When the sensitivity values are plotted against the (1-specificity) values, a receiver-operating characteristics curve is obtained. The Area Under the Curve (AUC) measures the performance of a screening test. The maximum area of 1.0 says that the screening test is as good as the gold standard; an area of 0.5 says that the test is totally useless. One can use special statistics to compare the areas of two different ROC curves generated by applying two screening test to same individuals, so as to determine whether the performance of one test is statistically different (better or worse) from another one.

A more updated method to assess the performance of a screening test is the likelihood ratio, defined as sensitivity/(1-specificity). It is used to generate post-test odds and probabilities which can be compared with pre-test odds and pretest probabilities (prevalence of disease). A large change between the pre-test and post-test values would suggest that screening is worthy as it would change a clinical practice. Positive and negative predictive values are defined as the probabilities that someone with a positive or negative screening result is actually a case or a non-case. They are intuitively important. However, as they are dependent on prevalence of disease; they cannot be used to assess accuracy of screening test.