CVintra= 100*sqrt(exp(y)1); CVinter= 100*sqrt(exp(x)1); However, In proc GLM, I don't get the residual value to use in CVintra equation as I know that if CVintra > 30%, we have to use proc glm no proc Mixed. The code I am using for proc GLM is: ods html; proc glm data=ee; class subject treat per sequence; model auc0_t=treat per sequence subject(sequence); random subject(sequence); means treat /hovtest=levene(type=abs); output out=resids r=residual; run; Using this, I get the sum of the squared residuals but I dont know which to use?!!. Merged from followup post [Helmut]: 2) For Outlier Analysis using Residuals, Do I have to check for the outliers of each parameter separated? As for example; In AUC0INF I have missing data values. However, in AUC0T and Cmax I have no missing ones. Thanks. Yours, M.Yehia Edit: Relax; see also this post #8. 21 hours ≠ 14 days. You can edit you original post for 24 hours; see there. Followup post deleted. [Helmut] 