Heterogeneity of sleep quality based on the Pittsburgh Sleep Quality Index in a community sample: a latent class analysis

Original Article

DOI: 10.1007/s41105-017-0097-7

Cite this article as:
Yildirim, A. & Boysan, M. Sleep Biol. Rhythms (2017). doi:10.1007/s41105-017-0097-7
  • 24 Downloads

Abstract

This study aims to assess the latent dimensional structure of sleep quality as measured by seven components of the Pittsburgh Sleep Quality Index (PSQI) and evaluate its diagnostic utility in discriminating individual differences on circadian preferences. Three hundred sixty-seven subjects, aged 17–58 years (mean 22.3 ± 6.3) and 55% female, participated in the study. The PSQI and Morningness–Eveningness Questionnaire (MEQ) were administered, and latent class analysis was run to assess latent homogeneous subgroups according to seven components of the PSQI. Latent class analysis revealed that sleep quality is multifaceted, and data distribution fits best to two-class model. About two-thirds of the subjects (n = 289) were classified into poor sleep quality class and 78 participants were grouped into good sleep quality class. A PSQI total ≥5 was identified as the cut-off value for an optimal discrimination between these two latent classes. Three-step regression analysis did not demonstrate a significant relationship between circadian preferences and sleep quality. Finally, signal detection analysis showed that the PSQI total cut-off value had low diagnostic utility with respect to the individual variation in circadian preferences. Sleep quality is a distinct psychological construct from circadian preferences. Sleep problems were prevalent in the study population and developmentally sensitive sleep programs are required.

Keywords

Diagnostic utility Diurnal preferences Mixture analysis Sleep disturbance assessment 

Copyright information

© Japanese Society of Sleep Research 2017

Authors and Affiliations

  1. 1.Department of PsychiatryYuzuncu Yil University School of MedicineVanTurkey
  2. 2.Department of PsychologyYuzuncu Yil University Faculty of Social SciencesVanTurkey

Personalised recommendations