Accuracy of Antibody Testing -- and Denialist Arguments
Constantine(3) sums up the general consensus among experts and institutions such as the CDC when he says "The antibody tests are nearly 100 percent sensitive (unless a person is in the window period) and about 99 percent specific." Such levels of accuracy have been documented in a number of studies, including periodic evaluations of commercially available test kits conducted by the World Health Organization.
Still, AIDS denialists (the self-styled "AIDS dissidents" who claim that HIV is either harmless or doesn't exist) continue to claim that HIV antibody tests are unreliable. Many of their arguments seem to derive from a series of articles written by Christine Johnson in the mid-1990s, several of which are available on denialist web sites.(4,5,6)
Johnson's argument boils down to two key points: 1) HIV has never been properly isolated, so the HIV proteins used in the tests haven't been proven to actually come from HIV, 2) Even if HIV is real, the proteins are not unique and cross- react with many other antigens, rendering a positive result meaningless. Johnson's list of some 60 factors she describes as "known to cause false-positive HIV-antibody test results" turns up regularly in denialist literature.
The claim that HIV has never been properly isolated, based on the writing of a group in Perth, Australia, is too technical and complex to examine thoroughly here. However, it is elegantly demolished in Michael Coon's article, "HIV, AIDS and the Distortion of Science," available on the AEGIS web site.(7) In short, Coon argues that the Perth Group set up artificial, phony criteria for "proof" of HIV's isolation that bear no relation to how virology works in the real world.
The second argument, though, contains a grain of truth. Cross-reactions are possible, and a number of factors can, on occasion, produce false-positive HIV antibody test results. What Johnson fails to address in any detail is that such effects are typically transient and rare, affecting few individuals.
For example, one well-known cause of false-positives Johnson lists is influenza vaccination.(8,9) But she neglects to mention that a key reference she cites described the phenomenon as "infrequent" and "of short duration," while in another only 10 false-positives were found among 133,000 individuals who had flu shots prior to testing, with half of those reverting to negative within six months.(9)
Constantine adds, "I doubt very much that it has been firmly documented that 60 factors can interfere with antibody tests. In fact, it has been long sought to try and identify the causes of false positive results, and only a few have really been documented to consistently interfere (e.g. pregnancy, certain autoimmune diseases, some infectious diseases). However, even these do not consistently cause problems with the tests... There are very few false positives that can't be resolved with further testing."
Consent, Anonymity, and Counseling
(1) Anonymous Testing
Prior to HIV, blood testing was considered a routine procedure, with such minimal dangers that formal informed consent was rarely required. But because HIV presented massive psychosocial risks, from employment discrimination to rejection by family and reduced access to health care, special procedures were widely adopted.(10) These included specific informed consent and pre- and post-test counseling. Many states set up test sites where people could get tested anonymously, without ever giving their name.
Anonymous testing (other than the home test, below) was never universally available, Morin notes, but was and is offered in many places, despite the recent move by numerous states to adopt a system of names-based HIV reporting (a few, including California, are implementing HIV reporting via codes that don't reveal the person's name). The CDC and others urged that the option for anonymous testing should be kept available, believing fear of disclosure would keep some from being tested, and most states have followed this recommendation.
Because local laws and procedures vary, Morin recommends that anyone concerned about anonymity or disclosure contact their local health department to check. A number of AIDS service organizations operate hotlines, which should also be able to provide this information.
Those living in areas with no anonymous test sites can still be tested anonymously via home collection test kits, which are sold in many drug stores. Introduced in the mid- '90s, the kits were controversial because counseling is provided by telephone rather than in person. Morin says fears that telephone counseling would prove inadequate haven't been borne out, but sales of the kits have been less than expected. Still, "the FDA ruled that you cannot bar their sale in any state, so even in states that don't have anonymous testing people can use home test kits to anonymously be tested," he says.
But, he adds, things change when the individual seeks treatment: "If you go to your doctor and the doctor does a viral load test, you get reported through the viral load test to the health department. So there's no way to keep treatment for HIV anonymous."
(2) Consent and Counseling
As with anonymity, requirements for consent and counseling vary from state to state. Most, but not all, states require specific informed consent -- sometimes in writing -- for HIV testing. Approximately one-fifth of states require pre- test counseling, with many listing specifically what that counseling must include. The U.S. Department of Health and Human Services recommends that all HIV testing include counseling that covers the test itself, basic information about HIV and AIDS, how to avoid spreading the virus, the confidentiality of the results, the possible impact of the results on the person being tested, and discussion of to whom results should be disclosed, such as sex or needle- sharing partners.(11) Counselors should also be able to give referrals to medical and psychosocial support services.
Counseling and consent procedures vary greatly, remain controversial and may continue to change. Even states that require informed consent may allow HIV testing without consent in special circumstances. For example, many permit involuntary testing of a patient when health workers have been exposed to the person's blood. Some test prisoners or people accused of sex crimes, and at least two, New York and Connecticut, require mandatory testing of newborns, which indirectly reveals the mother's HIV status, but does not tell if the infant has been infected.
In October 2000 the Institute of Medicine recommended that HIV testing be included as a routine part of prenatal care. Women would be informed of the test and could opt out, but specific consent would not be required. Thus far the U.S. Public Health Service has stopped short of urging an end to informed consent in such cases, simply suggesting that providers recommend HIV testing to all pregnant patients.