Background and aims: Requirements for diagnostic tests include high validity and good repeatability.
Methods and results: The validity of a test refers to what extent the test measures what it is supposed to measure and is often expressed in terms of sensitivity and specificity. A graph may be constructed to incorporate both the sensitivity and specificity of a given test, the receiver operator characteristic (ROC) curve. The ROC curve actually compares the true positive fraction (sensitivity) vs the false positive fraction for a series of cut-off points for a given test. Use of the curve requires that the diagnostic test and the gold standard must be independent of each other. Repeatability is the ability of a test to give the same results in repeated measurements.
Conclusion: When designing a study to examine a diagnostic test, the following principles should be followed. First, the order of performance of the diagnostic test and the gold standard in repeated measurements should be randomised. Second, in repeated measurements, the test situations should be equal. Third, the technicians performing the tests should be blinded to the data obtained in the other test or to the result of the gold standard. A pilot study is recommended before conducting the study to examine the diagnostic test.