Predictive Value of a Single Diagnostic Test in Unselected Populations

WHEN a new test for a disease is being evaluated it is customary to perform the test in two selected groups of subjects: those with an indisputable diagnosis of the disease by other criteria; and those from the normal population who have no evidence of the disease and in whom all the factors known to result in a higher than normal risk of the disease can be excluded.1 2 3 The test results may be expressed dichotomously as "positive" or "negative," or by some numerical units along a scale, usually with the bulk of values on either side of an arbitrary dividing . . .