Article Text

This article has a correction. Please see:

other Versions

Geographic variations in sleep duration: a multilevel analysis from the Boston Area Community Health (BACH) Survey
  1. Shona C Fang1,2,
  2. S V Subramanian3,
  3. Rebecca Piccolo1,
  4. May Yang1,
  5. H Klar Yaggi4,
  6. Donald L Bliwise5,
  7. Andre B Araujo1
  1. 1New England Research Institutes, Inc, Watertown, Massachusetts, USA
  2. 2Department of Environmental Health, Harvard School of Public Health, Boston, Massachusetts, USA
  3. 3Department of Social and Behavioral Sciences, Harvard School of Public Health, Boston, Massachusetts, USA
  4. 4Department of Internal Medicine, Yale University School of Medicine, New Haven, Connecticut, USA
  5. 5Department of Neurology, Emory University School of Medicine, Atlanta, Georgia, USA
  1. Correspondence to Dr Shona C Fang; sfang{at}


Background Sleep plays an important role in health and varies by social determinants. Little is known, however, about geographic variations in sleep and the role of individual-level and neighbourhood-level factors.

Methods We used a multilevel modelling approach to quantify neighbourhood variation in self-reported sleep duration (very short <5 h; short 5–6.9 h; normative 7–8.9 h; long ≥9 h) among 3591 participants of the Boston Area Community Health Survey. We determined whether geographic variations persisted with control for individual-level demographic, socioeconomic status (SES) and lifestyle factors. We then determined the role of neighbourhood SES (nSES) in geographic variations. Additional models considered individual health factors.

Results Between neighbourhood differences accounted for a substantial portion of total variability in sleep duration. Neighbourhood variation persisted with control for demographics, SES and lifestyle factors. These characteristics accounted for a portion (6–20%) of between-neighbourhood variance in very short, short and long sleep, while nSES accounted for the majority of the remaining between-neighbourhood variances. Low and medium nSES were associated with very short and short sleep (eg, very short sleep OR=2.08; 95% CI 1.38 to 3.14 for low vs high nSES), but not long sleep. Further inclusion of health factors did not appreciably increase the amount of between-neighbourhood variance explained nor did it alter associations.

Conclusions Sleep duration varied by neighbourhood in a diverse urban setting in the northeastern USA. Individual-level demographics, SES and lifestyle factors explained some geographic variability, while nSES explained a substantial amount. Mechanisms associated with nSES should be examined in future studies to help understand and reduce geographic variations in sleep.

  • Neighborhood/place

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Corrections
    BMJ Publishing Group Ltd