A latent variable approach to multifactorial associations with sleep problems in individuals with mild cognitive impairment

Abstract

Background: This study examines key associations of psychological symptoms, cognitive functioning, and age of onset with sleep problems, especially insomnia, in individuals with mild cognitive impairment (MCI), using a latent-variable approach. Methods: A total of 127 Greek individuals with MCI, consecutive visitors of an Alzheimer’s day clinic, underwent comprehensive assessments of cognitive functioning, psychological symptoms, and sleep disturbances. The model was examined using a Partial Least Squares approach to structural equation modeling, incorporating three latent constructs: ‘psychological burden’ (depressive symptoms, anxiety, and stress), ‘cognitive status’ (general cognition, phonemic and semantic fluency, executive functioning/mental flexibility, and episodic memory), and ‘sleep problems’ (nocturnal awakenings, early awakenings, insufficient duration, impaired daytime functioning, and reduced well-being). Results: The model demonstrated acceptable fit, predictive relevance, and superior predictive performance compared to linear regression. ‘Psychological burden’ was the strongest predictor of sleep problems, whereas ‘cognitive status’ showed a significant negative association, with lower cognitive abilities linked to greater sleep disturbances. Earlier MCI onset was also associated with more severe sleep problems. Conclusion: The latent-variable approach revealed the multidimensional complexity of sleep problems in MCI, highlighting the need for integrative evaluations and interventions that target cognitive status and psychological well-being as clinically relevant targets for assessment and management.