370 Chronic Fatigue Syndrome
others like them, underrepresented low-income and under-
served minorities shown to manifest higher levels of chronic
illness while also being less likely to have access to the health
care system, and thus less likely to be counted in prevalence
rates derived from treatment sources. These studies also un-
derestimated the prevalence of CFS in the general population
because they depended on diagnoses by health care providers
who discounted the existence of the illness and would thus fail
to diagnose it (Richman, Flaherty, & Rospenda, 1994).
In 1991, Jason and Wagner (1998) directly surveyed
nurses, and 202 nurses (6%) out of a sample of 3,400 indi-
cated that they had experienced debilitating fatigue for six
months or longer. In 1992, these nurses were recontacted and
interviewed by phone. Based on those interviews, a high-risk
group of 82 who either reported having CFS or had most of
the symptoms of CFS was selected. These nurses were
interviewed using a structured psychiatric test, and their med-
ical records and other collected information were evaluated
by a Physician Review Team. Thirty-seven nurses met case
criteria for current CFS, yielding a prevalence rate of 1.088%
(1,088 per 100,000; Jason & Wagner, 1998). Because nurses
are a high-risk population for CFS (e.g., nurses• work
is stressful and schedules often disrupt circadian rhythms),
it may not be possible to generalize these data to the entire
population.
In 1993, Jason and colleagues (1995) screened a random
community sample (n1,031) for the presence of CFS.
Sixty-four percent of the fatigued group indicated that they
had no current medical doctor overseeing their illness. The
individuals who reported having CFS or who had many of the
symptoms of CFS were examined by a physician and inter-
viewed by a psychiatrist to determine whether they could be
diagnosed with CFS. The DePaul University research team
diagnosed two patients with current CFS (0.2% of the sam-
ple, or a CFS prevalence rate of 200 per 100,000), a number
higher than we would have expected, given rates from past
epidemiological studies.
The CDC conducted a community-based survey in San
Francisco (Steele et al., 1998). Using telephone interviewing
between June 1 and December 1, 1994, these investigators
surveyed 8,004 (87%) of 9,155 households, providing data
on 16,970 adult and minor residents. Based on this popula-
tion-based telephone survey, the authors estimated the preva-
lence of CFS-like illness to be between .076 and .233% (76
and 233 per 100,000). Unfortunately, this study involved
only self-reported data. In the absence of medical and psy-
chiatric examinations, it was not possible to estimate the
prevalence of actual CFS cases from this study.
Another prevalence study involved a random sample of
4,000 individuals in a health maintenance organization roster
in the Seattle area (Buchwald et al., 1995). Seventy-seven
percent (n3,066) of the individuals surveyed responded.
Three individuals were determined to have CFS, for a preva-
lence rate of .075% (75 per 100,000). If the rates among non-
respondents and nonparticipants were similar to the rates
among study participants, the prevalence rate would rise to
.267% (267 cases per 100,000). A limitation in this study
is that only individuals with access to health care were
represented.
In Great Britain, Pawlikowska et al. (1994) sought re-
sponses from 15,283 people recorded in the general practice
register as having mild fatigue. Based on this sample,
Wessely et al. (1995) later ascertained the presence of CFS
in 1,199 people who presented to the physician with symp-
toms of infections and 1,167 for other reasons. Using health
and fatigue questionnaires (Wessely et al., 1997), it was de-
termined that 2.6% of the sample had CFS according to the
Fukuda et al. (1994) criteria. These rates are within the
range of prevalence of several mood disorders. (Mood disor-
ders are the most prevalent psychiatric illness after anxiety
disorders: For major depressive episode, the one-month
prevalence is 2.2% and lifetime prevalence is 5.8%; Regier
et al., 1988.) These CFS prevalence rates are considerably
higher than those reported in other recent epidemiological
studies.
In a subsequent study by Euba et al. (1996), those indi-
viduals with a CFS diagnosis in the community sample de-
scribed previously were compared with people diagnosed in
a hospital unit specializing in CFS. Whereas 74% of the
community sample had a psychiatric diagnosis before the
onset of their fatigue, only 21% of the hospital sample had
a previous diagnosis. The community sample had signi“-
cantly worse mental health scores and were more likely to
be impaired in their work. Fifty-nine percent believed that
their illness might be due to psychological or psychosocial
causes (compared with 7% for the hospital sample). Wes-
sely et al. (1997) did indicate that, of the 2.6% with CFS in
the community sample, only 0.5% had no psychological
disorder. In another study of CFS and psychiatric symptoms
(Wessely, Chalder, Hirsch, Wallace, & Wright, 1996), 36 in-
dividuals were diagnosed as having CFS from a cohort of
1,985 primary care patients. Among the CFS subgroup, only
64% had sleep disturbances and 63% had postexertional
malaise. These percentages are rather low, given that both
symptoms are critical features of CFS. Wessely and col-
leagues• “nding of an increased prevalence of CFS re”ects
the use of the new Fukuda et al. (1994) case de“nition and
possibly indicates the inclusion of pure psychiatric cases in
the category of CFS disorders. Thus, using a broad or nar-
row de“nition of CFS has an important in”uence on CFS