PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of jepicomhInstructions for authorsCurrent TOCJournal of Epidemiology and Community Health
 
J Epidemiol Community Health. Mar 2006; 60(3): 267–269.
PMCID: PMC2465560

Socioeconomic position and sleep quantity in UK adults

Abstract

Study objective

To investigate the association between markers of socioeconomic position (years of full time education, access to a car or van, and occupational social class) and mid‐range sleep quantity (6.5 to 8.5 hours).

Data source

Office of National Statistics Omnibus Survey, May 1999.

Participants

3000 households were randomly selected from the UK postcode address file and one person aged over 16 years from each household was randomly selected and invited to take part in the survey. All respondents aged 25 and over (n = 1473) were included in the analyses where appropriate data were available.

Main results

Total sleep quantity was greater in more deprived women as measured by years in full time education. There was no linear association between total sleep quantity and any of the markers of socioeconomic position in men. Mid‐range sleep quantity was more common in more educated women. Similar, non‐significant, trends were seen in men.

Conclusion

There is little evidence that more socioeconomically deprived people obtain less sleep than more advantaged ones—indeed, the reverse may be true—but some evidence that more advantaged women are more likely to report mid‐range sleep. Any hypothesis implicating sleep in socioeconomic inequalities in health should take into account mid‐range, rather than total, sleep quantity. Further work should focus both on mid‐range, or “healthy” sleep, quantity as well as sleep quality.

Keywords: socioeconomic position, inequalities, sleep

There is evidence that sleep deprivation is associated with impaired glucose tolerance and hypertension.1,2 As these disease markers, as well as numerous disease outcomes, vary according to socioeconomic position (SEP),3 it has been suggested that variations in sleep quantity may mediate the relation between SEP and health.4 Theoretical reasons for this include known socioeconomic variations in stress,5 housing quality,6 and overcrowding.7 However, the association between sleep quantity and both morbidity and mortality seems to be U shaped with lowest risk at around six to eight hours of sleep per night.8,9,10,11

Published data on SEP and sleep quantity are sparse but have failed, to date, to report a linear association.12,13,14 In fact, there is some evidence of an inverse linear association between SEP and sleep.14 One recent analysis of US data investigated socioeconomic variations in short (<6.5 hours), mid‐range (6.5–8.5 hours), and long (>8.5 hours) sleep finding that lower educational attainment was associated with increased risk of being a short or long sleeper.15

To provide further information on the association between SEP and sleep quantity, and the first such data from the UK, data from the UK Office for National Statistics' (ONS) Omnibus Survey were analysed.

Methods

The ONS Omnibus Survey is a regular, multi‐purpose, survey.16,17 Each month 3000 private households in the UK are randomly selected using the postcode address file and one randomly selected person aged over 16 from each household is invited to take part in the survey. In May 1999, respondents were asked to complete a time use survey for one 24 hour period from the past seven days, recording their main activity in each of 96 15‐minute periods.

Total sleep quantity was calculated by assuming that each 15‐minute period where the main activity reported was sleep or rest consisted entirely of sleep. People were also dichotomised into mid‐range and other sleepers with healthy sleepers defined as those reporting between 6.5 and 8.5 hours of sleep.15

Age when respondents left full time education (collapsed into three groups), whether or not a car or van was available to the household, and occupational social class of the respondent, or their head of household if they were not currently working (collapsed into three groups using the customary groupings of classes I and II, class III, and classes IV and V), were used as markers of SEP—all known to be useful measures of SEP in the context of socioeconomic inequalities in health.18,19 The association between markers of SEP and sleep quantity, after control for age, was assessed using linear regression. The odds of people in different social groups being mid‐range sleepers, after control for age, were assessed using logistic regression.

As SEP may not be stable in young adulthood, people aged less than 25 were excluded throughout. All analyses were conducted separately in men and women as there are a number of examples of the relation between SEP and health variables differing between men and women.20,21,22 People were excluded from specific analyses if the required data for those analyses were missing. All analyses were conducted in Stata v8.0 using population sampling weights from the original dataset.

Results

A total of 1626 people responded to the survey, of whom 1473 (647 men and 826 women) were aged 25 or over. Total sleep was approximately normally distributed (mean 8.40 hours, standard error 0.05). Women reported significantly more sleep than men (mean 8.59 compared with 8.17 hours; p<0.001). There was no difference in the proportion of men and women reporting mid‐range sleep (54.3% of men and 51.0% of women, p = 0.203).

Table 11 shows mean hours of sleep according to age groups and markers of SEP for men and women separately, along with regression coefficients of the ability of the markers of SEP to predict sleep quantity after control for age where appropriate. The association between age and sleep quantity was U shaped, particularly in women, with men aged 45–54 and women aged 35–44 reporting the least amount of sleep. There was evidence that education could predict sleep quantity in women (p<0.05), but not men, with those leaving education earlier reporting more sleep. There was no evidence of a linear association between social class or car or van availability and sleep quantity in either men or women.

Table thumbnail
Table 1 Mean weighted hours of sleep in previous 24 hours by markers of socioeconomic position with regression coefficients

Table 22 shows the proportion of men and women reporting mid‐range sleep quantities by age group and markers of SEP along with odds ratios of the chance of reporting mid‐range sleep quantities compared with the most advantaged group. There was a general trend for the likelihood of reporting mid‐range sleep to decrease with age with men and women aged over 75 being least likely to report mid‐range sleep. While the odds of reporting mid‐range sleep showed an overall decrease with decreasing advantage in both men and women for all markers of SEP, this was only significant (p<0.05) in relation to education in women.

Table thumbnail
Table 2 Proportion of people reporting mid‐range sleep (6.5–8.5 hours/24 hours) by markers of socioeconomic position with odds ratios

What is already known about this topic

It has been suggested that sleep deprivation may mediate the association between socioeconomic position and health. The association between sleep quantity and morbidity and mortality seems to be U shaped. Previous work has failed to show a linear association between SEP and sleep quantity but few have investigated the socioeconomic distribution of mid‐range sleep quantity (around 6.5–8.5 hours).

What this paper adds

Socioeconomic deprivation seems to be associated with greater total sleep but less likelihood of mid‐range sleep quantity—particularly in women.

Discussion

Data from the 1999 UK ONS Omnibus Survey show evidence, in women but not men, that total sleep quantity increases with increasing deprivation (as measured by age at completion of education, but not occupational social class or car or van access), but that the chances of sleeping for a mid‐range length of time decreases with increasing deprivation. Similar trends, for mid‐range sleep quantity, were seen in men but these did not reach conventional levels of statistical significant, perhaps because of smaller numbers.

The overall response rate for the May 1999 ONS Omnibus Survey was 54%. Any bias that this apparently low response rate may have introduced will be, at least partially, overcome by the population weights used. Respondents were asked to assign the main activity they were engaged in for every 15‐minute period of a single day to one of 33 pre‐defined time use codes with sleep and rest included in a single code. It is, therefore, possible that not all time that has been included in the calculation of sleep quantity was actually spent asleep with both rest and attempting to sleep included. However, self report may be the only feasible method of assessing sleep quantity at a population level and time use data are likely to be more accurate than simply asking respondents to estimate the average amount of sleep they get per night—as done in previous work.12,13 Self report information from more than one 24 hour period would, however, be likely to increase the accuracy of such data. In addition, no account was taken of sleep quality and it is possible that people increase time spent in sleep, or trying to sleep, to compensate for poor sleep quality.

As factors such as household composition, working conditions, and housing quality have been proposed as possible explanatory variables in a possible relation between SEP and sleep, these variables were not controlled for. Control for such variables in future analyses may help explain the pattern of results reported here.

These results corroborate previous reports that there is little evidence that more deprived people obtain less sleep than more advantaged ones12,13—and that the reverse may be true14—but there is some evidence that mid‐range levels of sleep do seem to be socioeconomically patterned—particularly in women.15 These findings suggest that any hypothesis implicating sleep in socioeconomic inequalities in health should take into account mid‐range, rather than total, sleep quantities. Further work should focus both on healthy sleep quantity as well as sleep quality and make use of longitudinal data to explore the possibility of a causal relation between SEP and sleep.

Policy implications

If sleep quantity is confirmed as a mediator of inequalities in health, interventions to promote and enable healthy sleep patterns, particularly among more deprived people, may help to reduce socioeconomic inequalities in health.

Acknowledgements

This article used data from the ONS Omnibus Survey, May 1999 (UKDA study number 4224). Principal Investigator: Office for National Statistics Social Survey Division. Sponsors: Department of the Environment, Transport and the Regions; Office for National Statistics; University College London Department of Epidemiology and Public Health; Department of Customs and Excise. Distributed by: UK Data Archive, University of Essex, Colchester, 12 February 2001. JA is supported by a Wellcome Trust Value in People award from Newcastle University.

Footnotes

Competing interests: none declared.

Ethics: this analysis of publicly available routine data did not require ethical approval

References

1. Spiegel K, Leproult R, Van Cauter E. Impact of sleep debt on metabolic and endocrine function. Lancet 1999. 3541435–1439.1439. [PubMed]
2. Tochikubo O, Ikeda A, Miyajima E. et al Effects of insufficient sleep on blood pressure monitored by a new multibiomedical recorder. Hypertension 1996. 271318–1324.1324. [PubMed]
3. Drever F, Whitehead M. eds. Health inequalities decennial supplement. London: The Stationery Office, 1997.
4. Van Cauter E, Spiegel K. Sleep as a mediator of the relationship between socioeconomic status and health: a hypothesis. Ann New York Acad Sci 1999. 896254–261.261. [PubMed]
5. Brunner E. Socioeconomic determinants of health: stress and the biology of inequality. BMJ 1997. 3141472.
6. Lowry S. Health and housing. Plymouth: BMJ Books, 1991.
7. Townsend P, Phillimore P, Beattie A. Health and deprivation: inequality and the North. Bristol: Croom Helm, 1988.
8. Kripke D, Garfinkel L, Wingard D. et al Mortality associated with sleep duration and insomnia. Arch Gen Psychiatry 2002. 59131–136.136. [PubMed]
9. Ayas N T, White D P, Manson J E. et al A prospective study of sleep duration and coronary heart disease in women. Arch Intern Med 2003. 163205–209.209. [PubMed]
10. Wingard D, Berkman L. Mortality risk associated with sleeping patterns among adults. Sleep 1983. 6102–107.107. [PubMed]
11. Tamakoshi A, Ohno Y. Self‐reported sleep duration as a predictor of all‐cause mortality: results from the JACC study, Japan. Sleep 2004. 2751–54.54. [PubMed]
12. Guerin N, Reinberg A, Testu F. et al Role of school schedule, age, and parental socioeconomic status on sleep duration and sleepiness of Parisian children. Chronobiol Int 2001. 181005–1017.1017. [PubMed]
13. Moore P, Adler N, Williams D. et al Socioeconomic status and health: the role of sleep. Psychosom Med 2002. 64337–344.344. [PubMed]
14. Biddle J, Hamermesh D. Sleep and the allocation of time. Journal of Political Economy 1990. 98922–943.943.
15. Hale L. Who has time to sleep? J Public Health 2005. 27205–211.211.
16. Valdes A, Andrew T, Gardner J. et al Obesity, cigarette smoking, and telomere length in women. Lancet. 2005;doi: 10, 1016/S0140–6736(05)66630–5.
17. Office for National Statistics Social Survey Division ONS Omnibus Survey, May 1999. Colchester. Essex: UK Data Archive, 2001.
18. Krieger N, Williams D, Moss N. Measuring social class in US public health research: concepts, methodologies and guidelines. Annu Rev Public Health 1997. 18341–378.378. [PubMed]
19. Macintyre S, Ellaway A, Der G. et al Do housing tenure and car access predict health because they are simply markers of income or self esteem? A Scottish study. J Epidemiol Community Health 1998. 52657–664.664. [PMC free article] [PubMed]
20. Adams J, White M, Pearce M. et al Lifecourse measures of socioeconomic position and self reported health at age 50: prospective cohort study. J Epidemiol Community Health 2004. 581028–1029.1029. [PMC free article] [PubMed]
21. Walker B, McConnachie A, Noon J. et al Contribution of parental blood pressures to association between low birth weight and adult high blood pressure: cross sectional study. BMJ 1998. 316834–837.837. [PMC free article] [PubMed]
22. Stocks N, Ryan P, McElroy H. et al Statin prescribing in Australia: socioeconomic and sex differences: a cross‐sectional study. Med J Aust 2004. 180229–231.231. [PubMed]

Articles from Journal of Epidemiology and Community Health are provided here courtesy of BMJ Group