The most common initial screening test for HIV is the enzyme immunoassay or EIA, also known as ELISA. The test detects proteins the body makes in response to HIV. These proteins, called antibodies, are defensive molecules made by the B-cells of the body’s immune system. The most common fluid to check for antibodies is the blood, and rapid tests on saliva and urine are available as well. The EIA test has a long history of helping to screen patients for HIV. The accuracy of a test is measured in several ways, though. Accuracy may refer to either the test or the test result. As for the test, a person might want to know how well the EIA is at finding people with and without HIV. As for the test result, he might want to know how much to trust a positive or negative test result. The measures of accuracy for the common EIA were established by large, well-designed studies in the late 1980s and early 1990s. More recently, studies have looked at the newer fourth-generation of tests as well.
According to the book “Medical Management of HIV Infection” by Johns Hopkins physician John Bartlett, the EIA has a sensitivity of 99.3 to 99.7 percent. This means that for every 1,000 people infected with HIV who are tested by EIA, 993 to 997 will test positive, and three to seven will test negative, which makes the test very sensitive and useful. In addition, new fourth-generation tests combine the EIA with a test for antigens, which are proteins made by HIV. According to a study in 2009 in the journal Transfusion Medicine, the sensitivity of these fourth-generation tests is 100 percent.
The specificity of EIA is at least 99.7 percent and may be higher, according the landmark studies on the EIA by the Centers for Disease Control in the late 1980s and later reconfirmed. So, for every 1,000 people without HIV who are tested, at least 997 will test negative, and three or fewer will test positive, making this measure of the test's accuracy very high. The 2009 study in Transfusion Medicine found the specificity of the fourth-generation tests to range from 99.91 percent to 99.97 percent. When these tests are used, about 5 out of 10,000 people would be expected to have a false negative.
Positive Predictive Value
The positive predictive value indicates the likelihood that a positive result is correct. Unlike the above measures of accuracy, positive predictive value depends on the number of cases in the population. More cases means the positive predictive value will be higher. The landmark study in the New England Journal of Medicine in 1988 found that in a population with a relatively low number of AIDS cases, the number of false positives was about 1 in 135,000. This low rate would be improved even more with the fourth-generation tests, according to a study in 2009 in the Journal of Clinical Microbiology.
Negative Predictive Value
The negative predictive value indicates the likelihood that a negative test result is correct, and it also depends on the number of cases. More cases lead to lower negative predictive value. A study in the Journal of Infectious Disease in 1993 showed that the number of false negatives in a population with about 30 percent of people infected with HIV is about 0.3 percent. A study in the New England Journal of Medicine in 1991 showed that in a population with a lower number of HIV cases, those of U.S. blood donors, the rate of false negative was about 0.001 percent. The authors of the 2009 article in Transfusion Medicine note improvement of false negative rates with the fourth-generation tests. When the HIV antibody levels were very low, the fourth-generation test detected seven times the number of cases compared with the standard test.
The EIA is a very accurate test. However, the test may be falsely negative during what is known as the “window period.” This is the time after infection with HIV when the body has not yet created antibodies, and it may take several weeks for the test to detect the antibodies; the later generations of tests are more sensitive for this period. The EIA is also a screening test. A positive test on the EIA is not definitive. Instead, it is followed up with a repeat of the EIA test, plus another confirmatory test called a Western blot. If uncertainty still exists, an RNA test is usually performed next.