Dear all,

I have a data set of 70 patient in a intervention foloow up study. 4 of the patient did drop out.

1. In my long data set I calculate the interaction between groups at follow up on the depended variable pcwp, however the pcwp at baseline was by random significantly different between the two groups, now is this model adjusting for the differnece when i use id_no as random intercept? As i know, using mixed model give the advantige of using my holde population even with missing values.

Code:
mixed pcwp i.group##i.visit || randomization_no:
Array




2. The next question is calculating the relative change of pcwp by calculating (followup pcwp - baseline pcwp/baseline pcwp) * 100 to get the percantage.
However i found out that the variable relchangepcwp is skewed, giving me a broad different median from mean: Array




My question is how do i get the most right confidence interval and p value of the relative change?

Thank you very much.