Sensitivity and specificity are the indices most commonly reported when describing the performance of a screening or diagnostic test. These indices, and their corresponding predictive values or likelihood ratios, are fundamental test properties since they allow the user to determine the consequences of selecting a particular cut off criterion for referral or further investigative tests. (Sensitivity is the proportion of diseased individuals correctly identified as diseased and specificity is the proportion of non-diseased individuals correctly identified as non-diseased. The positive predictive value (PPV) is the proportion of patients with positive screening test results who are found to have disease and the negative predictive value (NPV) is the proportion of patients with negative screening test results who are found not to have disease, based on the gold standard. Given a test result, a likelihood ratio describes how many times more likely a patient with disease is to have that test result, compared with a patient without the disease.)

A recent article1 has highlighted the importance of complying with methodological standards2-4 when evaluating diagnostic or screening tests, in order that the findings of a study can be applied with confidence to clinical practice. These standards need to be considered at both the study design stage and the reporting stage (see Table 1). Standards 1 and 2 are closely related, since they are both concerned with the way in which the sensitivity and specificity of a test may vary depending on the clinical and demographic characteristics of a population (for example, disease stage, age, sex). These standards allow clinicians wishing to use a test to judge whether the sensitivity/specificity reported by the evaluation can be applied to their own population of patients. Standard 3 also relates to a study's population, being concerned with the bias which can arise if only a …