A Latent Variable Approach for the Investigation of the Multi-factor Impact on Sleep Problems in Individuals with Mild Cognitive Impairment

The present study aims to explore the impact of neuropsychiatric symptoms, cognitive functioning, and medical status on sleep problems in individuals with Mild Cognitive Impairment (MCI), focusing on insomnia, by implementing a latent variable approach. The 131 Greek patients with MCI, who are consecutive visitors of the Alzheimer’s day clinic, underwent comprehensive assessments in cognitive functioning, neuropsychiatric symptoms, sleeping disturbances, and medical health status. The model was estimated by applying a Partial Least Squares (PLS) approach to structural equation modeling (SEM), incorporating the following latent variables: “psychological burden” (3 indicators: depressive symptoms, anxiety, and stress), “cognitive status” (5 indicators: general cognitive status, phonemic and semantic fluencies, psychomotor speed, and episodic memory), and “sleep problems” (5 indicators: overnight and earlier final awakening(s), decreased daytime functioning and well-being, and insufficient sleep duration). The results supported the SEM effectiveness, with good model fit and high predictive power. The “psychological burden” played the most significant role in the development of sleep problems within the clinical group of MCI, while “cognitive status” appeared to be the second most important SEM predictor, showing a negative association with “sleep problems”. Additionally, MCI age of onset and sleep problems severity were negatively associated. This latent-variable approach allowed us to capture the multidimensional complexity of sleep problems in MCI. Thus, applying integrative evaluations and interventions that target the cognitive status and psychological well-being, may improve the sleep quality of individuals with MCI, and further contribute to a slower cognitive decline course.