The Psychometric Properties of the Medical Outcomes Study Sleep Scale in Older Adults

Abstract

Decline in both sleep and gait is common in aging and is associated with a multitude of adverse outcomes. The Medical Outcomes Study Sleep Scale (MOS-SS; Hays, & Stewart, 1992) is a commonly used self report subjective sleep scale. The present study aimed to establish the psychometric properties of this scale in an older adult population. We specifically examined the internal consistency reliability, the factor structure, and the test-retest reliability of this scale. We also aimed to follow up on previous findings of associations between sleep and gait velocity in the elderly to determine whether there are associations between sleep and other outcome measures of gait. Participants were 673 non-demented individuals (mean age = 80.7 yrs; 61% female) enrolled in the Einstein Aging Study. Participants were administered the MOS-SS and were asked to walk on a 12 foot instrumented walkway (GAITRite, CIR systems, Havertown, PA) at their "normal walking speed" for two trials. Internal consistency reliability as measured by Cronbach's alpha was good ranging from .61-.72 on the various subscales. Principal component factor analysis (PCA) yielded three significant factors on this scale which accounted for 51% of the variance. Good test retest reliability was found as evident by the high correlations between scores at two separate time intervals. The correlation coefficient between the two time intervals ranged from .48-.71 on the various subscales. Separate linear regression analyses found that nearly all of the sleep subscales and summary measures were significantly associated with gait velocity and stride length, however only one subscale was significantly associated with stride length variability and cadence. The MOS-SS was overall found to be a reliable measure of subjective sleep in older adults although certain items on this scale may be better assessed via alternative methods.