Abstract
Introduction
Sleep health, a construct introduced to characterize the multidimensional attributes of sleep, has been explored in a variety of populations; however has not been adequately examined for middle-aged and older adults. As attributes of sleep may change with age, the dimensional structure of sleep health may differ in this population. This study aimed to validate a composite measure of sleep health among middle-aged and older adults using data from the Health and Retirement Survey (HRS).
Methods
Data from the 2014 Core survey of the HRS was used to create a composite measure of Sleep Health including sleep efficiency, duration, timing, satisfaction, and alertness. We standardized and averaged the original variables before transforming to T scores. Sleep Health T scores (ranging 0-100, higher scores indicating better sleep health) were examined using exploratory and confirmatory factor analysis (EFA; CFA).
Results
Our sample included 6,095 adults with mean age of 68 years (SD=10.1; range 50-99 years). The majority were female (59.7%), white (77%), with high school education (53.9%). Sleep Health T scores ranged from 27-61 (mean=50; SD=6.7). EFA identified one factor. Timing was removed due to low factor loading (<0.4). The revised four-dimension composite Sleep Health measure had acceptable reliability (Cronbach’s alpha 0.6). CFA showed a well-adjusted model (REMSA=0.097; NFI=0.964; RMR=0.035; GFI=0.990; AGFI=0.951).
Conclusion
These results suggest that the composite measure was valid for assessing sleep health among middle-aged and older adults. Limitations include the use of secondary data, as sleep health dimensions were based on variables not created specifically for our research question. Future research should further examine the role of sleep timing in overall sleep health among middle-aged and older adults.
Support
This study was supported by the University at Buffalo Clinical and Translational Science Institute (CTSI) funded by the National Institutes of Health (Lorenz, PI).