[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address Please contact the publisher to request reinstatement.
[Skip to Content Landing]
April 8, 1983

Reference Groups: Comparing Oranges With Oranges and Apples With Apples

Author Affiliations

Veterans Administration Medical Center West Roxbury, Mass Harvard Medical School Boston; Department of Pathology LSU Medical Center New Orleans

JAMA. 1983;249(14):1890. doi:10.1001/jama.1983.03330380078033

The article by Ryder et al, "Serum Angiotensin Converting Enzyme Activity in Patients With Histoplasmosis" (p 1888), provides a good opportunity to discuss a common error in the clinical assessment of any proposed new diagnostic test. Such a test is often developed for the diagnosis of a specific disease. It is then found to show positive results in a high percentage of patients with the disease. Next, a control group, often consisting of a group of volunteers, healthy medical students, some blood donors, or a randomly selected group of laboratory technicians, is selected for testing. The results are found to be normal, or negative, in this control group; on that basis, the laboratory test is then claimed to be specific for the disease. Results of the clinical evaluation of this new laboratory test are published in a medical journal; there is a resultant flush of enthusiasm for the test. Clinicians