[Skip to Navigation]
October 1991

Percutaneous Injuries Among Health Care Workers: The Real Value of Human Immunodeficiency Virus Testing of 'Donor' Blood

Author Affiliations

From the Department of Pediatrics, the Division of Infectious Diseases, The Hospital for Sick Children (Drs Allen and Read), the Departments of Medicine and Health Administration, University of Toronto (Dr Detsky), and the Division of General Internal Medicine and Clinical Epidemiology, the Toronto Hospital (Drs Guerriere and Detsky), Toronto, Ontario. Dr Allen is now with the Children's Hospital of Eastern Ontario, University of Ottawa.

Arch Intern Med. 1991;151(10):2033-2040. doi:10.1001/archinte.1991.00400100107018

A decision analysis was conducted to examine whether health care workers should receive short-term (42 days) zidovudine treatment following percutaneous exposure to blood, as well as to determine the value of testing "donor" (patient's) blood. Three alternative options were analyzed: treat all, treat none, and test. In the treat all option, all health care workers receive short-term zidovudine therapy immediately after exposure; in the treat none option, no one receives zidovudine; and in the test option, donor blood is tested, and if it is human immunodeficiency virus (HIV) positive, zidovudine is given. Baseline variables were obtained from the literature. Each outcome was expressed as a utility; this is a method of quantifying the values that persons place on different health states. The results showed that the test option was preferred. Sensitivity analyses indicated that even if the risk of seroconversion were zero or the effectiveness of zidovudine were zero or the drug were withheld, this option was preferred, thus indicating some value of testing other than merely identifying health care workers who should receive zidovudine. In the baseline analysis, this was derived from the fact that approximately 95% of the health care workers would be reassured by a negative test; ie, only approximately 5% of donors are HIV positive. If the prevalence of HIV seropositivity exceeded 42%, the treat none option was preferred. This was found to be due to the fact that increased numbers of health care workers would be told that they were exposed to HIV-positive blood. The "worrying factor" associated with such an exposure was such that above 42% HIV seropositivity, the treat none option was preferred overall. Thus, the real value of testing donor blood is in identifying those persons (>95%) who could be told that they were exposed to HIV-negative blood, that is, reducing their worrying factor to zero. Because acquired immunodeficiency syndrome is a fatal disease, and given that zidovudine is the only available therapeutic option at present, the drug has an important role to play if its effectiveness is greater than zero.

(Arch Intern Med. 1991;151:2033-2040)

Add or change institution