Analysis of case-control studies of the efficacy of screening for cancer: How should we deal with tests done in persons with symptoms?

Many tests used to screen asymptomatic individuals for the presence of cancer also are used in symptomatic persons to help determine the presence and/or extent of that cancer. For example, mammography can be used both to screen for breast cancer in a woman coming for an annual checkup and to help in the diagnosis of a woman who has recently noticed a lump in her breast. Case-control studies that seek to evaluate the efficacy of a screening test of this type in reducing cancer mortality need to consider how to deal with tests done in persons with symptoms possibly due to cancer. On the surface, the resolution to this issue seems obvious. If a test had been done in the absence of symptoms of the cancer, it should be categorized as a screening test; otherwise, it should be considered a diagnostic test (1). However, in the conduct of a specific study, circumstances may be present that make the choice among analytic options not at all obvious.