[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 54.163.173.253. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Article
November 1983

Biases in the Laboratory Diagnosis of Depression in Medical Practice

Author Affiliations

From the Department of Medicine, University of California, Los Angeles (Drs Shapiro and Greenfield), and the Department of Psychiatry, University of Rochester (NY) (Dr Lehman). Dr Shapiro is a Henry J. Kaiser Family Foundation Faculty Scholar in general internal medicine.

Arch Intern Med. 1983;143(11):2085-2088. doi:10.1001/archinte.1983.00350110063017
Abstract

• Several studies, conducted in psychiatric settings, have reported that the dexamethasone suppression test (ST) is useful in the diagnosis of endogenous depression. To determine whether the test has clinical utility in internal medicine practice, data were reviewed and reanalyzed from all studies that evaluated the dexamethasone ST in the diagnosis of depression. In these 11 studies, the mean positive predictive value reported for the test was 84%, reflecting high prevalence of disease (50%), sensitivity (43%), and specificity (92%) in those specialized populations studied. When estimates of the prevalence and severity of conditions associated with depressed mood seen in internal medicine were used, the sensitivity dropped to 30%, and the specificity to 85%. Assuming a prevalence of 20%, a figure based on available epidemiologic data, the predictive value for a positive test would be 33%, too low to have value in most clinical settings. Data from studies currently available do not support the use of dexamethasone ST in internal medicine practice. More generally, before any test is adopted in a general medical setting, it should be studied in that setting, or available data should be reanalyzed to correct for biases affecting prevalence, sensitivity, and specificity that may inflate the test's value.

(Arch Intern Med 1983;143:2085-2088)

×