You called? Check out Box 1 in this paper: journals.sagepub.com/doi/10.1177/...

I first read a paper on within-arm comparison written by late prof Altman and colleague many years ago and advocated no within-arm comparison for RCT data analysis proposed at all. unfortunately some PIs cravingly requested statistician including myself doing the within-arm comparison.

I actually fancy 1 stage if the whole data was treated as data for a mlm regression.