Fatigue in the U.S. Workforce: Prevalence and
Implications for Lost Productive Work Time
Judith A. Ricci, ScD, MS
Elsbeth Chee, ScD
Amy L. Lorandeau, MA
Jan Berger, MD
Learning Objectives
• Recall the prevalence of fatigue in this national cross-sectional telephone
survey of US workers, and how its presence affected workers’ health status
and quality of life.
• Outline the ways in which fatigue interacted with other health disorders to
increase lost productive work time (the sum of self-reported absenteeism
and presenteeism) and its monetary cost.
• List possible mechanisms by which fatigue may increase functional
impairment caused by other adverse health conditions.
Abstract
Objective: The objective of this study was to estimate fatigue prevalence and
associated health-related lost productive time (LPT) in U.S. workers. Methods:
Fatigue prevalence, LPT due to fatigue, and LPT for any health-related reason (in
hours and dollars) were measured in a national cross-sectional telephone survey
of U.S. workers. Results: The 2-week period prevalence of fatigue was 37.9%. Of
workers with fatigue, 65.7% reported health-related LPT compared with 26.4% of
those without fatigue. Workers with fatigue cost employers $136.4 billion annually
in health-related LPT, an excess of $101.0 billion compared with workers without
fatigue. Fatigue frequently co-occurs with other conditions and, when present, is
associated with a threefold increase, on average, in the proportion of workers with
condition-specific LPT. Conclusions: Fatigue is prevalent in the U.S. workforce.
When occurring with other health conditions, it is associated with significantly
more condition-specific LPT. (J Occup Environ Med. 2007;49:1–10)
F atigue is a common symptom withreported prevalence in the populationranging from 7% to approximately45%.1–8 Fatigue is diagnostically
nonspecific and associated with
many health conditions.2,5 Broadly
defined as “a feeling of weariness,
tiredness or lack of energy,”9 fatigue
is best viewed on a continuum.5 At
the milder end, fatigue occurs fre-
quently and generally comprises
acute circumstance-based episodes
that can resolve quickly after inter-
vention such as rest or the improve-
ment of an environmental stressor.
At the more severe end, it is less
prevalent and potentially symptom-
atic of a more chronic and disabling
condition such as major depressive
disorder, fibromyalgia, or chronic fa-
tigue syndrome.
Fatigue impairs work ability. Work-
ers with fatigue are significantly more
likely to miss work and experience
long-term work absence than workers
without fatigue.10 In addition, health
conditions in which fatigue is a pri-
mary symptom such as chronic fatigue
syndrome11 and depressive disorders12
also negatively impact work ability. In
economic terms, the total annual cost
of lost labor force participation result-
ing from unemployment among indi-
viduals with chronic fatigue syndrome
was estimated at $6.8 billion.11 The
total annual cost of lost productive work
time among U.S. workers with depres-
sion was estimated at $31 billion.12
The prevalence of fatigue in U.S.
workers and its relation to productive
work time have not been studied
previously. The Caremark American
Productivity Audit provides data on
a large nationally representative
From Caremark (Dr Ricci, Dr Chee, Ms Lorandeau), Hunt Valley, Maryland; and Caremark (Dr
Berger), Northbrook, Illinois.
The research was financially supported by Caremark.
Address correspondence to: Judith A. Ricci, ScD, MS, Caremark, 11311 McCormick Rd, Suite 230,
Hunt Valley, MD, 21031; E-mail: judi.ricci@caremark.com.
Copyright © 2007 by American College of Occupational and Environmental Medicine
DOI: 10.1097/01.jom.0000249782.60321.2a
CME Available for this Article at ACOEM.org
JOEM • Volume 49, Number 1, January 2007 1

sample of U.S. workers with infor-
mation linking health conditions, in-
cluding fatigue, to health-related lost
productive time (LPT). We describe
the results of research to estimate the
prevalence of fatigue among U.S.
workers, characterize the health sta-
tus and quality of life of workers
with fatigue, and quantify worker
health-related LPT and associated
costs, including both time absent
from work and reduced performance
while at work.
Materials and Methods
The Caremark American Produc-
tivity Audit (or the audit) is a U.S.
national population-based random-
digit-dial telephone survey of the
noninstitutionalized U.S. population
that measures the relation between
health and work productivity.13 The
Caremark Work and Health Inter-
view (or the WHI) is the validated
data collection instrument adminis-
tered in the audit survey. Both the
audit13 and the WHI14,15 have been
described in detail elsewhere and are
summarized briefly here.
Work and Health Interview and
Measurement of Lost
Productive Time
The WHI is a computer-assisted
telephone data collection instrument
that measures LPT and its health-
related causes in the 2 weeks before
interview.14,15 The interview captures
information on self-reported employ-
ment status, occupational characteris-
tics, health conditions and symptoms,
lifestyle factors, health-related quality
of life, and demographic characteris-
tics, including annual salary. LPT is
measured as the sum of self-reported
hours per week absent from work for a
health-related reason (ie, absenteeism)
and the hour-equivalent per week of
self-reported health-related reduced
performance while at work (ie, presen-
teeism). Presenteeism is quantified by
measuring the average frequency of
engaging in five specific work behav-
iors and the average amount of time
between arriving at work and starting
to work on days not feeling well. The
five work behaviors included losing
concentration, repeating a job, work-
ing more slowly than usual, feeling
fatigued at work, and doing nothing at
work. Response options are all of the
time, most of the time, half the time,
some of the time, and none of the time.
The primary health-related reason for
LPT was attributed directly by the
respondent.
Sample Selection and
Data Collection
Audit households were selected as
a random sample of residences with
telephones in the 48 contiguous
states and the District of Columbia.
Residents were eligible to participate
if they were 18 to 65 years of age,
reported in the affirmative to the
Current Population Survey (CPS)
question on employment status (ie,
“Last week, did you do any work for
either pay or profit?”),16 and were a
permanent member of the household
contacted. Audit data collection
began on August 1, 2001, and con-
tinued through May 31, 2003. Qual-
ity-of-life data were collected only
between August 1, 2001, and August
24, 2002. During this period, inter-
views were completed at a rate of
approximately 2500 per month with
an estimated participation rate of
66%.13 Up to two eligible respon-
dents were interviewed per house-
hold. The Essex Institutional Review
Board (Lebanon, NJ) approved the
research protocol and data collection
instrument. Oral informed consent
was obtained from each participant
before initiating the interview.
A two-step weighting method ac-
counted for selective participation (ie,
noncoverage and nonresponse).13
In the first step, a weight was
applied to individuals to account
for the unequal probability of select-
ing households. In the second step, a
population weighting adjustment ac-
counted for selection bias due to
incomplete coverage of the U.S. pop-
ulation and ensured that estimates of
certain sample demographic sub-
groups’ totals conformed to the CPS,
an external database providing high-
quality data on a nationally represen-
tative sample of the U.S. workforce.
A raking method was used for the
population weighting adjustment,
benchmarking to four variables com-
mon to both the audit and the CPS.
Benchmarking and weighting vari-
ables with missing data were im-
puted using a previously described
procedure.13
Analysis
The sampling frame for this study
included the 28,902 adults 18 to 65
years of age who participated in the
first year of the audit survey and
reported working for pay or profit in
the week before interview. Because
no standard definition or assessment
method for fatigue currently exists,17
we used an affirmative response to
the following question: “Did you
have low levels of energy, poor
sleep, or a feeling of fatigue in the
past 2 weeks?” to define the presence
of fatigue in our sample. Applying
this definition, 11,719 workers
screened positive for fatigue in the
previous 2 weeks.
Analyses were completed to esti-
mate the prevalence of fatigue and cost
of fatigue-related LPT in the U.S.
workforce. Prevalence estimates were
derived based on the sampling fraction
of workers and projected to the U.S.
workforce using the previously de-
scribed benchmarking procedure.
Health-related LPT was derived
from the Work and Health Interview
as described previously.13 LPT was
examined in three ways. Initially, we
estimated LPT attributed specifically
to fatigue by workers with fatigue.
Second, we estimated excess health-
related LPT (ie, LPT attributed to
fatigue in addition to other health
conditions) in workers with fatigue
compared with a 1:1 group-matched
(by age and gender) random sample
of workers without fatigue. Excess
LPT was calculated as the difference
in total annual hours of LPT for any
health-related reason between the
two groups. Third, we estimated ex-
2 Fatigue and LPT in the U.S. Workforce • Ricci et al