Although the data on the nurse form would be collected retrospectively, the nurses who completed the form were acquiring the information prospectively. The data include not only prevalence, but severity as well, and a large amount of additional data about demographics and the disease could be collected from the charts. A chart review could yield a large sample size, facilitating a multivariate statistical analysis that could describe differences in factors associated with fatigue in cancer and fatigue in CHF. The chart review is consistent with the resources we have. The data we would get are scientifically credible and would duplicate any cross-sectional prospective study that we could do. We should definitely do the chart review.

INCORRECT

The correct answer is (c).

It may well be a reasonable decision to do the chart review, but your analysis of the plusses and minuses is flawed. You do highlight some of the potential advantages of the chart review study: It would be relatively quick and easy to do, and it would require fewer resources. The information available in the chart can allow interesting analyses that include the type of information found in medical records, specifically demographics, disease-related information, and some information about relevant comorbidities (e.g. blood test results). You were not correct in saying, however, that the approach would be as scientifically credible as a prospective survey. Although the information about fatigue came from the patients, which is essential, the methodology used does not duplicate the effort to reduce bias that should be part of a prospective survey. For example, there was no control over the number or training of interviewers (the nurses doing the admissions). The instrument to query fatigue is only face valid and asks for patient recall from a period before the illness that brought them to the hospital. Recall bias is a real risk in this setting, and in the absence of validation data, the ability of the question to ascertain fatigue status cannot be accepted. In short, the chart review could provide data that is of interest, and could certainly be used as pilot data for a grant proposal, but it is not true that the information would duplicate a well designed prospective survey.