Abstract
STUDY OBJECTIVES: To shed light on understanding sleep duration trajectories (SDTs) using different classification methods and their outcomes, this study aimed to (1) identify common SDTs among different age groups, (2) investigate the alignment versus differences between SDTs identification by group-based trajectory modeling (GBTM) and clinical standards, and (3) examine the impacts of SDTs on health outcomes. METHODS: A systematic literature search from four databases yielded 34 longitudinal SDT studies with GBTM analyses spanning three or more data waves. Apart from the proportion meta-analysis, a three-level meta-analysis was conducted with 14 of the studies that examined the association between SDT groups and health outcomes. Assessment of study quality was performed using the Guidelines for Reporting on Latent Trajectory Studies checklist. RESULTS: Qualitative analysis identified four age-related SDT classes based on longitudinal trends: "persistent sleepers," "increase sleepers," "decrease sleepers," and "variable sleepers." Meta-analysis also showed differential proportions of "GBTM-defined shortest sleepers" across age groups and sample regions, as well as significant discrepancies in the prevalence of short sleep identified by clinical standards (=50% vs. 15% per GBTM). Overall, SDTs predicted emotional and behavioral outcomes, neurocognitive problems, and physical health (OR = 1.538, p < 0.001), in GBTM-defined "short," "fluctuating," "long," and "decreasing" sleepers as compared to the "adequate" group. The effects were stronger in adolescents and in datasets with more waves. CONCLUSIONS: The identification of the GBTM-defined "short," "fluctuating," "long," and "decreasing" SDT groups and their associations with various health outcomes supported longitudinal investigations, as well as the development of interventions focusing on both the length and stability of sleep durations, especially in younger populations. Study registration: PROSPERO registration number CRD42023412201.