The research questions should guide the study protocol. Data can be collected retrospectively from the medical record or other sources, or prospectively from the patient. In the latter case, complementary data also can be obtained from the medical record or other sources. Retrospective studies of phenomena that are inherently subjective, such as fatigue, usually are problematic because fatigue is often inadequately assessed and the notes written by physicians and nurses are unlikely to provide a valid and reliable record of a patient's experience. For this reason, retrospective studies of the phenomenon are not preferred.

Retrospective collection of information about fatigue would yield credible data only if the symptom was systematically evaluated as an integral aspect of clinical practice. For example, the routine administration of a symptom assessment scale to all patients admitted to a hospital unit might yield some credible information about fatigue that could be captured through a retrospective review.

Prospective data collection from the patient is the "gold standard" for surveys of subjective phenomena. The challenge in these surveys is to design a questionnaire packet and an assessment protocol that will yield high recruitment and completion rates, and will also adequately address the specific aims of the study.

Eligibility criteria include a case definition for fatigue and other information that defines the study population. Eligible patients complete one or more questionnaires. In most cases, it is appropriate to include a multidimensional fatigue scale, supplemented by other questions about fatigue or related phenomena, as discussed previously.