Patterns of sleep quality during and after postacute rehabilitation in older adults: a latent class analysis approach

Abstract

Sleep quality is related to emotional, physical, psychological and cognitive functioning and functional independence in later life. After acute health events, older adults are likely to utilize postacute rehabilitation services to improve functioning and facilitate return to independent living. Patterns of how sleep changes with postacute rehabilitation, and predictors of such patterns, are unknown. The current investigation employed latent class analysis (LCA) methods to classify older adults (n=233) into groups based on patterns of self-reported sleep quality pre-illness, during postacute rehabilitation and up to 1year following postacute rehabilitation. Using LCA, older adults were grouped into (1) consistently good sleepers (46%), (2) good sleepers who transitioned into poor sleepers (34%), (3) consistently poor sleepers (14%) and (4) poor sleepers who transitioned into good sleepers (6%). In three planned analyses, pain was an independent predictor of membership in classes 1 or 2 (good pre-illness sleep quality) versus classes 3 or 4 (poor pre-illness sleep quality), and of membership in class 1 (consistently good sleep) versus class 2 (good sleep that transitioned to poor sleep). A lower Mini-Mental State Examination score was a predictor of membership in class 1 versus class 2. There were no statistically significant predictors of membership in class 3 versus class 4. Demographics, comorbidities and depressive symptoms were not significant predictors of class membership. These findings have implications for identification of older adults at risk for developing poor sleep associated with changes in health and postacute rehabilitation. The findings also suggest that pain symptoms should be targeted to improve sleep during postacute rehabilitation.