Sample size calculations for the primary and key secondary endpoints have been previously described [23]. Sleep onset and maintenance outcomes reported here include changes from base line in sSOL, sSE, sWASO and sTST assessed from sleep diary data, uptosixmonthsforsubjectswhoreceivedplacebo,LEM5orLEM10 in Treatment Period 1, and up to twelve months for subjects with continuous LEM5 or LEM10 treatment across Treatment Periods 1 and 2. Changes from baseline for quality of sleep rating and ratings of morning sleepiness/alertness were similarly analyzed. The p values for statistical comparisons (Treatment Period 1 only) were based on mixed-effect model repeated measurement analysis as previously described [23]. For sleep diary endpoints, reported values were calculated as the mean of the final seven nights before study visit, unless otherwise stated. Statistical analyses were per formed using SASv9.4 orother validated software by Firma Clinical Research as designated by the sponsor (Eisai). 335 J. Yardley, M. K arpp a, Y. Inoue et al. Sleep Medicine 80 (2021) 333e3

what is the statistical analysis of the study?