Case Definition 367these four minor symptoms also are de“ning criteria for
CFS. Is it possible that some patients with a primary affective
disorder could be misdiagnosed as having CFS? Some CFS
investigators would not see this as a problem because they
believe that high rates of psychiatric comorbidity indicate
that CFS is mainly a psychiatric disorder and that distinctions
between the two phenomena are super“cial and merely a
matter of nomenclature.
However, several CFS symptoms, including prolonged
fatigue after physical exertion, night sweats, sore throats, and
swollen lymph nodes are not commonly found in depression.
In addition, while fatigue is the principal feature of CFS, fa-
tigue does not assume equal prominence in depression
(Friedberg & Jason, 1998; Komaroff et al., 1996). Also,
illness onset with CFS is often sudden, occurring over a few
hours or days, whereas primary depression generally shows a
more gradual onset. In summary, CFS and depression are two
distinct disorders, although they share a number of common
symptoms. Most importantly, the erroneous inclusion of peo-
ple with primary psychiatric conditions in CFS samples will
have detrimental consequences for the interpretation of both
epidemiologic and treatment ef“cacy “ndings.
The reliability of current CFS criteria (Fukuda et al.,
1994) needs to be improved. To accomplish this important
task, it is relevant to examine the signi“cant improvements
made in the reliability of clinical diagnoses in the “elds of
psychology and psychiatry over the past 50 years. In the
1950s, researchers in the “eld of diagnostic reliability recog-
nized that one of the key factors contributing to the problem
of low interrater reliability in psychiatric diagnosis was the
inability of two or more examiners to achieve a consensus on
the symptoms or behaviors that characterized a speci“c diag-
nosis (Matarazzo, 1983). This is not unlike the current state
of affairs regarding diagnostic criteria for CFS. Because a
diagnosis or classi“cation can be no more accurate than the
classi“er •s knowledge and understanding of what he or she is
classifying, it was determined that the “rst step to improving
diagnostic reliability was the development of operationally
explicit and objectively denotable criteria (Feighner et al.,
1972).
By the 1970s, researchers in the “eld of diagnostics also
recognized that the provision of operationally explicit, objec-
tively denotable criteria was not enough to ensure that clini-
cians would know how to elicit the necessary information
from a clinical interview to permit them to apply it to the re-
liable criteria (J. Endicott & Spitzer, 1977). These concerns
led to the development of a series of structured interview
schedules. Structured interview schedules ensure that clini-
cians in the same or in different settings conduct clinical
interviews using standardized questions that maximize the
accuracy of clinical diagnosis (J. Endicott & Spitzer, 1977).
Thus, structured interview schedules serve to remove unreli-
ability introduced by differences in the way clinicians elicit
clinical information. Together, the provision of operationally
explicit, objectively denotable criteria and standardized inter-
views were found to signi“cantly improve the reliability of
clinical diagnosis for a number of psychological and psychi-
atric conditions (Leckliter & Matarazzo, 1994). It is possible
that similar strategies might be used to enhance the reliability
of CFS criteria.
Diagnostic and epidemiological research requires diag-
nostic categories that are both reliable and valid (Cantwell,
1996). The criteria used in different case de“nitions must be
clearly operationalized. Field tests must be conducted to
determine the reliability and validity of these nosologies. In
the determination of psychiatric diagnosis, considerable im-
provements were made to the DSM-IV(American Psychiatric
Association [APA], 1994) when committees were appointed
to make recommendations concerning different features
of the overall diagnostic system (Leckliter & Matarazzo,
1994). These recommendations were implemented in nation-
wide “eld trials to establish diagnostic reliability. This ap-
proach might be used to bring greater precision to the case
de“nition of CFS.
Several investigators have tried to validate or con“rm ap-
proaches for the classi“cation of fatigue using statistical
methods (Haley, Kurt, & Hom, 1997; Hall, Sanders, &
Repologle, 1994), or by attempting to distinguish psycholog-
ical from physical fatigue (Katerndahl, 1993). Others have
tried to clinically con“rm the CFS criteria established by the
Centers for Disease Control (CDC; Komaroff et al., 1996).
Nisenbaum, Reyes, Mawle, and Reeves (1998) found that
three correlated factors (fatigue-mood-cognition symptoms,
”u-type symptoms, and visual impairment symptoms) ex-
plained a set of additional correlations between fatigue last-
ing for six or more months and 14 interrelated symptoms. No
factor explained observed correlations among fatigue lasting
for one to “ve months and other symptoms. Findings like
these are of great interest because they indicate that only fa-
tigue lasting six or more months (with selected symptoms)
overlaps with published criteria to de“ne CFS.
In another study, Friedberg, Dechene, McKenzie, and
Fontanetta (2000) found three factors (cognitive problems,
”u-like symptoms, and neurologic symptoms) in a sample of
patients with CFS, and those with longer duration had a
larger number of cognitive dif“culties. Hadzi-Pavlovic et al.
(2000) used latent class analysis to classify patients with CFS
into three classes: those with multiple severe symptoms,
those with lower rates of cognitive symptoms and higher
rates of pain, and those with a less severe form of multiple