Principles of Diagnostic Testing Series
Part 1 – Test Characteristics: Sensitivity and Specificity
The performance characteristics of a diagnostic test influence the circumstances that the test is most useful and how the results might be interpreted. Ideally, every diagnostic test would perfectly distinguish diseased from non-diseased individuals, however, this is rarely the case. All diagnostic tests make errors. It is useful to classify the possible results of a diagnostic test into four possible outcomes: test positive and diseased (true positive; TP), test positive and not diseased (false positive, FP), test negative and not diseased (true negative; TN), or test negative and diseased (false negative, FN). Commonly, a 2 x 2 table is created to visualize these outcomes.Diagnostic tests are evaluated by conducting the test on a population with the disease and from a population without the disease.
Sensitivity is the proportion of individuals with a disease who test positive on a diagnostic test.
Specificity is the proportion of individuals without a disease who test negative on a diagnostic test.
Sensitivity and specificity are characteristics of a diagnostic test. Diagnostic tests with outcomes of continuous numbers have different sensitivity and specificity for each numerical value or cut-off point. The selection of a cut-point to increase test sensitivity reduces test specificity and vice versa. Therefore, a cut-off point is chosen to optimize either test sensitivity or test specificity depending on the relative cost of a false-positive or false-negative result.