As noted in Dr. Massie’s review, th

As noted in Dr. Massie’s review, the sponsor’s goals for the phase 3 studies wereambitious, in that they aimed to demonstrate effects for both elderly and non-elderly patientson both sleep maintenance and sleep onset, in terms of both an objective and a subjectiveassessment for each, in the same study. The sponsor also powered the studies for the highdose (the low dose had 30- to 40% less patients), and the multiplicity method tested the highdose first. The studies were thus underpowered for the low dose, such that non-statisticallysignificant findings for some endpoints at some time points for the low dose is unsurprising,and in no way provides interpretable evidence against efficacy. Likewise, the phase 2 studywas powered for objective sleep maintenance, but not latency or subjective sleep endpoints,such that non-statistically significant findings can not be taken as meaningful evidence oflack of efficacy at lower suvorexant doses. The question, discussed below, then becomes ifenough evidence has been provided in the overall database for appropriate dose-selection,even though studies were underpowered to provide this data in the form of consistentlystatistically positive endpoints.