Minimal Clinically Important Difference for the Visual Analog Scale for Sleep Inertia Using Data from a Phase 3 Trial of Low-sodium Oxybate for Idiopathic Hypersomnia
Richard Bogan1, Douglas Fuller2, Marisa Whalen2, Cristina Casstevens2, Wayne Macfadden2, Logan Schneider3
1University of South Carolina School of Medicine, 2Jazz Pharmaceuticals, 3Stanford University Center for Sleep Sciences and Medicine
Objective:
Propose a minimal clinically important difference (MCID), anchored in the Patient Global Impression of Change (PGIc), for the visual analog scale for sleep inertia (VAS-SI).
Background:
The VAS-SI assesses severity of sleep inertia (difficulty awakening), a common, debilitating idiopathic hypersomnia symptom; however, the MCID has not been determined.
Design/Methods:
Participants from a phase 3 trial (NCT03533114) of low-sodium oxybate (LXB) for idiopathic hypersomnia began open-label LXB treatment (10–14 weeks), followed by a 2-week stable-dose period (SDP). Participants were randomized to placebo or continued LXB during a 2-week, double-blind, randomized-withdrawal period (DBRWP). Using the VAS-SI, participants rated their difficulty awakening, on a 100-mm line anchored at 0 (very easy) and 100 (very difficult), at baseline, end of SDP, and end of DBRWP. Participants also rated their change in condition on the PGIc, a 7-point Likert-type scale (very much improved to very much worse), at end of SDP and DBRWP. The MCID was estimated based on change in VAS-SI and PGIc scores assessed via Kruskal-Wallis test and a linear mixed model (LMM) with repeated measures.
Results:
Participants (N=109) were mean (SD) 40.8 (14.1) years of age, primarily female (70%) and White (82%). Median (quartile 1, quartile 3) changes in VAS-SI for PGIc levels were 52.4 (31.6, 59.6) for very much worse; 18.8 (2.4, 34.3) for much worse; 2.5 (–3.2, 14.0) for minimally worse; 0.4 (–3.0, 6.7) for no change; –12.3 (–17.4, 2.2) for minimally improved; –15.6 (–35.7, –5.1) for much improved; and –28.7 (–51.6, –17.6) for very much improved (Kruskal-Wallis test statistic, 110.2; P<0.001). With an LMM, the mean (SE) difference in VAS-SI scores between consecutive PGIc levels was 10.9 (0.8).
Conclusions:
Using an anchor-based approach, a minimal clinically important difference of 10–12 mm is proposed for the visual analog scale for sleep inertia.