Formal approaches
to the assessment of medical comorbidities by either chart review
(Charlson
et al., 1987) or patient interview (Katz
et al., 1996) may be adapted for use in surveys or
clinical trials. Alternatively, a checklist can be developed
that can codify the existence of the most important of these
phenomena. The checklist can be completed using information
from the medical record (e.g. recent hemoglobin level), patient
history and physical examination (e.g. degree of physical inactivity),
or brief screening tools that are added to the questionnaire
packet (e.g. depression screen).
In some cases, additional
relevant information can be obtained from an intervention that
would usually be considered to pose minimum burden, such as
pulse oximetry or blood sampling to measure hemoglobin or electrolytes.
Other information is obtainable only with more elaborate testing,
such as pulmonary function tests or electrodiagnostic studies.
Like the use of multiple questionnaires, these tests can be
burdensome for some fatigued patients. Respondent burden must
be balanced against the anticipated value of the data when designing
the assessment protocol for a study.
Most investigators
are biased toward the collection of as much information about
comorbidities as possible. In all types of studies, surveys
and controlled trials, these data may provide the means to perform
post-hoc analyses that could potentially clarify important relationships
and generate new hypotheses. In controlled trials, one or more
of these variables may be considered so likely to influence
the primary outcome (usually fatigue severity) that they either
are selected for stratification among the study groups or are
prospectively included in the data analysis plan.