DAILY SLEEP, WEEKLY WORKING HOURS, AND RISK OF WORK-RELATED INJURY: US NATIONAL HEALTH INTERVIEW SURVEY (2004-2008)
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Chronobiology International, 27(5): 1013–1030, (2010)
Copyright © Informa UK Ltd.
ISSN 0742-0528 print/1525-6073 online
DOI: 10.3109/07420528.2010.489466
DAILY SLEEP, WEEKLY WORKING HOURS, AND RISK
OF WORK-RELATED INJURY: US NATIONAL HEALTH
INTERVIEW SURVEY (2004–2008)
David A. Lombardi,1 Simon Folkard,2,3 Joanna L. Willetts,1 and
Gordon S. Smith4
Chronobiol Int Downloaded from informahealthcare.com by 143.115.159.53 on 07/19/10
1
Center for Injury Epidemiology, Liberty Mutual Research Institute for Safety, Hopkinton,
Massachusetts, USA
2
Laboratoire d’Anthropolgie Appliquée, Université Paris Descartes, Faculté de Médecine,
Paris, France
3
Body Rhythms and Shiftwork Centre, University of Wales Swansea, Swansea, UK
4
National Study Center Trauma and EMS, University of Maryland, Baltimore,
Maryland, USA
For personal use only.
The impact on health and safety of the combination of chronic sleep deficits and
extended working hours has received worldwide attention. Using the National
Health Interview Survey (NHIS), an in-person household survey using a multistage,
stratified, clustered sample design representing the US civilian, non-institutionalized
population, the authors estimated the effect of total daily self-reported sleep time and
weekly working hours on the risk of a work-related injury. During the survey period
2004–2008, 177,576 persons (ages 18–74) sampled within households reported that
they worked at a paid job the previous week and reported their total weekly work
hours. A randomly selected adult in each household (n = 75,718) was asked to report
his/her usual (average) total daily sleep hours the prior week; complete responses
were obtained for 74,415 (98.3%) workers. Weighted annualized work-related injury
rates were then estimated across a priori defined categories of both average total daily
sleep hours and weekly working hours. To account for the complex sampling design,
weighted multiple logistic regression was used to independently estimate the risk of a
work-related injury for categories of usual daily sleep duration and weekly working
hours, controlling for important covariates and potential confounders of age, sex,
race/ethnicity, education, type of pay, industry, occupation (proxy for job risk), body
mass index, and the interaction between sleep duration and work hours. Based on
the inclusion criteria, there were an estimated 129,950,376 workers annually at risk
and 3,634,446 work-related medically treated injury episodes (overall injury rate
2.80/100 workers). Unadjusted annualized injury rates/100 workers across weekly
work hours were 2.03 (≤20 h), 3.01 (20–30 h), 2.45 (31–40 h), 3.45 (40–50 h), 3.71
This paper was presented at the 19th International Symposium on Shiftwork and Working
Time, August 2–6, 2009, Venice, Italy.
Address correspondence to David A. Lombardi, PhD, Senior Research Scientist, Center for Injury
Epidemiology, Liberty Mutual Research Institute for Safety, 71 Frankland Road, Hopkinton, MA 01748,
USA. Tel.: (508) 497-0210; Fax: (508) 435-3456; E-mail: david.lombardi@libertymutual.com
10131014 D. A. Lombardi et al.
(50–60 h), and 4.34 (>60 h). With regards to self-reported daily sleep time, the esti-
mated annualized injury rates/100 workers were 7.89 (10 h). After
controlling for weekly work hours, and aforementioned covariates, significant
increases in risk/1 h decrease were observed for several sleep categories. Using 7–7.9
h sleep as reference, the adjusted injury risk (odds ratio [OR] for a worker sleeping a
total of 10
h of usual daily sleep, the OR was marginally significantly elevated, 1.82 (95% CI:
0.96–3.47). These results suggest significant increases in work-related injury risk with
decreasing usual daily self-reported sleep hours and increasing weekly work hours,
independent of industry, occupation, type of pay, sex, age, education, and body mass
Chronobiol Int Downloaded from informahealthcare.com by 143.115.159.53 on 07/19/10
(Author correspondence: david.lombardi@libertymutual.com)
Keywords Epidemiology; Fatigue; Injury; Sleep; Work hours
INTRODUCTION
The impact on health, safety, and well-being of sleep deficits and
extended working hours has received increasing worldwide attention
For personal use only.
(Caruso et al., 2006). With respect to sleep, the evidence is growing that
short sleep duration is associated with several chronic disease outcomes,
such as diabetes (Gottleib et al., 2005), hypertension (Gangwisch et al.,
2006), cardiovascular disease (Ayas et al., 2003), obesity (Di Milia &
Mummery, 2009; Marshall et al., 2008; Singh et al., 2005), and sickness
absences from work (Westerlund et al., 2008). In contrast, long sleep dur-
ations have been associated with adverse mental health conditions, such
as depression (Patel et al., 2006). With respect to safety, lack of sleep has
been associated with traffic crashes (Connor et al., 2001; Teran-Santos
et al., 1999), and work-related injuries (Dembe et al., 2005; Folkard &
Lombardi, 2005). Likewise, long working hours and demanding work
schedules are associated with various adverse impacts on worker safety
and health (Biggi et al., 2008; Dembe et al., 2005; Esquirol et al., 2009;
Hanecke et al., 1998; Lin et al., 2009; Suwazono et al., 2009). One study
of a large representative sample of US workers found that working >60
h/wk increased injury hazard rates by 23% (Dembe et al., 2005), and
another large German study reported an exponential increase in work
“accident” risk observed beyond the 9th h of work (Hanecke et al., 1998).
Injury risks associated with long work hour schedules and overtime
have been shown to vary widely across a range of occupations in the USA
(Dembe et al., 2008). This effect has been demonstrated across multiple
industries; among manufacturing workers, for example, those who worked
>64 h the week before the shift in which they were injured had an 88%
excess risk compared to those who workedSleep, Working Hours and Risk of Injury 1015
workers, long work hours and irregular work schedules were significantly
associated with a higher work-related injury rate after controlling for
potential confounders (Dong, 2005). With respect to health outcomes, long
working hours has also been associated with hypertension, cardiovascular
disease, diabetes, and sleep disorders (Härmä, 2003, 2006). However, only a
limited number of published studies that have examined the association
among long work hours, sleep duration, and injury risk have controlled for
important potential confounders, e.g., occupation, industry, and body mass
index (BMI). Many have examined only small study populations making
them vulnerable to the healthy worker effect and limited external validity
(Folkard & Lombardi, 2004). Despite these limitations, there appears to be a
Chronobiol Int Downloaded from informahealthcare.com by 143.115.159.53 on 07/19/10
clear relationship between work hours and sleep duration, with an increase
in work hours being associated with both a reduction in the opportunity to
sleep and sleep quality. In one study involving 5720 Stockholm workers,
high work demands and physical effort at work were found to be risk indi-
cators for disturbed sleep (Åkerstedt et al., 2002), and another study of 367
forestry workers reported that long working hours led to reduced sleep,
compromised recovery time, and potentially reduced work safety (Lilley
et al., 2002). More recently, a study by Virtanen et al. (2009) reported that
working >55 h as compared to 35–40 h/wk significantly increased by 2-fold
For personal use only.
the risk of shortened sleeping hours (odds ratio [OR] = 1.98, 95% confi-
dence interval [CI] = 1.04, 3.77), after adjusting for demographic factors.
With respect to other potential confounders in the relationship
between work hours, sleep duration, and injury risk, a study of the US
population using the National Health Interview Survey (NHIS) data
showed a descriptive association between usual sleep duration and obesity
among adults (Schoenborn & Adams, 2008). The findings of this study
suggested that those sleeping 40 h/wk rose from 26% to 31%.
At the same time, data from the US Centers for Disease Control and
Prevention (CDC) indicates that the percent of males and females reporting
an average of1016 D. A. Lombardi et al.
Sweden (Westerlund et al., 2008). We used the National Health Interview
Survey (NHIS) and pooled across a large population of workers from the 5
most recent years of available data (2004–2008) to estimate the independent
effect of usual daily sleep duration and weekly working hours on the risk of
a work-related injury in the USA, controlling for important factors such as
age, sex, race/ethnicity, education, type of pay, industry, occupation (as a
proxy for job risk), body mass index (BMI), and the interaction among
usual sleep duration and work hours.
MATERIALS AND METHODS
Chronobiol Int Downloaded from informahealthcare.com by 143.115.159.53 on 07/19/10
Study Population
The data analyzed in this study are from the NHIS, which is an in-
person household survey with a multistage, stratified, clustered sample
design that provides estimates on health indicators, health care utilization
and access, and health-related behaviors of the civilian, non-institutiona-
lized population residing in the 50 states and the District of Columbia in
the USA. The NHIS has been a continuous nationwide sample survey
since 1957 conducted by the National Center for Health Statistics
For personal use only.
(NCHS) to monitor the health of the population of the USA on a broad
range of health topics (NCHS-NHIS, 2008). The current study was
approved by the Liberty Mutual Research Institute for Safety’s
Institutional Review Board and complies with the ethical requirements of
the journal (Portaluppi et al., 2008).
Pooling
The data for this study were pooled across the 5 most recent years of
the NHIS survey (2004–2008). In general, the survey questions and
sampling design for these years is nearly identical. However, starting in
2006, the size of NHIS was reduced by about 13% due to budget con-
straints (Chen et al., 2009), and a further cut of another 13% was made in
2008 (NCHS-NHIS, 2008). NHIS data are obtained through a complex
sample design that allows for valid population estimates, assuming appro-
priate adjustments are made to the statistical analysis that take into con-
sideration the stratification, clustering, and weighting in the study design
(NCHS-NHIS, 2008). Pooling NHIS data across years for estimation is con-
sidered reliable when the proper adjustments to sampling weights are
made. Sample weights were assigned by the NCHS for each respondent on
the basis of the number and composition of households and included
adjustment for non-response (Chen et al., 2009; NCHS-NHIS, 2008).
Figure 1 describes the selection criteria and sample sizes for the pooled
data used in the current study. In summary, for the years 2004–2008, theSleep, Working Hours and Risk of Injury 1017
Chronobiol Int Downloaded from informahealthcare.com by 143.115.159.53 on 07/19/10
For personal use only.
FIGURE 1 National Health Interview Survey (2004–2008).
NHIS data includes 417,390 persons with complete sampling design infor-
mation reported in the family survey by an adult in the household. This
family survey includes injury information on 281,451 adults aged 18–74
yrs, of which 177,576 (63%) were reported to be working the previous
week and average weekly working hours was available. Within each
sampled household, one adult was then randomly chosen for a more
detailed interview. Among these adults, 75,718 (93%) provided complete
data on their usual sleep duration, and 74,415 provided complete data on
both sleep and work-hour information. The primary analysis of the
current study includes 69,248 subjects (91.5% of the subsample) who
reported working the previous week, their weekly work hours, usual daily
sleep duration, and complete covariate information.1018 D. A. Lombardi et al.
Injury Data
As with the sleep duration and work hours data, injury data are ascer-
tained each year (2004–2008) in the administration of the Family
Questionnaire, which captures any injury or poisoning episode that a
member of the household may have experienced in the 3 months prior
to the interview that required medical attention (Chen et al., 2009; Smith
et. al., 2006; Warner et al., 2000). This includes a phone call or visit to a
doctor or a visit to a hospital, clinic, or the emergency room (outpatient).
Multiple injury reports are captured in that each family member is
allowed to report more than one injury or poisoning episode. Verbatim
text information on each injury includes the body part and how the
Chronobiol Int Downloaded from informahealthcare.com by 143.115.159.53 on 07/19/10
injury happened, which are recorded for further coding at NCHS. Only
injury descriptions which can be coded to the International Classification
of Diseases, Clinical Modification (ICD)-9CM nature of injury codes
800–999 are included in the injury file (includes acute musculoskeletal
conditions but not gradual onset conditions such as tendonitis) (NCHS,
1997; Smith et al., 2006). Although each person is allowed up to 10 sep-
arate injury episodes, for the purposes of the current analysis, a person
was considered injured if they had at least one injury episode reported.
For personal use only.
An injury was further classified as work-related based upon the response
to the question, “What were you doing when the injury happened?” in
which the response was “Working at a paid job.”
Work Hours
An employed person was defined as a subject who responded that
they were “employed in a job or business during the week” before the
interview. This includes those who were engaged in unpaid work in the
family business or farm, excluding personal work in home. Work hours
were estimated as average weekly working hours for each “employed
person” reported by the randomly selected adult respondent of each
household. A priori categories of weekly working hours were defined as:
≤20, 21–30, 31–40, 41–50, 51–60, and >60 h.
Sleep Duration
Sleep duration was introduced into the NHIS in the Sample Adult
Core in 2004 using the question “On average, how many hours of sleep
do you get in a 24-hour period?” The large sample size in this study
allowed us to examine the effects of sleep in workers across a range of
seven a priori categories defined as: ≤4.99, 5–5.99, 6–6.99, 7–7.99,
8–8.99, 9–9.99, and >10 h.Sleep, Working Hours and Risk of Injury 1019
Covariates and Potential Confounders
Our analyses controlled for sociodemographic, physical, and job
factors (Dembe et al., 2008; Schoenborn & Adams, 2008; Vgontzas et al.,
2008). Sociodemographic factors included age, sex, education, and race/
ethnicity (Bohle et al., 2008; Di Milia et al., in press; Folkard et al., 2008;
Gander & Signal, 2008). Education was defined as the highest level of
education achieved. Race/ethnicity was categorized as Non-Hispanic
White, Hispanic, Non-Hispanic Black, and Non-Hispanic Other. For
physical factors, BMI was obtained from the Sample Adult Core and
calculated from self-reported height and weight. Job-related factors,
obtained from the Sample Adult Core, included industry, occupation,
Chronobiol Int Downloaded from informahealthcare.com by 143.115.159.53 on 07/19/10
and type of pay (in response to the question “Were you paid hourly?”).
Industry and occupation were coded by the NHIS using the North
American Industrial Classification System (NAICS) and Standard
Occupation Classification (SOC), respectively, based on a verbatim
description of the person’s industry and occupation. We also controlled
for the interaction between a priori work-hour and sleep-hour categories.
For personal use only.
Data Analysis
To analyze the complex survey design data in the NHIS, SAS version
9.1 (SAS Inc., Cary, NC), procedures for survey data were used to incor-
porate the sampling design attributes and weighting into the analyses. For
descriptive analyses, SurveyMeans and SurveyFreq with corrected standard
errors for mean estimation, 95% confidence intervals, weighted frequency,
and percents were calculated. Weighted annualized work-related injury
rates were estimated across a priori defined categories of total daily sleep
hours and weekly working hours also using SAS Proc SurveyFreq.
Multiple logistic regression using the SAS procedure Proc Survey
Logistic was used to estimate the risk of a work-related injury (as a binomial
variable) as a function of daily sleep time, controlling for weekly working
hours, age, sex, race/ethnicity, education, type of pay, industry and occu-
pation, BMI, and interactions. Parameter and odds ratio (OR) estimates
were computed after adjustment for stratification, clustering, and unequal
weighting using the method of maximum likelihood. Taylor series approxi-
mations were used to compute the corrected standard errors and confidence
intervals of the OR (Binder, 1983). An alpha level of .05 was use for statistical
significance for parameter based on the Wald chi-square; however, sociode-
mographic variables were kept in the model for adjustment, regardless of
their statistical significance. To examine the interaction between total daily
sleep hours and weekly working hours, a hierarchically well-formulated
(HWF) logistic regression model (Kleinbaum, 1992) approach, including
both lower order terms of the two-way higher order interaction, was used.1020 D. A. Lombardi et al.
RESULTS
Subject Characteristics
Among subjects interviewed over the 5-yr period who reported both
working the previous week and average weekly work hours (n = 177,576),
54% were male and the mean (±SEM) age of the subjects was 40.6 (±0.03)
yrs (Table 1). The majority of subjects were White, non-Hispanic (69.5%),
followed by Hispanic (14.0%), and Black, non-Hispanic (11.2%).
Table 2 presents the top 10 industries and occupational categories.
Industry employment was primarily distributed among Health Care and
Social Assistance (13.1%), Manufacturing (11.1%), Retail Trade (10.5%),
Chronobiol Int Downloaded from informahealthcare.com by 143.115.159.53 on 07/19/10
TABLE 1 Characteristics of National Health Interview Survey (NHIS): all subjects (2004–2008)
Number of subjects Weighted frequency Weighted
Characteristic interviewed (000s) percent
Total 177,576 649,752 100.0
Sex
Male 94,472 350,902 54.0
For personal use only.
Female 83,104 298,850 46.0
Age (yrs)
18–24 22,321 84,669 13.0
25–34 39,944 145,705 22.4
35–44 44,850 158,742 24.4
45–54 42,809 156,371 24.1
55–64 22,952 86,396 13.3
65 + 4,700 17,869 2.8
Mean ± SEM 40.6 ± 0.03
Combined Ethnicity
Non-Hispanic White 103,623 451,322 69.5
Hispanic 39,571 91,245 14.0
Non-Hispanic Black 23,620 72,946 11.2
Non-Hispanic Other 10,762 34,239 5.3
Education
Did not complete High School 24,874 71,854 11.2
High School Graduate 49,687 180,545 28.2
Some College 51,151 192,979 30.2
Bachelor’s Degree 32,489 129,438 20.2
Master’s Degree 11,682 46,634 7.3
Professional Degree 4,463 17,959 2.8
Missing 3,230
Body Mass Index (BMI)a
Underweight (30) 18,527 67,843 24.9
Mean ± SEM 27.3 ± 0.02
a
BMI was asked only on the Sample Adult Core questionnaire.Sleep, Working Hours and Risk of Injury 1021
TABLE 2 Top 10 occupations and industries of study subjectsa
Number of subjects Weighted Weighted
Characteristic interviewed frequency (000s) percent
Total 74,415 278,650 100.0
Industry
Health Care and Social Assistance 9,641 35,458 13.1
Manufacturing 8,037 29,957 11.1
Retail Trade 7,441 28,255 10.5
Education Services 6,608 25,374 9.4
Construction 5,404 19,914 7.4
Accommodation and Food Services 4,487 16,065 6.0
Professional, Scientific, and Tech Services 4,443 18,177 6.7
Chronobiol Int Downloaded from informahealthcare.com by 143.115.159.53 on 07/19/10
Public Administration 3,746 14,214 5.3
Other Services (except Public Administration) 3,662 13,239 4.9
Finance and Insurance 3,496 13,733 5.1
Otherb 14,885 55,366 20.5
Occupation
Office and Administrative Support 10,238 38,294 14.2
Sales and Related 7,459 29,004 10.8
Management 6,328 25,286 9.4
Production 5,162 18,180 6.7
Construction and Extraction 4,376 15,572 5.8
For personal use only.
Education, Training, and Library 4,303 16,832 6.2
Transportation and Material Moving 4,148 14,987 5.6
Healthcare Practitioners and Technical 3,734 14,606 5.4
Food Preparation and Serving Related 3,685 13,230 4.9
Building and Grounds Cleaning and Maintenance 3,234 10,341 3.8
Otherc 19,192 73,439 27.2
Paid Hourly
No 30,595 119,341 42.5
Yes 44,349 161,428 57.5
Missing 552
a
Persons aged 18–24 yrs reporting working the previous week, weekly work hours, usual daily and
sleep duration.
b
Other industries include Administrative and Support and Waste Management/Remediation
Services; Transportation and Warehousing; Wholesale Trade; Information; Real Estate and Rental/
Leasing; Arts, Entertainment, and Recreation; Agriculture, Forestry, Fishing, and Hunting; Utilities;
Mining; Armed Forces; Management of Companies/Enterprises.
c
Other occupations include Business and Financial Operations; Installation, Maintenance, and Repair;
Personal Care and Service; Computer and Mathematical; Healthcare Support; Protective Service; Arts,
Design, Entertainment, Sports, and Media; Architecture and Engineering; Community and Social
Services; Legal; Life, Physical, and Social Science; Farming, Fishing, and Forestry; Military Specific.
Education Services (9.4%), and Construction (7.4%); however, most indus-
tries were represented in the NHIS data. The most frequent occupations
were Office and Administrative Support (14.2%), Sales and Related
(10.8%), Management (9.4%), Production (6.7%), and Construction and
Extraction (5.8%).
Most injuries were due to overexertion or strenuous movements
(24.5%), falls (23.3%), being cut or pierced (15.5%), and being struck by1022 D. A. Lombardi et al.
Chronobiol Int Downloaded from informahealthcare.com by 143.115.159.53 on 07/19/10
For personal use only.
FIGURE 2 Percent of subjects injured by cause (ICD-9-CM E-codes).
object or person (13.0%) (Figure 2). Machinery and transportation
injuries accounted for 3.4% and 2.3% of the injuries, respectively.
Weekly Working Hours and Injuries
Over the 5-yr study period, there were an estimated 3,634,446 work-
related injury episodes, among the approximately 130 million estimated
workers at risk annually (incidence = 2.80/100 workers) (Table 3a).
Unadjusted annualized injury rates generally increased across the seven
categories of weekly work hours: 2.03 (≤20 h/wk), 3.01 (20–30 h), 2.45
(31–40 h), 3.45 (40–50 h), 3.71 (50–60 h), and 4.34 (>60 h).
Usual Sleep Duration and Injuries
The unadjusted estimated annualized injury rates/100 workers gener-
ally increased with decreased duration of usual daily sleep (Table 3b): 7.89
(10 h). Thus, the crude (or unadjusted)Sleep, Working Hours and Risk of Injury 1023
TABLE 3a Estimated annual injury incidence/100 workers by weekly work hours
Average Number of Est. number of
Hours number workers work-related Est. number Est. annual
worked/ of hours interviewed injuries of workers incidence/
week worked 2004–08a annuallyb annually 100 workers
Total 40.3 177,576 3,634,446 129,950,376 2.80
≤20 h 14.9 14,785 229,343 11,286,527 2.03
21–30 h 27.2 13,333 298,900 9,929,180 3.01
31–40 h 39.3 101,442 1,746,467 71,388,048 2.45
41–50 h 47.4 28,396 761,163 22,042,456 3.45
51–60 h 57.9 13,448 387,346 10,433,206 3.71
>60 h 72.1 6,172 211,227 4,870,959 4.34
Chronobiol Int Downloaded from informahealthcare.com by 143.115.159.53 on 07/19/10
a
Includes individuals reporting employment at a job or business and who had reported weekly
work hours.
b
A work-related injury was reported by 1,178 (0.70%) workers interviewed.
TABLE 3b Estimated annual injury incidence/100 workers by usual daily sleep duration
Average Number of Est. number of
number of workers with work-related Est. number Est. annual
Usual hours hours slept sleep hours injuries of workers at incidence/
For personal use only.
slept/day (± SE) 2004–2008a annuallyb risk annually 100 workers
Total 7.00 ± 0.004 75,271 1,669,142 56,349,239 2.96
≤4.99 h 3.80 ± 0.010 1,431 83,730 1,061,759 7.89
5–5.99 h 5.00 4,407 172,462 3,310,950 5.21
6–6.99 h 6.00 17,251 469,756 12,988,074 3.62
7–7.99 h 7.00 25,950 448,458 19,774,245 2.27
8–8.99 h 8.00 22,604 413,942 16,571,317 2.50
9–9.99 h 9.00 2,361 39,101 1,759,342 2.22
>10 h 10.50 ± 0.030 1,267 41,694 883,551 4.72
a
Includes individuals reporting employment at a job or business and who had been asked and
reported usual sleep hours.
b
A work-related injury was reported by 544 (0.74%) workers interviewed.
rate-ratio comparing those sleeping a total of 10 h of usual daily sleep
had a marginally elevated OR. The adjusted injury risk OR for a worker1024 D. A. Lombardi et al.
TABLE 4 Multivariate logistic regression model, adjusted∗ odds ratio estimates
Standard Est. odds
Parameter Estimate error p value ratio
Intercept −5.3859 0.4737Sleep, Working Hours and Risk of Injury 1025
Chronobiol Int Downloaded from informahealthcare.com by 143.115.159.53 on 07/19/10
FIGURE 3 Multivariate logistic regression adjusted odds ratio estimates for usual sleep duration.
For personal use only.
FIGURE 4 Multivariate logistic regression adjusted odds ratio estimates for weekly work hours.
further examination (Figure 4) suggests that there is generally a 2-fold
risk across work hour categories in comparison to working ≤20 h/wk,
except for the 30–40 h/wk category, where the OR = 1.54 or a 54%
increase in work-related injury risk is observed. These can also be seen in1026 D. A. Lombardi et al.
Table 3a, in examining the unadjusted annualized injury rates/100
workers across categories of weekly work hours: 2.03 (≤20 h/wk),
3.01 (20–30 h), 2.45 (31–40 h), 3.45 (40–50 h), 3.71 (50–60 h), and 4.34
(>60 h).
DISCUSSION
This study of the adult population of the USA pooled data across
the 5 most recent years of the NHIS survey (2004–2008) and found that
both sleep duration and weekly working hours are significantly and
independently (i.e., there was no significant interaction between these
Chronobiol Int Downloaded from informahealthcare.com by 143.115.159.53 on 07/19/10
factors) associated with the risk of a work-related injury. Decreases in
usual sleep duration and increases in weekly working hours both
increased injury risk, even after controlling for several important socio-
demographic, job-related, and physical factors. Thus, reduced sleep
increased injury risk, regardless of the number of hours worked/week
and, conversely, hours/week worked increased injury risk, regardless of
usual daily sleep duration. If the potential impact of these risk factors is
considered simultaneously, the independent additive risk of a work-
For personal use only.
related injury could be substantial. For example, comparing a person
working >60 h/wk and a usual daily sleep duration ofSleep, Working Hours and Risk of Injury 1027
response speed typically being reduced and the frequency of errors and
omissions increased when individuals are fatigued (Williamson et al.,
in press).
Strengths and Limitations
An important strength of the current study is its strong external val-
idity due to the use of pooled NHIS data, which are representative of the
adult population of the USA. Pooling of 5 yrs of NHIS data also provided
a large sample size, providing strong statistical power to both evaluate our
primary hypotheses and to effectively control for several important
Chronobiol Int Downloaded from informahealthcare.com by 143.115.159.53 on 07/19/10
potential confounders. The use of NHIS data, and the ability to identify
each injury for work-relatedness, overcomes many of the shortcomings
inherent in other large national data sources (Smith et al., 2006). Our
study used a 3-month recall period for self-reported medically treated
injuries, which may underreport some injuries (Warner et al., 2005).
However, although injury recall improves with increasing severity, it is
unlikely to result in differential underestimation of injuries (Harel et al.,
1994; Smith et al., 2006). Self-reported sleep duration has been reported
For personal use only.
in some studies to systematically overestimate more objectively measured
sleep time (Lauderdale et al., 2008; Lockley et al., 1999), and daily varia-
bility in sleep time makes it difficult to integrate over long periods. In
addition, usual sleep patterns may not be representative of the sleep dur-
ation or sleep quality at the time of a reported injury; however, we have
no information in the current study on the timing of the sleep episode(s),
which might be important with respect to injury risk. With respect to
weekly work hours, a weekly average may not reflect the highest risk
periods for an injury, e.g., time on task, and in the current cross-sectional
study, we have no estimate of exposed and nonexposed person’s time to
other potential work hazards. Another potentially important limitation of
this study is that the NHIS does not include information on the particu-
lar work shift of each individual, e.g., permanent night versus rotating
shiftworkers; thus, we cannot evaluate whether shiftwork mediates the
relationship between sleep duration and injury risk. However, the
Sample Adult Core of the NHIS collects individual-level data on industry,
occupation, and type of pay, which were controlled for in the logistic
regression model. Although imperfect, these work-related variables may
provide limited control for shiftwork as a variable, as they are likely
correlated with typical worker exposure to shiftwork (e.g., construction
workers), however not for the type of work shift. Also, BMI was calculated
from self-reported height and weight, there may be some inaccuracy in
our measure of body mass. However, it is unlikely that height and weight
would be differentially reported by injured and uninjured persons
(Breslow & Smothers, 2005).1028 D. A. Lombardi et al.
Conclusions
Workplace health and safety can be improved by reducing worker
fatigue and risk. In the current study both sleep duration and weekly
working hours were independently associated with the risk of a work-
related injury. However, when considering the impact of various work
schedules on injury risk, it is important to take into account the com-
ponents of weekly work hours, such as the length of the individual shifts,
the type of shift, breaks, and number of successive shifts worked (Folkard
& Lombardi, 2005), and importantly, the opportunity for sufficient sleep
duration needed for the adequate rest and recovery of the worker.
Chronobiol Int Downloaded from informahealthcare.com by 143.115.159.53 on 07/19/10
ACKNOWLEDGMENTS
This work was supported by the Liberty Mutual Research Institute
for Safety. The authors are grateful to Dr. Mary F. Lesch and Dr. Yueng-
hsiang (Emily) Huang for their thoughtful comments on this manuscript.
Declaration of Interest: The authors report no conflicts of interest.
For personal use only.
The authors alone are responsible for the content and writing of the
paper.
REFERENCES
Åkerstedt T, Knutsson A, Westerholm P, Theorell T, Alfredsson L, Kecklund G. (2002). Sleep disturb-
ances, work stress and work hours: a cross-sectional study. J. Psychosom. Res. 53:741–748.
Ayas NT, White DP, Manson JE, Stampfer MJ, Speizer FE, Malhotra A, Hu FB. (2003). A prospective
study of sleep duration and coronary heart disease in women. Arch. Intern. Med. 163:205–209.
Biggi N, Consonni D, Galluzzo V, Sogliani M, Cosat G. (2008). Metabolic syndrome in permanent
night workers. Chronobiol. Int. 25:443–454.
Binder D. (1983). On the variance of asymptotically normal estimators from complex surveys. Int.
Stat. Rev. 51:279–292.
Bohle P, Di Milia L, Fletcher A, Rajaratnam S. (2008). Introduction: aging and the multifaceted influ-
ences on adaptation to working time. Chronobiol. Int. 25:155–164.
Botman SL, Moore TF, Moriarity CL, Parsons VL. (2000). Design and estimation for the National
Health Interview Survey, 1995–2004. National Center for Health Statistics. Vital Health Stat. 2.
130:1–31.
Breslow RA, Smothers BA. (2005). Drinking patterns and body mass index in never smokers:
National Health Interview Survey, 1997–2001. Am. J. Epidemiol. 161:368–376.
Caruso CC, Bushnell T, Eggerth D, Heitmann A, Kojola B, Newman K, Rosa R, Sauter SL, Vila B.
(2006). Long working hours, safety, and health: toward a National Research Agenda. Am. J. Ind.
Med. 49:930–942.
Center for Disease Control. (2008). QuickStats: percentage of adults aged >18 years who reported an
average ofSleep, Working Hours and Risk of Injury 1029
Connor J, Whitlock G, Norton R, Jackson R. (2001). The role of driver sleepiness in car crashes: a
systematic review of epidemiological studies. Accid. Anal. Prev. 33:31–41.
Dembe AE, Erickson JB, Delbos R, Banks S. (2005). The impact of overtime and long work hours on
occupational injuries and illnesses: new evidence from the United States. Occup. Environ. Med.
62:588–597.
Dembe AE, Delbos R, Erickson JB. (2008). The effect of occupation and industry on the injury risks
from demanding work schedules. J. Occup. Environ. Med. 50:1185–1194.
Di Milia L, Mummery L. (2009). The association between job related factors, short sleep and obesity.
Ind. Health 47:363–368.
Di Milia, Smolensky MH, Holwarth H, Costa G, Ohayon M, Philip P. (2010, in press). Demographic
factors, fatigue and driving accidents: an examination of the published literature. Accid. Anal.
Prev. doi:1016/j.aap.2009.12.018.
Esquirol Y, Bongard V, Mabile L, Jonnier B, Soulat J-M, Perret B. (2009). Shift work and metabolic
syndrome: respective impacts of job strain, physical activity and dietary rhythms. Chronobiol. Int.
Chronobiol Int Downloaded from informahealthcare.com by 143.115.159.53 on 07/19/10
26:544–559.
Dong X. (2005). Long working hours, work scheduling, and work-related injuries in construction.
Scand. J. Work Environ. Health 31:329–335.
Folkard S. (2008). Shift work, safety, and aging. Chronobiol. Int. 25:183–198.
Folkard S, Lombardi DA. (2006). Modeling the impact of the components of long work hours on
injuries and “accidents.” Am. J. Ind. Med. 49:953–963.
Folkard S, Lombardi DA, Tucker PT. (2005). Shiftwork: safety, sleepiness and sleep. Ind. Health
43:20–23.
Gander P, Signal L. (2008). Who is too old for shift work? Developing better criteria. Chronobiol. Int.
25:199–213.
Gangwisch JE, Heymsfield SB, Boden-Albala B, Buijs RM, Kreier F, Pickering TG, Rundle AG,
For personal use only.
Zammit GK, Malaspina D. (2006). Short sleep duration as a risk factor for hypertension: analyses
of the first National Health and Nutrition Examination Survey. Hypertension 47:833–839.
Gottlieb DJ, Punjabi NM, Newman AB, Resnick HE, Redline S, Baldwin CM, Nieto FJ. (2005).
Association of sleep time with diabetes mellitus and impaired glucose tolerance. Arch. Intern.
Med. 165:863–867.
Hanecke K, Tiedemann S, Nachreiner F, Grzech-Sukalo H. (1998). Accident risk as a function of
hour at work and time of day as determined from accident data and exposure models for the
German working population. Scand. J. Work Environ. Health 24(Suppl. 3):43–48.
Härmä M. (2003). Are long work hours a health risk? Scand. J. Work Environ. Health 29:167–169.
Härmä M. (2006). Workhours in relation to work stress, recovery and health. Scand. J. Work Environ.
Health. 32:502–514.
Harel Y, Overpeck MD, Jones DH, Scheidt PC, Bijur PE, Trumble AC, Anderson J. (1994). The
effects of recall on estimating annual nonfatal injury rates for children and adolescents.
Am. J. Public Health 84:599–605.
Kleinbaum DG. (1994). Logistic regression: a self-learning text. New York: Springer Verlag.
Knauth P. (2007). Extended work periods. Ind. Health 45:125–136.
Lilley R, Feyer AM, Kirk P, Gander P. (2002). A survey of forest workers in New Zealand. Do hours
of work, rest, and recovery play a role in accidents and injury? J. Safety Res. 33:53–71.
Lin Y-C, Hsiao T-J, Chen P-C. (2009). Persistent shift-work exposure accelerates development of
metabolic syndrome among middle-aged female employees: a five-year follow-up. Chronobiol. Int.
26:740–755.
Lockley SW, Skene DJ, Arendt J. (1999). Comparison between subjective and actigraphic measure-
ment of sleep and sleep rhythms. J. Sleep Res. 8:175–183.
Marshall NS, Glozier N, Grunstein RR. (2008). Is sleep duration related to obesity? A critical review
of the epidemiological evidence. Sleep Med. Rev. 12:289–298.
National Center for Health Statistics. (1997). Official ICD-9-CM guidelines for coding and reporting.
Hyattsville, MD: National Center for Health Statistics.
National Center for Health Statistics. (2008). Provisional report. Summary health statistics for U.S.
adults: National Health Interview Survey. Vital Health Stat. Series 10 (242):2–3.
National Center for Health Statistics. Data file documentation, National Health Interview Survey
(NHIS), 2004, 2005, 2006, 2007, 2008. Hyattsville, MD: National Center for Health Statistics,1030 D. A. Lombardi et al.
Centers for Disease Control and Prevention. Available at: http://www.cdc.gov/nchs/nhis.htm.
Accessed September 10, 2009.
Patel SR, Malhotra A, Gottlieb DJ, White DP, Hu FB. (2006). Correlates of long sleep duration. Sleep
29:881–889.
Pleis JR, Lucas JW. (2009). Summary health statistics for U.S. adults: National Health Interview
Survey, 2007. National Center for Health Statistics. Vital Health Stat. Series 10 (240):1–159.
Portaluppi F, Touitou Y, Smolensky MH. (2008). Ethical and methodological standards for laboratory
and medical biological rhythm research. Chronobiol. Int. 25:999–1016.
Roberts G., Rao NK, Kumar S. (1987). Logistic regression analysis of sample survey data. Biometrika
74:1–12.
Singh M, Drake CL, Roehrs T, Hudgel DW, Roth T. (2005). The association between obesity and
short sleep duration: a population-based study. J. Clin. Sleep Med. 1:357–363.
SAS Institute Inc. (2000–2004). SAS 9.1.3 help and documentation. Cary, NC: SAS Institute Inc.
Schoenborn CA, Adams PF. (2008). Sleep duration as a correlate of smoking, alcohol use, leisure-time
Chronobiol Int Downloaded from informahealthcare.com by 143.115.159.53 on 07/19/10
physical inactivity, and obesity among adults: United States, 2004–2006. Hyattsville, MD:
National Center for Health Statistics, Centers for Disease Control and Prevention. Retrieved
11 January, 2010 from: http://www.cdc.gov/nchs/products/pubs/pubd/hestats/sleep04-06/
sleep04-06.pdf.
Skinner CJ, Holt D, Smith TMF. (1989). Analysis of complex surveys. New York: John Wiley & Sons,
Inc.
Smith GS, Sorock GS, Wellman HM, Courtney TK, Pransky GS. (2006). Blurring the distinctions
between on and off the job injuries: similarities and differences in circumstances. Inj. Prev.
12:236–241.
Suwazono Y, Dochi M, Oishi M, Tanaka K, Kobayashi E, Sakata K. (2009). Shiftwork and impaired
glucose metabolism: a 14-year cohort study on 7104 male workers. Chronobiol. Int. 26:926–941.
For personal use only.
Teran-Santos J, Jimenez-Gomez A, Cordero-Guevara J. (1999). The association between sleep apnea
and the risk of traffic accidents. N. Engl. J. Med. 340:847–851.
Vegso S, Cantley L, Slade M, Taiwo O, Sircar K, Rabinowitz P, Fiellin M, Russi MB, Cullen MR.
(2007). Extended work hours and risk of acute occupational injury: a case-crossover study of
workers in manufacturing. Am. J. Ind. Med. 50:597–603.
Vgontzas AN, Lin H-M, Papaliaga M, Calhoun S, Vela-Bueno A, Chrousos GP, Bixler EO. (2008).
Short sleep duration and obesity: the role of emotional stress and sleep disturbances. Int. J. Obes.
32:801–809.
Virtanen M, Ferrie JE, Vahtera J, Elovainio M, Singh-Manoux A, Marmot MG, Kivimäki M. (2009).
Long working hours and sleep disturbances: the Whitehall II prospective cohort study. Sleep.
32:737–745.
Warner M, Barnes PM, Fingerhut LA. (2000). Injury and poisoning episodes and conditions:
National Health Interview Survey, 1997. Vital Health Stat. Series 10 (202):1–38.
Warner M, Schenker N, Heinen MA, Fingerhut LA. (2005). The effects of recall on reporting injury
and poisoning episodes in the National Health Interview Survey. Inj. Prev. 11:282–287.
Williamson A, Lombardi DA, Folkard S, Stutts J, Courtney TK, Connor JL. (in press). The link
between fatigue and safety. Accid. Anal. Prev.You can also read
