[Skip to Navigation]
September 1989

Validation of a Diagnostic Test: Epidemiologic Principles in Dermatology

Author Affiliations

Department of Dermatology of Brown University Veterans Administration Medical Center 650-111 Davis Park Providence, RI 02908

Arch Dermatol. 1989;125(9):1260-1264. doi:10.1001/archderm.1989.01670210098018

How should we monitor methotrexate therapy in psoriatic patients to avoid hepatic cirrhosis? The standard approach requires liver biopsies after each 1.5 g of methotrexate, in addition to performing liver function tests three to four times per year. In this issue of the Archives, O'Connor and his colleagues1 suggest that we may be able to avoid exposing many of our patients to the risk of morbidity and mortality associated with multiple liver biopsies by using simple and relatively inexpensive liver function tests. They derive a prediction rule for hepatic fibrosis based on the results of these tests.

The evaluation of the validity of such a prediction rule involves several basic epidemiologic concepts related to diagnostic testing.2-10 These are reviewed below in the context of predicting hepatic fibrosis with a liver function test. Mentioned below are some general pitfalls in the assessment of test validity, and a discussion of more