This Article  • References  • Full text PDF  • Correction  • Send to a friend  • Save in My Folder  • Save to citation manager  • Permissions  Citing Articles  • Citation map  • Citing articles on HighWire  • Contact me when this article is cited  Related Content  • Similar articles in JAMA  Social Bookmarking   What's this?

Michael J. Barry, MD; Albert G. Mulley, MD, MPP; Daniel E. Singer, MD
Massachusetts General Hospital Boston

JAMA. 1985;253(23):3395-3396.

	Since this article does not have an abstract, we have provided the first 150 words of the full text PDF and any section headings.

To the Editor.—

We are concerned that in the recent article by Weiss et al,¹ some of the test operating characteristics reported may have been misleading. Specifically, the authors conclude that their screening test for human T-cell leukemia (lymphotropic) retrovirus type III (HTLV-III) antibodies (as a marker for the ability of a unit of blood to transmit acquired immunodeficiency syndrome [AIDS]) was "highly sensitive and specific" based on values calculated after "excluding borderline results." Neglecting these "borderline" results is inappropriate, as test results in the overlap zone between distributions for disease and nondisease actually define a test's sensitivity and specificity.²

Sensitivity-specificity pairs for different cutoffs can be determined from data in the article's Table 1 (combining all patients with AIDS to define "disease" and Burlington blood donors to define "nondisease"). However, it may be more informative in examining a screening test to scrutinize the ratio of false-positives to . . . [Full Text PDF of this Article]

Footnotes
Edited by Drummond Rennie, MD, Senior Contributing Editor.

CiteULike    Connotea    Del.icio.us    Digg    Reddit    Technorati    Twitter     What's this?