Latent profiles of actigraphy-derived sleep regularity, timing, and duration in adults with type 2 diabetes: clinical and psychosocial correlates

Abstract

Objectives: This study aimed to identify distinct latent sleep profiles based on sleep regularity, timing, and duration in adults with type 2 diabetes (T2D), and examine associations with sociodemographic, clinical, and psychosocial factors and subjective sleep quality. Methods: Seven consecutive days of actigraphy data for 183 adults with T2D (mean age 53.9 ± 13.9 years) were analyzed. Five metrics were computed: Sleep Regularity Index, mean and standard deviation (SD) of mid-sleep timing, and mean and SD of total sleep time. Latent profile analysis (LPA) was applied to the five sleep indices to identify distinct latent sleep profiles. One-way ANOVA with Tukey post-hoc and Pearson chi-square examined sociodemographic, clinical, and psychosocial correlates of profile membership and differences in subjective sleep quality (Pittsburgh Sleep Quality Index; PSQI). Results: LPA identified four profiles: Moderately Irregular Short Sleep (36.1%), Moderately Regular with Inconsistent and Insufficient Duration (37.7%), Highly Irregular Sleep (14.2%), and Highly Regular Sleep (12.0%). Moderately Irregular Short Sleep was the oldest profile (59.4 ± 12.1 years, p < .001), with the highest HbA1c level (7.7%, p < .001), lowest perceived social support (p = .020), and poorest subjective sleep quality (global PSQI = 7.5, p = .004). Highly Regular Sleep was the youngest (46.2 ± 14.5 years), with the best glycemic control (HbA1c = 6.8%) and best subjective sleep quality (global PSQI = 4.0, p = .004). Conclusions: These findings suggest that sleep regularity, timing, and duration co-occur in clinically distinct patterns in adults with T2D. Person-centered approaches such as LPA may offer a more precise framework for identifying high-risk individuals and tailoring behavioral interventions in diabetes care.