If you have issues viewing or accessing this file contact us at NCJRS.gov. . .7 --

" U.S. Department of Justice National Institute of Ju~tice

ASQH~Z . All'V<£Y' ill n 1f 4- at t f J t~W\l&7'tJ:~~INr.'8.fjl . 1iJ) Na t·ll.onal .llmill\,l U e 0 us Ice "." Bulletin

October 1988

lHIKV AlThtibody T~§ting~

JPIroce((}lulres 9 lIIDlteIrlPHret~tiolTIi9 2InHdl ReRiabnHnty of lReslUlfits

Theodore M. Hammett, Ph.D., Abt Associates, Inc.

Introduction The test was originally developed to screen In addition, a few States instituted the blood supply and it has virtually In 1983 ar.d 1984, scientists at the screening of all inmates in their correc­ eliminated transfusion-associated HIV in­ Institut Pasteur in Paris and at the tional institutions. fection. Quickly, however, requests were National Institutes of Health identified In the spring of 1987, calls for mass made for use of the test to screen groups of and isolated the cause of AIDS: a virus screening-and, increasingly, for people. The most widely publicized early now generally known as the human mandatory mass screening-began to be application was the Defense Department's immunodeficiency virus (HIV). In early heard. Pursuant to Presidential directive, screening of all potential military recruits. 1985, a commercial test for antibodies to the screening of all Federal prisoners and HIV became available. immigrants was begun. Subsequently,

From the Director continue to perfonn their duties in a safe published and widely disseminated. A and professional manner. third report has just been published that Acquired Immunodeficiency Syndrome­ addresses AIDS as it impacts probation AIDS-has been called the most serious Since 1985, the National Institute of Justice and parole services. public health problem in the United States has worked with the Centers for Disease and worldwide today. Since it ftrst Control and other public health officials to President Reagan has said that the AIDS appeared in 1981, there has been an provide important authoritative medical crisis "calls for urgency, not panic ... enonnous amount of uncertainty and fear infonnation about AIDS to criminal justice compassion, not blame. .. understanding, about this fatal disease. Because they may professionals. not ignorance." The National Institute of be in contact with intravenous drug users Justice is working to ensure that criminal and others at high risk for the disease, This AIDS Bulletin is part of a new series justice professionals have the accurate criminal justice professionals understand­ designed to infonn criminal justice infonnation they need to understand and ably are concerned about becoming professionals about the disease and its deal with the risks created by AIDS. Until infected with the AIDS virus While implications for criminal justice agencies. medical science can bring this deadly carrying out their duties. Future bulletins will summarize agency disease under control, our best defense is a policies relating to AIDS, education well-infonned citizenry. Until a vaccine or cure for AIDS is found, programs, and legal and labor relations education is the cornerstone of society's issues. response to this deadly disease. Accurate James K. Stewart infonnation can help dispel misinfonna­ In addition, two special reports on AIDS­ Director tion about the disease and its transmission, as AIDS relates to corrections and law thus enabling criminal justice personnel to enforcement agency procedures-have been there has been a proliferation of propos­ === nonmandatory, or less extensive, testing als for screening State prisoners, mar· ~ .. before deciding to conduct progr2:.'!1. If policymakers cannot clearly riage license applicants, persons admitted mandatory screening in any reach these determinations, they ought to to hospitals, and other populations. decide against mandatory screening population, policymalcers should programs. The position of the Centers for Disease determine (1) the precise objec­ Control (CDC) on testing underwent a This AIDS Blllietin will not discuss the subtle-but important-change during tives of the screening program, debate over mass screening in any further 1987. The CDC shifted from advocating (2) exactly how the test results detail. The interested reader will find the voluntary counseling and testing of will be used to reduce transmis­ issues of concern well covered in an persons and their sexual contacts who sion of the virus, and (3) that it extensive and rapidly growing literature.2 have engaged in high-risk behaviors, to would be impossible to achieve Instead, this Bulletin summarizes some of favoring "routine" counseling and testing the key technical issues surrounding mv of such individuals.! "Routine" means the same objectives through a antibody testing: What is an mv anti­ that testing will be performed as part of nonmandatory, or less extensive, body test? What do the test results mean? normal medical procedures unless the testing program. §9 What methods are used in the basic and individual specifically declines to be .. confirmatory tests? How much do the tested. In other words, CDC has shifted tests cost? How reliable are confirmed the individual's responsibility from results? requesting the test to declining the test. infected individuals underground and subject those who are tested to significant In reality, however, it is impossible to There continues to be widespread discrimination. Opponents also argue separate these technical issues from the controversy surrounding all policies and that the essential educational messages ethical and legal issues surrounding HIV proposals to conduct mandatory screen­ and prevention strategies remain the antibody screening. Thus, decisionmakers ing for HIV antibodies. Proponents same, regardless of antibody status. They in the criminal justice system and else­ generally argue that effective prevention feel that differentially targeted educa­ where must consider all areas as they seek measures and medical care depend upon tional programs may tend to create a fEtlse to develop rational and effective AIDS identifying infected individuals. They sense of security in seronegative popula­ policies. point to the value of targeted educational tions and undermine the essential efforts; the utility of segregating infected message that everyone must be careful. The nature and meaning persons in hospitals, correctional institu­ Precautions against unprotected exposure of HIV antibody testing tions, and other custodial facilities; the to blood and body fluids should be taken importance of health care, correctional, by all persons, regardless of antibody There appears to be significant confusion and custodial staffs knowing the HIV status. Opponents of mandatory mass about the nature and meaning of HIV antibody status of the persons with whom testing fear that incomplete information antibody testing. Although terms such as they have regular contact; and the on HIV status may undermine consistent "AIDS testing" are often used, and possible benefits of antibody status application of these precautions. Finally, individuals are commonly said to have information to diagnosis and medical there is no effective therapeutic treatment "tested positive for AIDS," the fact is that intervention. for persons infected with HIV who there is no test for AIDS. The available tests do not determine whether or not all Opponents counter that mandatory remain asymptomatic. The available individual has AIDS; rather, AIDS can screening will be counterproductive drugs are generally only prescribed when only be diagnosed through identification because it will drive many potentially symptoms appear. Clinical trials regard­ ing the benefit of AZT in treating asymp­ of opportunistic infections or malignancies tomatically infected persons are currently indicating an underlying immune defi­ Points of view or opinions expressed in this ciency caused by HIV infection, or central publication are those of the author and do not in progress but the results will probably not be known until 1990. nervous system disorders now known to necessarily represent the official position or be caused directly by HIV infection. policies of the U.S. Department of Justice. In general, before deciding to conduct mandatory screening in any population, Readily available tests do not even detect The Assistant Attorney General, Office of the presence of HIV itself-only the Justice Programs, coordinates the activities of policy makers should determine (1) the the following program Offices and Bureaus: precise objectives of the screening presence of antibodies to the virus. National Institute ofJustice, Bureau ofJustice program, (2) exactly how the test results Antibodies in the blood are evidence of Statistics, Bureau ofJustice Assistance, Office will be used to reduce transmission of the the immune system's attempt to fight off of Juvenile Justice and Delinquency Preven­ an infection. Actual culturing of the virus tion, and Officefor Victims of Crime. virus, and (3) that it would be impossible to achieve the same objectives through a (i.e., growing the virus from a specimen of

2 body fluid or tissue) is very difficult, time consuming, and expensive. It is currently perfonned only in a small number of What is an HIV antibody test? How much do the tests cost? research laboratories. However, antigen Antibody tests detect the presence of anti­ Testing costs vary considerably, depending tests, which detect a part of the virus, bodies to the AIDS virus-they do not de­ on how the testing is conducted. When may be available within the year. termine whether or not an individual has buying test kits in high volume, testing may AIDS. There is no test/or AIDS. cost as little as $2 to $3 per subject, which A properly confirmed positive result (see would include ELISA tests and any Western below) for HIV antibodies means that the What do the test results mean? Blot confirmatory tests necessary. Other­ wise, the reported range for ELISA testing individual was infected at some time in A properly confirmed positive result for the past, although the test cannot pinpoint is from $5 to $40, depending on whether the HIV antibodies means that the individual agency draws the blood and sends it out to a the date of infection. However, CDC rec­ was infected with the virus at some time in laboratory for testing or sends Individuals to the past, although the test cannot pinpoint ommends that all persons with confirmed private physicians to be tested. The the date of infection. However, CDC rec­ Western Blot test is much more expensive positive test results be considered ommends that all persons with confirmed infected and capable of transmitting the than the ELISA. The average cost per test oositive test results be considered infected is $75 with an approximate range of $25 to virus to others. Md capable of transmitting the virus to $150. others. According to the latest guidelines A great deal of uncertainty surrounds the from CDC, an individual should be consid­ questions of whether and when an ered positive for antibodies to HIV when a How reliable are confirmed results? infected individual will develop AIDS. sequence of three tests is consistently There are two areas of serious concern The average time between initial HIV positive. about the reliability of HIV antibody tests: (1) the problem of lag time between infection and the appearance of symp­ infection and the appearance of detectable toms appears to be about 8 years, but the What methods are used in the basic and confirmatory tests? antibodies; and (2) technical problems with range in individual cases is extremely the tests and testing procedures that may The basic test is an Enzyme-Linked Im­ produce incorrect results. The lag time broad. Because of this long incubation munosorbent Assay (ELISA or EIA). The period and the relative newness of the makes it impossible to guarantee detection ELISA test is generally performed on of all infected members of a population disease, it is not known precisely what blood drawn through venipuncture, through one-time screening. The technical percentage of infected individuals will go although a version based on a blotted problems cause currently a\'ailable HIV on to develop AIDS, or even evidence fingerprick is now being used as well. If antibody tests to be subject to error, even the first ELISA test is positive, a second when recommended confirmatory proce­ milder symptoms of HIV infection, ELISA is performed on the same speci­ sometimes called AIDS-Related Complex dures are used. Both problems should be of men. If that test is also positive, a contir­ serious interest to society and must be (ARC). Current estimates are that 65 to matory test-usually the Western Blot carefully considered before any testing 100 percent of infected persons will test-is performed, also on the same blood program-particularly a mandatory testing progress to diagnosed AIDS.3 sample. program-is instituted. Notably, a negative result on the HIV antibody test means only that the individ­ ual was not infected with HIV (or was infected but had not yet developed antibodies) at the time the blood sample 6 was taken. It says nothing about the Basic and confirmatory fingerprick is now being used as well. If likelihood of future infection or suscepti­ tests used the first ELISA test is positive, a second ELISA is performed on the same speci­ bility to infection. Indeed, this is one of According to the latest guidelines from the key messages to present in posttest men. If that test is also positive, a CDC, an individual should be considered confirmatory test-usually the Western counseling of seronegative persons. positive for antibodies to HIV when a Individuals who have engaged and are Blot test--is performed, also on the same sequence of three tests is consistently blood sample.7 continuing to engage in high-risk behav­ positive. 5 The basic test is an Enzyme­ iors should be told that their negative Linked Immunosorbent Assay (ELISA or The ELISA test was developed in the mid- result represents "pure luck" and that the EIA). The ELISA test is generally 1980's for the screening of blood supplies. only way to reduce their likelihood of performed on blood drawn through The presence of HIV antibodies is becoming infected in the future is to venipuncture (in which blood is drawn by signaled by a color reaction quantified discontinue these behaviors immediately puncturing a vein through the skin), through the use of a spectrophotometer. or, at least, to begin taking appropriate although a version based on a blotted The results are measured on a continuous precautionary measures.4 numerical scale representing a color

3 cutpoints be set relative to the presumed Lag time between infection and 65 The available tests do not deter­ level of actual infection in the population appearance of antibodies to be tested and to the consequences of mine whether or not an indi~idual CDC estimates that, on average, 6 to 12 inaccurate results; that is, a higher cutpoint weeks elapse between an individual's has AIDS; rather, AIDS can only for a population with a lower prevalence of infection with HIV and the appearance in be diagnosed through identifica­ infection and a lower cutpoint for a the blood of detectable antibodies to the population with a higher prevalence of tion of opportunistic infections or virus.l1 However, there have been infection.9 malignancies imlicating an isolated reports oflag times of up to 6 underlying immune deficiency Numerous efforts are currently underway months, and recent data suggest that even caused by HlV infection, or to develop testing methods for reliable longer delays in antibody appearance may central nervous system disorders direct detection of IllV, as well as to not be unusual. In addition, there is new improve the antibody tests. As noted, evidence that the virus may sometimes go now known to be caused directly improved testing technology may be into a latent period, during which it is still by HlV infection.!>9 available within a year.lO present (and the person is still infectious) but its antibodies are undetectable by available tests. 12 density reaction to the level of antibodies The costs of HIV antibody testing These facts are extremely important in the blood. The higher the antibody because infected individuals are capable of level, the greater the optical density or Testing costs vary considerably, depending transmitting the virus from the instant they color change. Therefore, a decision must on how the testing is conducted. At one are infected. Infectiousness, in other be made as to the "cutpoint" on this scale extreme, testing may be furnished at no words, does not await the appearance of that distinguishes positive and negative charge at numerous CDC-funded alterna­ detectable antibodies. Negative antibody results. Manufacturers of the test kit tive test sites. If many individuals are to test results based on blood drawn during recommend setting a specific cutpoint for be tested and blood-drawing and labora­ this lag time are, in effect, false negatives. each kit based on the degree of reaction to tory capabilities are available in-house, a Such instances have produced the very the known positive and negative control low-cost option is the direct purchase of small number of HIV infections associ­ samples supplied.8 Interpretation of ELISA test kits. One of these kits typi­ ated with transfusions administered results is a key factor in the ultimate cally can be used to perform several since universal screening of blood supplies reliability of the test. hundred individual tests. When buying began in 1985. The blood transfused in kits in high volume, testing may cost as The Western Blot test is used to confirm these cases was donated by infected little as $2 to $3 per subject, which would twice-repeated positive ELISA results. persons before the antibodies bad include ELISA tests and any Western Blot For this test, inactivated virus is separated appeared. confirmatory tests necessary. Otherwise, into component parts and "blotted" onto the reported range for ELISA testing is The lag time problem should also be of special paper. Complexes of viral protein from $5 to $40, depending on whether the concern to policymakers contemplating and antibodies are seen as spots or bands agency draws the blood and sends it out to any type of mass screening program. The in the final preparation. The Western Blot a laboratory for testing or sends individu­ lag time makes it impossible to guarantee test is not sold commercially as a kit nor als to private physicians to be tested. detection of all infected members of a does it have standardized interpretive The confirmatory Western Blot test is population through one-time screening. procedures. much more expensive than the ELISA. Leaving aside the other reliability prob­ Because the ELISA was initially devel­ The average cost per test is $75 with an lems (discussed below), repeated followup oped to screen blood. the recommendeq approximate range of $25 to $150. Of testing of populations would be necessary course, the Western Blot is only performed cutpoints are deliberately set quite low to to maximize the probability of detecting if two ELISA tests are positive. minimize false-negative results. When all infected individuals. This may have screening blood, it is better to discard serious cost and logistical implications. possibly un infected donations than to use The reliability of test results However, an effective antigen test, when it possibly infected blood. Of course, the There are two areas of serious concern becomes available, may eliminate the lag low cutpoint designed for blood screening about the reliability of mv antibody tests: time problem because it detects a portion produces a relatively high false-positive (1) the problem of lag time between of the virus itself, rather than just antibod­ rate when the test is used to screen people. infection and the appearance of detectable ies to the virus. As a result, it has been suggested that antibodies; and (2) technical problems with the tests and testing procedures that may produce incorrect results.

4 Technical issues negatives may also occur, particularly in who may subsequently be infected by regarding the tests the high-risk populations that are of individuals they believe to be free ofHIY. special interest to criminal justice agen­ Both problems should be of serious The currently available HIY antibody tests cies. False positives are of particular interest to society and must be carefully are subject to error, even when recom­ concern to persons being tested, who may considered before any testing program­ mended conflrmatory procedures are used. suffer mental anguish and be subjected to particularly a mandatory testing pro­ The major problem appears to be with severe discrimination. On the other hand, gram-is instituted. false-positive results, although false false negatives are of concern to persons The precision of a biomedical test is expressed in tenns of the consistency of its results-that is, it is highly precise if it Figure 1 always yields the same results when Hypothetical HIV Antibody Screening in a repeated under similar circumstances. Population of 500 With a 20 % True Prevalence of Infection However, HIV antibody test results have been shown to be affected by relatively True Antibody minor variations in temperature, humidity, Infection Status Test Results False Results and other factors. 13 as % of Negative Positive True Group n % Result Result Procedural variations fuid quality control deflciencies can also adversely affect the performance of HIY antibody tests. The True Infected 100 20 1.0 99.0 1%" Western Blot is particularly susceptible Groups Uninfected 400 80 396.0 4.0 l%b Total 500 100 397.0 103.0 to human error and variability of results because most laboratories use unlicensed False Results as % 0.3%< 3.9%d of all Test Results test kits. As a result, unlike the ELISA in Category test, the Western Blot is not usually based on a standardized commercial product. "This reflects the test sensitivity of 99%. However, the ELISA is also subject to "This reflects the test specificity of 99%. variation because of the possibility that "11lis is the percentage of all negative results that would be false. different testing facilities will use different dJ'his is the percentage of all positive results that would be false. criteria for setting the positive-negative cutpoint that is critical to interpreting the test results. The lack of uniform quality control and proflciency standards Figure 2 for laboratories also results in test variability. 14 Hypothetical HIV Antibody Screening in a Population of 500 With a 1 % True Prevalence of Infection The accuracy of biomedical tests is generally measured in terms of sensitivity and speciflcity. CDC estimates that the True Antibody Infection Status Test Results False Results sensitivity and speciflcity of currently as % of licensed ELISA tests are both 99 percent Negative Positive True Group n % Result Result or higher (assuming that a double ELISA test is performed), and these estimates do True Infected 5 I 0.05 4.95 1%" not appear to be in question. Ninety-nine l%b Groups Uninfected 495 99 490.00 5.00 percent sensitivity means that, on average, Total 500 100 490.05 9.95 the test will correctly identify 99 out of False Results as % 0.01%< 49.8%d every 100 infected individuals. Ninety­ of all Test Results in Category nine percent specificity means that, on

., average, the test will correctly identify 99 out of every 100 uninfected individuals. "This reflects the test sensitivity of 99%. "This reflects the test specificity of 99%. "11lis is the percentage of all negative results that would be false. ctrhis is the percentage of all positive results that would be false.

5 In other words, 1 percent of infected persons will be false negatives on the test, Figure 3 and 1 percent of uninfected persons will Hypothetical Application of Mass Screening for be false positives on the test. This does Antibodies to HIV in a Population of 25,000 Inmates not mean, however, that 1 percent of all positive or negative tests will be false. The percentage of positive (or negative) True Prevalence Faloo PO§ltlve!l by Toot Sequence SpecificitY" results that are false depends on the true of Infection 99.5% prevalence of infection in the tested ".9% 99.99% % n %b n %b n %b population and on the sensitivity and specificity of the test. O.S 124 50.0 25 16.7 2 2.0 Consider the two examples depicted in 1 124 33.2 25 9.0 2 1.0 Figures 1 and 2. In Figure 1, the true 3 121 14.0 24 3.1 2 0.3 prevalence of infection in a population of 5 119 8.7 24 1.9 2 0.2 10 113 4.3 23 0.9 2 0.1 500 is 20 percent. The sensitivity and 20 100 2.0 20 0.4 2 0.04 specificity of the test are both assumed to 30 88 1.2 18 0.2 2 0.03 be 99 percent. There are 400 un infected persons of whom about 1 percent, or four "Teet sequence sensitivity is assumed throughout to be 99.5%. With 99.5%, 99.9%. and 99.99% specificity, people, will have a false-positive test .5%, .1 %, and .0 I % respectively I of truly unlnfected persons will be false IXlsitives. bfalse positive results as a percentage of all positive results. In calculating thh pen:~ntage, positive results result. About 1 percent, or one person, of include all truly infected persons, minus false negatives, plus false positives. the 100 infected persons will have a false­ negative test result. Thus, a total of 103 people will test positive, of whom four in the hypothetical high-prevalence the number of false positives will be will be false positives-3.9 percent of all population (Figure 1) above would be unacceptably high in populations where positive results will be false. cut in half (to 2 percent) while, in the the actual incidence of infection is very Figure 2 shows that when the true preva­ lower prevalence in population (Figure 2), low (such as persons applying for mar­ lence of infection is lower, the rate of it would be reduced by about one-third to riage licenses or positions as police false positives will increase, simply 34 percent-still a very significant officers). These researchers have calcu­ because there will be a larger number of proportion. lated that the percentage of positive results truly uninfected individuals, about 1 that will be false after the entire test Unfortunately, as noted, the Western Blot percent of whom would test falsely sequence (including the Western Blot) in as performed in most laboratories is not a positive. In Figure 2, the true prevalence very low~risk populations could be in the standardized test like the ELISA. The of infection is 1 percent in the hypotheti­ range of 28 to 90 percent. 16 definition of a positive Western Blot has cal popUlation of 500, and the percentage changed over time and remains a matter of The prevalence of mv infection in most of positive results that are false rises to disagreement. Thus, application of the prison populations in the United States is almost 50 percent. The number of false Western Blot is more susceptible to probably between 0.5 and 5 percent, with positives would continue to rise with variation and its overall performance is a few jurisdictions sharply higher. Figures increases in the size of the tested less amenable to systematic evaluation. IS released in October 1987 by the Federal population. Nevertheless, these hypothetical results Bureau of Prisons indicate that about 3 Thus far, the discussion assumes that only underscore the importance of the Western percent of Federal prisoners are infected a double ELISA test has been performed. Blot test in reducing the number of false with mV17. Figure 3 represents a hypo­ Reducing false-positive rates depends positives. In any testing program, great thetical application of mass screening for heavily on the ability of the Western Blot care should be taken to maximize quality HIV antibodies to a population of 25,000 confirmatory test to eliminate falsely control in all phases, but particularly in inmates, when the true prevalence of positive results from ELISA tests and thus the Western Blot confirmatory phase. infection ranges from 0.5 to 30 percent increase the specificity of the entire test and the specificity of the entire testing Because of the apparent susceptibility of sequence. Properly peiformed, the sequence ranges from 99.5 to 99.99 these tests (particularly the Western Blot) Western Blot is more highly specific than percent-a range encompassing the values to quality control problems, and because the ELISA. Assuming it improves assumed by most researchers. The of the dramatic effect of losing even a specificity by about 1/2 of 1 percent, the sensitivity of the test sequence is held fraction of 1 percent in specificity to such percentage of positive results that are false constant at 99.5 percent. problems, several researchers contend that

6 The percentages of positive results that meaning and reliability of the tests. In would be false under this hypothetical s: .. testing should be viewed view of the ethical, legal, and practical application of mass screening show as only one part of an overall problems that surround the use of test dramatic variations. At one extreme-- results, the purpose of any testing program 99.5 percent specificity and 0.5 percent strategy for reducing the HlV should also be carefully considered. HIV true prevalence-fully one-half of transmission, and used only antibody screening should not be consid­ confirmed positive results will be false. with adequate safeguards for ered a panacea for the problem of AIDS More than 120 uninfected inmates would confldentiality and protection in our society. Instead, testing should be mislabeled as HIV infected, with all of be viewed as only one part of an overall the potential problems associated with against discrimination. ~ strategy for reducing HIV transmission, such a designation. Even at 99.9 percent = and used only with adequate safeguards specificity and 1 percent true prevalence, for confidentiality and protection against almost 10 percent of positive results percent. Although the percentage of false­ discrimination. would be false-not an insignificant negative results is very low (about 1/10 of proportion. At the other extreme, if we 1 percent), the absolute number of false Theodore M. Hammett, Abt Associates, assume 99.99 percent test sequence negatives would pose problems if efforts Inc., is Project Director and author of specification, the percentage of positive to reduce transmission were based on several NIJ-sponsored studies on AIDS. results that would be false is extremely segregating seropositive inmates. This low regardless of the true prevalence of should be of real concern to any policy­ infection. A hypothetical percentage of maker considering HIV antibody screen­ Notes false-positive results may be calculated ing as the basis for controlling the spread easily for any scenario by substituting the 1. Centers for Disease Control (CDC), of infection in a high-risk population. "Public Health Service Guidelines for population size, estimated true prevalence Counseling and Antibody Testing to Prevent of infection, and estimated test sequence Further complicating the picture is the possibility that intravenous drug users may HIV Infection and AIDS," Morbidity and Mor­ sensitivity and specificity. tality Weekly Report 37, August 14, 1987: be particularly prone to false-positive 509-515. Again, these results demonstrate the results because of their likely exposure to importance of maximizing the specificity mUltiple viruses that may create antibodies Z. See, e.g., T.M. Hammett, AIDS in Correc­ of the test sequence and ensuring quality mistakenly recognized by the ELISA test tional Facilities: Issues and Options (3d control in all testing procedures. The as HIV antibodies. Certain other groups edition: Washington: National Institute of potentially high percentages of false­ of both high- and low-risk individuals Justice, U.S. Department of Justice, 1988), esp. Chapter 4; R. Bayer et al., "HIV Antibody positive results in low-prevalence popula­ may be similarly prone to false-positive Screening: an Ethical Framework for Evaluat­ tions also underscores the importance of results. These include women who have determining in advance how test results ing Proposed Programs," Journal of the borne more than one child, persons who American Medical Association, October 3, are to be used. In most settings, the have received blood transfusions, persons 1986; 256: 1768-1774; L.O. Gostin et al., "The actions that can be taken in response to a with alcoholic hepatitis, and homosexual Case Against Compulsory Casefinding in Con­ positive test result are limited by legal, men who have participated in receptive trolling AIDS-Testing, Screening and ethical, and economic realities. If a high anal intercourse. 2o In general, these cross­ Reporting," American Journal of Law and percentage of positive results could also cutting potential problems suggest the Medicine 1987: 12: 7-53; M.A.R. Kleiman and be false, the negative consequences of need for real caution in decisions to R.W. Mockler, "AIDS, The Criminal Justice testing may far outweigh the presumed institute any large-scale testing program. System and Civil Liberties," Governance: benefits. Harvard Journal of Public Policy, Summer­ Fall 1987j 5: 48-54. Just as low-risk populations may present a Conch.llsion serious false-positive problem, in high-risk 3. Studies by CDC and the San Francisco popUlations, the problem of false nega­ In the past year, there have been numerous Department of Public Health, reported in AIDS Policy and Law, June 15 and June 29, 1988. tives may reach fairly serious propor­ proposals to extend HlV antibody mass tions. 16 For example, in the New York screening to various new populations. 4. D.C, Des Jarlais and S.R. Friedman, State prison population of about 35,000 Many proposals have called for mandatory "Target Groups for Preventing AIDS Among inmates, where the true prevalence of screening. Despite common statements to Intravenous Drug Users," Draft, Narcotic and infection is about 17 percent,I9 about 30 the contrary, there is no test for AIDS­ Drug Research, Inc., New York, 1986. infected persons would not be identified only tests for antibodies to HIV, the virus 5. CDC, "Public Health Service Guidelines through an HIV antibody screening which causes AIDS. Decisions to institute for Counseling and Antibody Testing," program, assuming test sensitivity of 99.5 or extend HIV antibody screening must be based on a full understanding of the

7 "u.s. G,P,Q, 1988-~41-71QlilOl0) 6. For the successful application of the Field, "Elimination of Source of Error in Parenteral Drug Abusers in 1982-1983" blotted fingerprick procedure to another bio­ HTLV-ill Antibody Testing" (letter), Lancet, (letter), New England Journal of Medicine, medical test, see M.D. Garrick et al., "Sickle 1986; 1: 1032-1033. January 9,1986; 314: 117-118; N.Z. Zafar, Cell Anemia and Other Hemoglobinpathenes: "Screening for Antibodies to HTLV-III/LAV Procedures and strategy for screening employ­ 14. Barry, "Screening for HIV Infection," 1.S. in our Population: A Word of Caution," ing spots of blood on filter paper as speci­ Schwartz et al., uHIV Test Evaluation, Journal of the Pakistani Medical Association, Performance and Use," JAMA, May 6, 1988; mens," New England Journal of Medicine, April 1987; 37:84-85. 1973;288: 1205-1268. 259: 2574-2579.

7. CDC, "Public Health Service Guidelines 15. CDC, "Update: Serologic Testing for for Counseling and Antibody Testing." Antibodies to HIV," MMWR 36, January 8, 1988: 833-840, 845. Schwartz, "HIV Test 8. M.J. Barry et al., "Screening for HIV Evaluation. " For additional information on HIV Infection: Risks, Benefits, and the Burden of antibody testing and other AIDS­ Proof," Law, Medicine and Health Care, 16. Barry, "Screening for HIV Infection," K.B. related issues, contact: Meyer and S.G. Pauker, "Screening for HIV: December 1986; 14: 259-267. Can We Afford the False Positive Rate?" New • NIJ AIDS Clearinghouse, 9. Barry, "Screening for HIV Infection." England Journal ofMedicine, July 23, 1987; 301-251-5500. This Clearinghouse has publications available to the criminal 10. See, e.g., "New Specific ELISA May Beat 317: 238-241. justice community, such as AIDS in Western Blot," Medical World News, October 17. M. Specter, "Some Inmates with AIDS Correctional Facilities: Issues alld 12,1987. Virus to be Isolated," Washington Post, Options, Third Edition, and AIDS in Pro­ 11. CDC, "Public Health Service Guidelines October 24,1987, p. A14; see also Hammett, bation and Parole Services, that explore for Counseling and Antibody Testing." AIDS in Correctional Facilities (3rd Edition), testing and its impact on institutional and Figure 2.7. community corrections settings. 12. A. Ranki et al., "Long Latency Precedes Overt Seroconversion in Sexually Transmitted 18. Barry, "Screening for HIV Infection." • National AIDS Information HIV Infection," Lancet, September 12, 1987; 19. Data from a I-month blind study of all Clearinghouse, 301-762-5111. To 2:589-593. Data on the possibly recurrent incoming inmates at the Downstate Correc­ request any of the several CDC publica­ latent period were presented at the Fourth tional Facility. Presented at the Fourth Interna­ tions that the Clearinghouse is distributing, International Conference on AIDS, Stockholm, tional Conference on AIDS, Stockholm, June such as If Your Test for Antibody to the June 1988. 1988. AIDS Virus Is Positive. or Understanding 13. J. Morgan et al., "Potential Source of Error 20. Barry, "Screening for HIV Infection;" R. AIDS, call 800-458-5231. in HTLV-ill Antibody Testing" (letter), D'Aguila et al., "Prevalence of HTLV-ill Lancet, 1986; 1: 739-740; 1. Cayzer and S. Infection Among New Haven, Connecticut,

U.S. Department of Justice National Institute of Justice BULK RATE POSTAGE & FEE PAID DOJjNll Washington, D.C. 20531 Permit No. G-9\

Official Business Penalty for Private Use $300