AIDS: The Second Decade (1990)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

1 The AIDS Epidemic in the Second Decade In its first report, the Committee on AIDS Research and the Social, Be- havioral, and Statistical Sciences reviewed what was known about the distribution of cases of acquired immune deficiency syndrome (AIDS) in the United States and the behaviors that transmit infection by the human immunodeficiency virus (HIV). In addition, the committee looked at ra- tional strategies for preventing further spread of infection. Since issuing that report, the committee has continued to monitor the progression of the epidemic and the nation's response to what is now clearly seen as an evolving and enduring health problem. This report identifies several important changes that point toward new and developing issues to be addressed during the second decade of the epidemic, as well as contin- uing problems from the first decade that compel the nation's sustained attention. AIDS surveillance data indicate gradual changes in the loci of the epidemic in the U.S. population and the emergence of either new popula- tions at risk or segments within already at-nsk populations that appear to be at higher risk than was previously thought. Whereas AIDS case data provide a sense of the scope and nature of the current problem, however, data on HIV infection portend the future of the epidemic in the second decade. The future is likely to bring other changes as well, including decreased morbidity associated with HIV infection as new treatments become available. This change in the character of AIDS, with the be- ginnings of a shift away from an emphasis on the acuteness of infection toward a view of AIDS as a long-term illness, will have important impli- cations for treatment and care and for the design of intervention strategies to facilitate behavioral change, still the only available means to prevent 38

STATE OF THE EPIDEMIC ~ 39 the spread of infection. This report considers the changing face of the epidemic and highlights recent data and research on several groups or populations whose risk for HIV transmission has emerged more clearly or changed over the past decade. It also presents an update on the im- plementation of prevention approaches discussed in the committee's first report and offers recommendations on directions for the future. INTRODUCTION In its earlier report, the committee noted that the AIDS epidemic is a social as well as a biomedical phenomenon. From a biomedical perspec- tive, AIDS is a disease caused by a virus, HIV-l, that is transmitted by anal or vaginal intercourse, by exposure to contaminated blood (either through shared injection equipment associated with intravenous drug use or transfusion), and from mother to fetus. Once it is acquired, HIV infection appears to persist for life. The AIDS virus destroys subpopu- lations of white blood cells that are crucial to normal functioning of the immune system. As the immune system deteriorates, infected individuals are no longer able to ward off infections, and some develop unusual cancers and other conditions that appear rarely among individuals with uncompromised immune functioning.) A sizable proportion of the people who were infected have, after variable periods of time, progressed to severe disease and death. At present, there is no cure for AIDS and no vaccine to prevent infection. In fact, the only means available to prevent further spread of the epidemic are strategies to alter the behaviors that transmit the virus. From a social perspective, AIDS is, for the most part, a preventable disease that is inextricably rooted in the behaviors that transmit HIV. Halting the progression of the epidemic will require a better understanding of the distribution of risk-associated behaviors, the social settings in which they are enacted, and the mechanisms that facilitate change in these behaviors. Like other epidemics in the past, AIDS will leave its mark on many aspects of the societies in which it be- comes prevalent.2 In the United States, the primary target of the disease is the nation's most productive population 20- to 40-year-old adults. As discussed below, however, the epidemic is moving in ever-widening circles to reach more people and geographic areas across the country. ~ For more information on the medical and biological aspects of AIDS and HIV infection, see Con- fronting AIDS: Directions for Public Health, Health Care, and Research, Washington, D.C.: National Academy Press, 1986; and Confronting AIDS: Update 1988, Washington, D.C.: National Academy Press, 1988. 2The committee's Panel on Monitoring the Social Impact of the AIDS Epidemic is currently exam- ining certain social consequences of the epidemic and will be preparing recommendations on specific methods to monitor these effects. The report is expected to be published in 1991.

40 ~ AIDS: THE SECOND DECADE In this country, gay men still bear the burden of most of the illness re- lated to AIDS. But as the epidemic progresses and the number of persons who are at risk increases, changes in disease prevalence are becoming apparent. For example, extensive studies of gay men conducted in urban epicenters of the epidemic over a period of several years have consistently shown lower incidence rates of HIV infection, but this downward trend in new infections is not uniform across the country. Moreover, young gay men report less behavioral change to prevent infection than has been reported among older gay men, a phenomenon that leaves adolescents and young adults who engage in male-male sex at potentially increased risk. Shifts in membership in the population of men who have sex with men may also produce changes in incidence as individuals enter or leave this group. Thus, the risk of HIV infection among day men continues to be an important concern. Other populations are also feeling the effects of the disease. Increas- ingly, AIDS has become a problem of intravenous (IV) drug users and heterosexuals. Indeed, the proportion of AIDS cases in the United States attributable to heterosexual contact is growing, and a significant fraction of these cases report contact with a drug user. The rise in reported AIDS cases among females especially minority women—occasions particular concern. As more women are infected, questions concerning the horizon- tal spread of the virus (to sexual and drug use partners) and its vertical transmission (from an infected mother to her unborn infant) have become prominent in national discussions on AIDS. In this chapter, the com- mittee presents data on AIDS and HIV infection among women as an example of an emerging problem that embraces a diverse subpopulation and that is likely to require new strategies for reaching the infected and for facilitating change in risky behaviors. Other epidemiological patterns show increasing geographic diffusion of the virus. Research in the first decade of the epidemic showed pockets of high seroprevalence in such cities as New York, Newark, San Fran- cisco, and Miami for a variety of populations. More recent data show increased numbers of AIDS cases in He central region of the country, malting the epidemic less a bicoastal phenomenon and more a prob- lem that may soon directly touch more and more individuals throughout this society. Data from U.S. military applicants support this finding of increasing geographic diffusion of HIV, showing increased rates of sero- prevalence among black and white applicants from nonepidemic regions of the country.3 Yet even though some patterns have changed, others Significant increases for the past 24 months are reported in California, Florida, Illinois, Ohio, and Texas (Gardner et al., 1989).

STATE OF THE EPIDEMIC ~ 41 have endured. One pattern that has remained dauntingly constant is the disproportionate magnitude of HIV infection and AIDS among minority men and women, who are overrepresented in every transmission category. Changes are also being seen in the patterns of behavior that transmit the AIDS virus. Many gay men with AIDS acquired HIV infection through unprotected sexual contact. In addition, since the earliest days of the epidemic it has been clear that sharing injection equipment posed a risk for the acquisition and spread of HIV infection. Today, there is growing appreciation of other factors that may influence risk, including the role in risk taking played by drugs such as alcohol. Indeed, the use of noninfected drugs, such as crack and alcohol, has been linked with both high-nsk sexual behavior and HIV infection in a variety of populations. Another potential risk comes from recent increases in the use of drugs such as opiates and cocaine, which in many areas are taken primarily through smoking. These drugs offer the possibility of wider spread of HIV infection if the route of their administration should shift to injection. Getting ahead of the epidemic requires foresight to prevent infection in populations and regions that currently have a low prevalence of AIDS and HIV infection. Such opportunities should not be overlooked, for, once lost, they cannot be recaptured. The third chapter of this report considers the scope of the AIDS problem among adolescents. In it, the committee notes opportunities for intervention with this group and the characteristics of this population that put adolescents at risk and that require consideration in the design of intervention strategies. Keeping pace with the epidemic requires perseverance an enduring, long-term commitment to surveillance, research, and improved interven- tion. Short-term approaches that result in one study, one report, one intervention will not suffice. The type of commitment required is one that includes a range of strategies and techniques, behavioral as well as biomedical. The response of the U.S. blood supply system to the threat of AIDS, which is discussed in Chapter 5, offers an example of such a commitment. The blood supply system has combined biomed- ical techniques (e.g., HIV blood screening) and behavioral approaches (e.g., self-exclusion procedures for high-risk donors, training programs for physicians to alter transfusion-related practices) to reduce dramati- cally the risk of ~ansfusion-related transmission and the number of cases associated with exposure to contaminated blood products. Yet despite the success these reductions represent, problems remain. Although the current prevalence of HIV infection in the general population is believed to be Tow, as the virus spreads and more people become infected, some parts of the country may find that the pool of uninfected, "safe," and

42 ~ AIDS: THE SECOND DECADE willing blood donors has shrunk to a point that could compromise the adequacy of the blood supply. The second decade of the epidemic thus brings continuing as well as new challenges in managing the effects of HIV infection and preventing its further spread. In its first report, the committee commented on the need for sys- tematic improvement of intervention strategies to facilitate change in the behaviors known to transmit HIV. It noted in particular that the interven- tion efforts described in that report had not resulted in data that offered a clear sense of which strategies worked best for specific subpopula- tions. The paucity of planned variations of intervention strategies and of information garnered through rigorous outcome evaluation that was cited by the committee has not been remedied. The gap is particularly problematic for innovative programs that have targeted groups such as {V drug users, in which infection has been shown to spread at an alarmingly fast rate in the absence of effective intervention efforts and in which potentially efficacious but politically sensitive efforts have come under attack. A year ago the committee recommended well-designed, carefully evaluated pilot tests of sterile needle exchange programs to ascertain the effectiveness of this strategy in preventing further infection among IV drug users. Although several programs were initiated, there are now crippling restrictions on the use of federal funds to support and evaluate these activities.4 This chapter relies extensively on counts of diagnosed AIDS cases reported to the Centers for Disease Control (CDC). In its last report, the committee noted some of the pitfalls that cause undercounting or misclassification of AIDS cases and deaths (Turner, Miller, and Moses, 1989:32-33~.5 It recommended that both vigilant quality control and spe- cial methodological studies be undertaken to improve our understanding of the errors and biases that affect the reporting of AIDS cases and deaths. The committee finds Hat the need for such methodological work continues, and it commends the Public Health Service for its plan to give pnonty to the funding of such research (PHS, 1988; CDC, 19891. As we embark on a second decade of understanding and coping with the AIDS epidemic, these efforts will provide needed attention to the quality of this key data system. More than 115,000 cases of AIDS have been reported in the United 4See Public Law 101-166 [H.R. 3566], Title General Provisions, Section 520. Depar~nents of La- bor, Health and Human Services, and Education and related agencies appropriations, 1990. November 21, 1989. Washington, D.C. 50ther reports have also looked at completeness of reporting of AIDS cases. See, for example, the estimates of Conway and colleagues (1989) and Modesitt, Hlllman, and Fleming (1990).

STATE OF THE EPIDEMIC ~ 43 States, and more than half of the people reflected by this statistic have died. Moreover, AIDS has become the leading cause of death for 30- to 50-year-old males and 20- to 40-year-old females in New York City (New York State Department of Health, 19891. Despite advances in the medical treatment of AIDS and other sequellae of HIV infection,6 it is likely that HIV and AIDS will remain significant threats to individual and public health for the foreseeable future. The persistence of risk in the environment and the potential for relapse among those who have taken protective action mean that, as we enter the 1990s and the epidemic's second decade, it is important to reaffirm the nation's commitment to preventing further spread of infection. To be more successful in future prevention endeavors than we have been in the past, we must commit talent and resources to the rigorous outcome evaluation of well-designed interventions while developing new approaches based on current under- standing of the evolutionary and persistent aspects of the problem. The task of devising intervention strategies to fulfill this commitment rests on obtaining sound information about the populations at risk and the behav- iors transmitting the virus to guide program planners and policy makers. The section that follows discusses the current profile of the disease in the United States on the eve of the epidemic's second decade. THE CHANGING EPIDEMIOLOGY OF AIDS IN THE UNITED STATES As of December 1989, the CDC reported 117,781 cases of AIDS in the United States.7 Of the adult and adolescent cases, a significant majority (70,093) are attributed to male homosexual or bisexual contact, and an additional 8,117 are ascribed to bow homosexual contact and rv drug use (CDC, 1990a). Of all men with AIDS who report same-gender sexual contact, more than one-quarter (27 percent) are black or I-atino~ (CDC, 1990a); moreover, homosexual contact is the predominant AIDS risk category among black and Latino males, accounting for 45 and 47 percent of cases, respectively (CDC, 1990a). 6GaiL Rosenberg, and Goedert (1990), for example, recently presented evidence suggesting that pro- phylactic use of zidovudine among seropositive homosexual and bisexual men (particularly in New York City, San Francisco, and Los Angeles) may be responsible for the lower than expected incidence of AIDS cases reported in this population since the middle of 1987. 7 although the degree of underreporting of AIDS cases is not known with any certainty, recent studies have judged reporting to be between 83 and 100 percent complete (Chamberland et al., 1985; Hardy et al., 1987; Lafferty et al., 1988). A recent survey of hospital discharge billing records in South Car- olina, however, found that only 59.5 percent of cases meeting CDC diagnostic criteria were reported; underreporting was significantly worse among black patients than among whites (Conway et al., 1989). 8In this report the teIms Latino and Latina are used interchangeably with Hispanic.

44 1 o oG c) := o x In ce .— o A: ~ - o so En ED c) of en - ~: in - c~ cO c) c~ ~ 'e - Do o o - . - . - m - o ~ O EN ~ ;^ = - C~ ~ ._ O Ct ~ ._ X C S~ O C) .0 O ec ~ X X U O Ct £ - X =o 'o + a O ~ £ X o ~ U. ._ ~ Cal o o ~ ~o as {C ~ a 00 _ C~ ~ ~ ~ ~ ~ X ~ — X X 00 00 00 X 00 00 X ~ C~ Ch 0\ ~ ~ ~ Cx O _ ~= C O — C~0 ~ 0 :~ e~ £ · = =; =, . ~ x ~_ 0 _ cL ~ ~o 3 c~ ~ ~ £ =~ ~ O CC 0 3 — C' ~ — o < - D C ~ ~ ~ o ~0 ~ e =, `., s -, c~ ~ s _ =, 0 ~ c' acJ=o~o~ v, s E = ~ E ~ ~ c~s ~ c' c~ n e ~ ° O j=` A 2 ~ S O ~ e-- ~ S C C.) ° X~ ~ = ~ c u .= ~ °` ~ = e O ~ ~0 ._ ~ 2 =. O ~ ~ ~ . c, ~ s . s ;? E O == e E .~ c!5 ~ ~ '~; ~ c~ c, c,, ~ C,~>S3 £ e s O ~ at e D~ O 0 = = ' O2 - , ~ ~ 3 e O oo - - c~ D C) C) C) a s 0 o s o c~ G) u: c~ v' - 6 - c, v, e~ o - o - c) ~ - - D U' a 6 V o c, : - o V . . o

STATE OF THE EPIDEMIC ~ 45 Data from the Multicenter AIDS Cohort Studies (MACS) and the San Francisco Men's Health Study suggest that incidence rates among gay and bisexual men have stabilized (Winkelstein et al., 1987, 1988), and there is some evidence that the number of new cases of AIDS in this population appears to be leveling off in New York City and Los Angeles (Berkelman et al., 19891. Indeed, the proportion of all AIDS cases attributed to male homosexual and bisexual contact has dropped from 64.9 percent in 1981 to 55.3 percent in 1989.9 At the same time, the proportions of cases attributable to heterosexual contact and to IV drug use have grown. In 1981 cases attributable to heterosexual transmission and {V drug use constituted 0.5 and 11.0 percent, respectively, of all reported cases; by 1989 these rates had increased to 5.0 and 23.2 percent (Table 1-1~. Among IV drug users, after several years of progressively worsening infection statistics, there is now evidence that HIV seroprevalence has stabilized in some areas, such as New York City (Des Jarlais et al., 1989; Stoneburner et al., 1990), San Francisco (Moss et al., 1989), Detroit (Ognjan et al., 1989), Amsterdam (van Haastrecht, van den Hoek, and Coutinho, 1989), and Stockholm (Olin and Kall, 19891.~° However, this stabilization reflects lower rates of new cases of HIV infection within a dynamic population and should not be confused with elimination of viral transmission. In fact, the evidence of stabilization in some regions is offset by data on the recent, very rapid spread of the disease in other areas. The proportion of cases attributed to drug use is increasing in Louisiana (Atkinson et al., 1989) and Maryland (Horman and Hamidi, 1989), as well as throughout cities in the northeast region of the United States (see Figure 1-11. The population of IV Hug users is mobile and quite capable of carrying the virus from one geographic area to another (Comerford et al., 1989, McCoy, Chitwood, and Page, 19891. Such mobility is one factor that can contribute to the possibility of rapid rises in seroprevalence rates in venous parts of the country. There is some evidence that homosexual and bisexual IV drug users may play a role in introducing HIV infection to other IV drug users in areas with 9Decreasing proportions, however' do not indicate decreasing numbers of cases. This smaller propor- tion of an ever-increasing whole still continues lo result in increasing morbidity. 10For some groups at highest risk for HIV infection, stable rates may signify that all vulnerable indi- viduals are already infected, a phenomenon known as saturation. This phenomenon does not appear to explain stable rates in New York City, however, where approximately 50 percent of IV drug users are estimated to be infected. If saturation had occurred, one would expect to see higher seroprevalence rates. For example, approximately 90 percent of IV drug users in New York City have been infected with hepatitis, a virus that is transmitted in the same manner as HIV and appears to have reached saturation in this population.

46 1 AIDS: THE SECOND DECADE 70 _ 60 __ ~ 50 an an C] o it G ILL 40 30 20' it 10 I- Homosexual- · Bisexual - IV Drug Use O 1 1 1 1 1 1 1 1982 1983 1984 1985 1986 1987 19~ 1989 YEAR FIGURE 1-1 Percentages of reported AIDS cases among men (aged 13 and older) in northeastern cities that have been attributed to homosexual contact or IV drug use, 1982-1989. (Cases reported as both IV drug use and homosexual/bisexual contact are not shown.) NOTE: Cities in the Northeast region include the OMB Metropolitan Statistical Areas (or New England County Metropolitan Areas) of Bergen-Passaic, N.J.; Buffalo, N.Y.; Boston, Mass.; Hartford, Conn.; Nassau-Suffolk, N.J.; New York City; and Newark, N.J. SOURCE: Tabulated from CDC's AIDS Public Infonnation Data Set for AIDS cases reported through December 31, 1989. low prevalences of HIV infection (Battjes, Pickens, and Amsel, 1989). Although the distribution of cases in the IV drug-using population has shown considerable variation, one constant remains: the burden of AIDS related to IV drug use falls most heavily on minorities (Selik and Petersen, 19891. Nearly half (47.6 percent) of all males reporting IV drug use as their only risk factor for AIDS are black, and 32 percent are Latinos (CDC, 1990a). Addressing the problem of IV drug use in relation to AIDS remains a high priority, but a focus solely on drug injection will no longer suffice to prevent further spread of HIV infection among individuals who use drugs. The evolving nature of the epidemic also requires a broadened perspective to encompass what appear to be new patterns of risk associated with drug use. In particular, the emergence of "crack" cocaine use as a risk factor for the transmission of HIV infection is a new and disturbing development in the epidemiology of AIDS in the United States. The risks associated with the use of crack are indirect (crack use is not in itself a mode of HIV transmission). However, the tendency of those who use crack to engage in unsafe sexual activity offers the potential for viral transmission.

STATE OF THE EPIDEMIC ~ 47 Unlike other forms of cocaine that are snorted or injected, crack is smoked.l1 It can produce a rapid, intense effect that for most users is a feeling of euphoria, including an increased sense of self-worth and power. The effect of the drug wears off rapidly, however, and may be followed by a craving to use more. As a result, some users will "binge," using crack every 10 to 20 minutes until they or their supply of drug is exhausted. Thus, the intensity and transience of the crack "high" encourage the development of dependence on the drug. For many males, crack is reported to produce intensified sensations of sexual arousal and sexual pleasure; indeed, in some places crack is marketed by highlighting the sexual effects of the drug. Some females who become dependent on crack show a willingness to exchange sex for the drug or for money to obtain the drug. (Chapter 4 discusses at greater length the relationship between crack use and female prostitution.) It is possible, of course, for crack users to practice safer sex, including condom use, but the effects of the drug and the perceived exigencies of repeating those effects reduce the likelihood that safer sex will be practiced consistently (Friedman et al., 19881. Small surveys of crack users have indicated an association between crack use and sexually transmitted diseases (STDs) (Fullilove et al., 1989), which supports the hypothesis that unprotected sex occurs in such settings. Because STDs appear to be cofactors in the acquisition of HIV infection, this association bodes ill for preventing the sexual transmission of HIV. Indeed, surveys of patients at STD clinics in New York have already found individuals with HIV infection whose only behavioral risk factor appears to be sexual activity associated with crack use (Chiasson et al., 1989; Hoegsberg et al., 1989~. The rapid emergence of crack as a risk factor for HIV infection highlights the need to be vigilant for signs of new patterns of nsk. Thus, the committee recommends that the Public Health Service establish mechanisms across its agencies for rapid identification and assessment of the relationship of new drug use problems to the spread of HIV. But changes in drug use are not the only new trends that are i 1 Usually, cocaine is sold as a hydrochloride Set; in this form it must be either "sniffed" and absorbed through the nasal mucosa or injected. To produce crack cocaine, the hydrochloride is removed; in this form the cocaine may be vaporized and thus inhaled. Inhaling any drug into the lungs permits it tO be absorbed into the blood stream rapidly and transported to the brain. Chemically, smoking crack is identical to "freebasing" the hydrochloride salt form of cocaine. Previously, persons who wanted tO freebase had to purchase cocaine hydrochloride and convert it themselves to the base form through a complicated and dangerous process that involved heating volatile chemicals that could easily catch fire. The rise in the use of crack has come from the development of new and safer methods for removing the hydrochloride so that the drug can be sold in its base fonn. 12 Some crack houses (the locations at which the drug is sold and used) employ women to provide sex to customers.

48 ~ AIDS: THE SECOND DECADE becoming apparent at the end of the first decade of the epidemic. There are also noticeable shifts in the groups being affected directly by the disease. Throughout its studies, the committee has been struck by the diversity of the populations now represented in the AIDS statistics. It has chosen to look at women and the risks HIV poses to them as an example of a heterogeneous population that will require considerably more attention in the second decade of this epidemic than it received in the first. Because most of the cases of AIDS reported in the early years of the epidemic were diagnosed in men, few women appreciated their potential risk of acquiring this disease. Yet as the epidemic has progressed, women of all races have begun to account for a greater proportion of cases. The majority of women represented by these statistics have a history of intravenous drug use; a subset report no drug use themselves but indicate a sexual relationship with an intravenous drug user in their risk profile. The range of women at risk for AIDS, however, goes beyond those involved through drug use. The sections that follow discuss the profile of risk among women in the United States, noting the varied subgroups of this population that are increasingly affected by the epidemic. A PICTURE OF EMERGING RISK: THE AIDS EPIDEMIC AMONG WOMEN Early in the epidemic, vertical or pennatal transmission of HIV infection (transmission from an infected pregnant woman to her offspring was of particular concern because the majority of female AIDS cases (75.4 per- cent) had been diagnosed in women between the ages of 20 and 39, prime childbearing years. Today, concerns about perinatal transmission must be joined with the recognition that the risk of HIV infection among women of all ages is increasing, and more and more women are confronting the disease in their own lives or in the lives of those around them. This increasing risk can be seen in the epidemiological data in Table 1-2, which show rising numbers of female AIDS cases over the past decade. By December 1989, a cumulative total of 10,611 cases of AIDS had been reported in women 13 years of age or older; however, this figure is likely to underestimate the scope of the problem owing to a failure to report cases and reporting delays. The upward trend shown in Table 1-2 is particularly ominous for minority women, who are disproportionately represented in almost every risk category (Table 1-31. In 1989 women also accounted for a larger proportion of all AIDS cases diagnosed in this

STATE OF THE EPIDEMIC ~ 49 country than they did in the early years of the epidemic, a temporal trend that is most striking in northeastern metropolitan areas (Figure 1-21.~3 Unlike men, among whom the majority of cases have been ascribed to same-gender sexual contact, the major transmission category for women with AIDS is {V drug use (see Table 1-2) (Guinan and Hardy, 1987; CDC, 1990a). As Table 1-2 shows, there has been a slight decrease in the proportion of female cases attributed to IV drug use since 1983, but this drop is offset by the doubling of the proportion of female heterosexual cases. AIDS case data show that women are at greater risk than men of acqu~nng infection through heterosexual contacti4 and that the proportion of U.S. AIDS cases attributable to heterosexual contact is growing.is A majority of female heterosexual cases, however, are related to sexual contact with an infected intravenous drug user.l6 Thus, women are at risk for HIV infection both directly through their own drug use and indirectly through sexual exposure to partners who inject drugs. Almost a thousand women with AIDSi7 (N = 934) have reported 13The proportion of all cases reported among women has grown from approximately 6 percent in 1982 to roughly 10 percent in 1989; accordingly, the male-to-female ratio of cases has dropped from 14.7:1 in 1982 to 8.6:1 in 1988. As the data in Table 1-2 show, the decline in the male-to-female ratio has been steady over the course of the epidemic and represents a statistically significant change (p < .001). Data from New York City also find a significant decline in the male-to-female ratio over time (Stoneburner et al., 1990). l4By December 1989, almost one-third of female cases but only 2 percent of male cases were attributed to heterosexual contact with an infected partner (CDC, 1990a). Unfortunately, these data probably underestimate the risk associated with heterosexual transmission, in part because of case classification criteria. In a case in which heterosexual risk is found in combination with some other risk factor, the case will be classified in the other risk category. Moreover, unless the index partner in a case is reported to have AIDS, to be at increased risk for HIV infection, or to be infected with HIV, the case will not be ascribed to heterosexual transmission but will be categorized as having no identifiable risk factor. Castro and colleagues (1988) reported that a significant portion of AIDS cases with no identifiable risk factor are likely to be heterosexually acquired, a conclusion supported by relatively high rates of other sexually transmitted diseases among cases in this category. 1SIn 1982, only 1.1 percent of cases were associated with heterosexual contact; by December 1989 the proportion had grown to 5 percent (Curran et al., 1988; CDC, 1990a). 16More than 67 percent of heterosexually acquired AIDS cases in women are ascribed to unprotected intercourse with an infected male intravenous drug user (see Guinan and Hardy. 1987; Curran et al., 1988). 17 Exposure to contaminated blood accounts for 3 percent of all AIDS cases. However, 10.3 percent of AIDS cases among women are related to transfusions, blood products, or tissue transplant whereas only 2.7 percent of cases among men are so categorized. The greater number of blood-related AIDS cases among men (N = 2,417) reflects both a greater burden of illness born by men to date and a greater prevalence of coagulation disorders found in males. Hemophilia, a complex of different co- agulation disorders, is a sex-linked hereditary disease that affects the body's ability to produce the various proteins needed to clot blood. Although a small subset of women have coagulation disorders, men constitute the vast majority of hemophiliacs.

so o ;~ - o ce c~ u: c~ c~ c) c~ u, - ~; s: o c) 'e c) c) - cq u, c~ c) o o >~ ~D oD - ~ av - o ~ c m ~ C~ U: ._ C~ o Ct ._ a 2 _ o o o o o o o o o o o o o o o o o o C~ ~ ~ oo ~ oo ~ ~ o. o. ~ ~ ~ —, ~ ~ , O 5 ~. _ ~ a~ V~ O O _ _ C~ U) o _ _ _ _ Ct o _ C) _ X o oe ~ V o C) C~ o c x ~ C~ o ~o o ~ C~ _ _ ~ _ ~ o O - _ ~ o ~ _ o oo C~ — . . . . . . . . . ~, C~-} (~ _ _ _ o O — (~ — ~t d- ~ O —) C~ — . . . . . . . . . O O O o o _ _ _ _ O ~D — ~ 5\ — oo C-l ~— . . . . . . . . . — — C~ ~ ~ ~ C~ + X _ C~ _ o :> o — ~ - — ee o ~ o ~ ~ ~ - c~ ~ ~ ~ c~ ~ ~ ~ ~ O. . . . . . . _ C~ oo oo _ ~ — O - . C) ~ U~ ~ V~ ~ ~ O O ;> ~ ~ ~ _ ~ U~ O oo ~ C~ O C~ ~ ~ ~ t— ~ ~D OO~ ~ — C) ~ t— oo ON C oo oo oc ~ C~ oo oo CS~ ~ CSx _ _ _ — — — — ;~.

0 ~ r~ ~ ~ ~ ~ ~D — oN ~ ~ ~ t - c~ ~ ~ ~ o ~ ~ ~ - - - - - ~ ~ ~ o - o o o o o o o o o o o o o o o o o o - - - - - - - - - ~ — d ~— \~0 —] ~ ~ ~ t~, . . . . . . . . ] ON ~ ~ o o o - - ~ c~ oO ~ ~ ~ C7- 0 - } r - - ~ ~ ~ o o - ~ t - o ~ ': ~ ~ ~ 6 6 6 ~ Z Z Z Z Z Z Z Z Z o ~ — ~ ~ _ o oo ~ . . . . . . . . . o o ~ ~ _ ~ _ ~ _ 6 6 ~ 6 ~ 6 ~ <5 ~ Z Z Z Z Z Z Z Z Z U~ ~ ~ ~ ~ ~ ~— oo Ch~ C~ oo oo oo X ao X oo oo Ce ox Ch C~ C) C,) _ _ _ _ _ _ _ _ >` C) ~ = _ ~ 6 C) ~L Ct ~ ~ ~ ~ ~ D 5:: ~ ·— o ;` V ~ ~ ~ _ o ~ t.) D ~ cc o~ ~ ce $_,, 3 O ~ ~ ~ .c~ ~ c~ C C) E~ ~ ·— _ O O ~ _ C~ ~ _ 6 .=, ~ c, ~ e e ¢, O V e O ~ ~ _ ~ E ~ ~ ~ ~ ° _ o ° 3 ~ ~ ~ ~ .= ~ ~ e _ _ _ . _ 3 ~ 3 O c' ~ ~ c~ E~ o ·- tts S <~ O ~ D ~ C: ~ ~ ~ ~ _ ~ co c~~ ~ ~ :;s u~ ~ ~ D ~ X 0 0 c~ ~ ) ~ 0 ~ c) {~ O ~ ~ c~ ~ 5 0 ~ ~ ~ <:s D ~ 0 5 - S ,0 ~ ~ t— <t S ~ :^ O D e~ ,_ C) ~ {~S ~ ~ c~ ~ s3 ~ ~ Z ~ c~ ._ O c~ ;^, ~ ·— ~C5 ~ ~ =,5\ ~ . - ~ ~ ~ ~ ~ ~ £~ o.s~ c~ ~ c~s, ~ ~ ~._ c, ,`0 >, · 00~ ~ ~ v~ 0 v v' ,_, - c~ ~, ~0 c' 3 0D c~ ~0 c~ - ~. oo - c' D £ c' c; a o - o C) C~ C) U: C) U) - ~: - C) U, Ct a 51

52 ~ AIDS: THE SECOND DECADE 18 Oh Oh co ~ Z O O ~ ~ C) Z ~ ~ O 16 14 12 10 8 6 2 _ O .._ 1 1 982 Northeastern Cities __ - - __ . _ _ _ National , _ _ -~~ Middle Atlantic Cities l J 1983 1984 1985 1986 1987 1988 1989 YEAR FIGURE 1-2 Percentages of reposed AIDS cases diagnosed among females, 1982-1989. (Tabulation includes only cases diagnosed among persons aged 13 and older.) NOTE: Cities in the Northeast region include the OMB Metropolitan Statistical Areas (or New England County Metropolitan Areas) of Bergen-Passaic, N.J.; Buffalo, N.Y.; Boston, Mass.; Hartford, Conn.; Nassau-Suffolk, N.J.; New York City; and Newark, N.J. Cities in the Mid-Atlantic region include Baltimore, Md.; Norfolk, Va.; Philadelphia, Pa.; Pittsburgh, Pa.; and Washington, D.C. SOURCE: Tabulated from CDC's AIDS Public Lnfonnation Data Set for AIDS cases reported through December 31' 1989. . exposure to contaminated blood as their only risk factor. This route of transmission accounts for only a small proportion of all AIDS cases, but it accounts for a greater proportion of AIDS cases among women than among men, a finding that is related to gender differences in the distribution of cases across transmission categories. Because all blood donors are now screened for risk factors and all rlonntinn~ urn. ~rrePn-A for antibodies to HIV, the decline in transfusion-related new infections that has been seen over the course of the epidemic is expected to continue. However, there remain a number of AIDS-related blood supply issues that are discussed in more detail in Chapter 5 of this report. Although the number of new AIDS cases related to contaminated blood is expected to continue to diminish, the impact of past exposures on the immediate future is uncertain because no one known the n,~mh~r ~ ^,_ ~^ ~4~= ~ ~ O~l ~~~ ~ . .. . . . at ~nc~v~0uals who are already infected with HIV. Unfortunatelv dome =~;~+~O ^~+ +~+ ~ ~~ ~ 1 _ ~- ~ ,~ ~ aLliliaL~ ~U~L U1~t ~ bUUSt~l~ld1 POH10n OI me hemophiliac population is infected. ]8 Women who have sexual contact with Moe hemophiliacs He 18Of the estimated 20,000 hemophiliacs in the United States (National Hemophilia Foundation, 1988),

STATE OF THE EPIDEMIC ~ 53 thus at indirect risk of acquiring HIV infection. (Because very few women have coagulation disorders, they have little direct risk of blood-related infection.) Small surveys of female sexual partners of hemophiliac men have found between O and 21 percent to be infected (Kim et al., 1988; Ragni et al., 1988; Smiley et al., 1988; Lawrence et al., 1989;19901. In addition, women who have received needle-stick injuries while helping to administer clotting factor treatments to an infected male partner may also be at risk (Smiley et al., 19881.~9 The AIDS case data thus present a picture of emerging risk for women, one that is strengthened by recently available seroprevaTence data on HIV infection ire this population. CDC and the National Institutes of Health (NIH) are currently conducting a neonatal survey throughout the United States that takes a small part of the dried blood samples collected from hospital-loom babies for metabolic disease testing and tests them for antibodies to HIV. Because newborns carry the mother's HIV antibodies (if present), seroprevalence rates derived from this survey can be projected to the population of childbearing women.20 Forty-four states, as well as the District of Columbia and Puerto Rico, currently participate in statewide surveys of newborns. Thus far, preliminary data have been aggregated for 18 states and territories in four regions of the country; state-specific data have also been reported for selected states. Rates of infection are highest in the Northeast and in the southern region and approach 0.7 and 0.5 percent, respectively, of cases among childbearing women in those areas.2i Data aggregated at this level, how- ever, do not clearly reflect the considerable variation in rates across the country. For example, pockets of "alarming HIV infection"22 are found Jackson and colleagues (1988) speculate that 15,000 may be infected with HIV; Mason, Olson, and Parish (1988) have estimated that 92 percent of people with hemophilia A, the most common form of the disease, are infected. 19In a survey of 32 hemophiliac couples, one woman who seroconverted (i.e., her antibody status changed from negative to positive) reported eight needle-stick injuries she had incurred while helping her spouse with clotting factor treatments (Smiley et al., 1988). 20It is important to note, however, that not all infants found to have antibodies for HIV will remain antibody positive. Moreover, not all women at risk for HIV infection will be captured by this survey. Specifically, this group will not include women who choose not lo have children, those who cannot have children, women who choose abortion or who miscarry prior to delivery, and postmenopausal women (Turner, Miller, and Moses, 1989). 21 Seroprevalence rates ranged from 0.03 to 0.66 percent nationally and regionally as follows: 0.03 to 0.66 percent for different locations in the Northeast, 0.09 to 0.46 percent in the South, 0.06 to 0.09 percent in the north central region, and 0.03 to 0.04 percent in the mountain zone (Pappaioanou et al., 1989). 22 In his article, "One in 61 Babies in New York City Has AIDS Antibodies, Study Says," Bruce

54 ~ AIDS: THE SECOND DECADE among women of childbearing age in New York City; the rate of infec- tion among women giving birth in this city is 1.7 percent.23 Palm Beach County, Flonda, has reported rates of slightly more than 1 percent among women delivenng infants, but the rate in Duval County (which includes Jacksonville, Flonda) is very much lower (0.28 percent) (Withum et al., 19891. Both AIDS case data (see Table 1-3) arnd neonatal seroprevalence rates reflect the differential risk of AIDS among minority women.24 Among births in New York City, racial and ethnic trends are spiking: 2.17 percent of black, 1.46 percent of Hispanic, arid 0.39 percent of white newborns were found to have antibodies to HIV (Novick et al., 19891.25 Disproportionate rates of infection among black arid Hispanic women have been seen in other parts of the country as well.26 An association has been noted between IV drug use and the risk of HIV infection for a substantial portion of pregnant women and their babies.27 In New York City, the neonatal seroprevalence rate was 2.2 Lambert reviewed the New York neonatal data for the New York Times (January 13, 1988, pp. Al and B4). The state health commissioner was quoted as saying "results in the first 19,157 newborn blood specimens demonstrate an alarming HIV infection rate among women of childbearing age in New York City.'' 23 Data from 276,609 consecutive births that occurred in New York between November 1987 and November 1988 showed substantially higher rates of HIV infection in New York City (1.25 percent) than in upstate New York (0.16 percent) (Novick et al., 1989). Moreover, there is substantial variation in infection rates within New York City. The highest rate was found in the Bronx (1.7 percent), fol- lowed by Manhattan (1.65 percent), Brooklyn (1.31 percent), Staten Island (0.68 percent), and Queens (0.62 percent) (Novick et al., 1989). A suIvey of 3,556 women giving birth and having abortions in New York City in 1987 and 1988 found that rates for both groups had stabilized (i.e., the rates demon- strated statistically insignificant fluctuations over this period of time) (Araneta et al., 1989). In this study, women between the ages of 25 and 29 had seroprevalence rates of 1.89 and 2.17 percent (for 1987 and 1988, respectively). Among Medicaid recipients, the rates for the two years were 3.36 and 2.31 percent. 24Black women account for 51.7 percent of all AIDS cases diagnosed among women, and Hispanic women account for an additional 20.0 percent of cases. However, blacks and Hispanics constitute only 11 percent and 8 percent of the U.S. population, respectively (U.S. Bureau of the Census, 1987). 25Statewide rates of infection in New York show similar disproportions: 1.8 percent- for black, 1.3 percent for Hispanic, and 0.13 percent for white newborns (Novick et al., 1989). 26Data from 65,007 infants born in Florida between July and December 1988 indicated that 1.2 percent of black babies were seropositive, whereas only 0.17 percent of white infants had circulating HIV antibody (Withum et al., 1989). In Texas, black infants accounted for 48.1 percent of all positive tests even though blacks constituted only 12.8 percent of all specimens tested (Thompson et al., 1989). Hispanics were not differentially represented among positive specimens from Texas: 14.8 percent of seropositive infants were Hispanic, but 28.3 percent of the specimens came from Hispanic babies. 27 Among female AIDS cases attributed to heterosexual contact, 51 percent of whites, 57 percent of blacks, and 83 percent of Latinas reported unprotected sex with an IV dmg user as their risk factor

STATE OF THE EPIDEMIC | 55 TABLE 1-3 D~stnbution of AIDS Cases Reported Through December 31, 1989, Among Women Aged 13 Years and Older by Exposure Category and Ethnic Group Exposure Category White Black Hispanic Others Total IV drug use N 1,175 3,171 1,110 35 5,491 % 21.4 57.7 20.2 0.6 100.0 Heterosexual N 781 1,297 769 36 2,883 % 27.1 45.0 26.7 1.2 100.0 Blood N 733 213 119 25 1,090 % 67.2 19.5 10.9 2.3 100.0 Pattern IIb N 1 433 2 3 439 % 0.2 98.6 0.5 0.7 100.0 Other nskC N 200 374 119 15 708 % 28.2 52.8 16.8 2.1 100.0 Total cases N 2,890 5,488 2,119 114 10,611 % 27.2 51.7 20.0 1.1 100.0 aThis category includes Asians and Pacific Islanders, American Indians, Alaskan natives, and persons whose race or ethnicity is unknown. bPattem II refers to individuals immigrating from those countries in central, eastem, and southern Africa and some Caribbean countries where the majority of cases are ascribed to heterosexual transmission. The male-to-female ratio is approximately 1:1 in these countries, pennatal transmission is more common than in other areas, and intravenous drug use and homosexual transmission occur at a very low level. CThis category includes cases currently under investigation for which no history of exposure has yet been reported and cases for which no exposure mode could ever be detennined. SOURCE: Computed from CDC's AIDS Public Information Data Set for AIDS cases reported through December 31, 1989. percent for ZIP Code areas with drug abuse rates (as determined by drug- related hospital discharge data) in the top quartile of the distribution for the city;28 the rate ire the remainder of the city's ZIP Code areas was 0.8 percent (Novick et al., 1989~. In New Jersey, the rates of seropositivity (CDC, 1990a). In a survey of 165 couples, black and Latina women were more likely than whites to report an intravenous drug user as their index case (Padian et al., 1989). 28 Indeed, more than half of the seropositive newborns resided in ZIP Code areas that fell within the top quartile of the drug use statistics for the city (Novick et al., 1989). . .

56 ~ AIDS: THE SECOND DECADE among newborns were highest in the two counties with the highest AIDS case rates and the largest portion of cases related to IV drug use (~.58 and 1.27 percent in the two counties versus a statewide rate of 0.49 percent) (Altman et al., 19891. Smaller serologic surveys show geographic variability in the distri- bution of HIV infection, a pattern that lends additional weight to the evidence provided by actual AIDS cases. Specifically, pockets of high seroprevalence rates and the strong association between HTV infection and IV drug use have been reported by large and small serologic surveys as well as national surveillance data on AIDS cases (see Table 1~. Five of the six studies in Table 1-4 that report seroprevalence rates of 10 per- cent or greater are based on samples of female intravenous drug users.29 It is difficult to draw conclusions from these studies, however, because of venous methodological constra~nts.30 Seroprevalence studies of female applicants for military service and women on active and reserve duty in the U.S. Army provide another view of the pattern of infection in a large population of women.3i Since October 1985, 383,112 women have been tested by the military for antibodies to HIV. The crude seroprevalence rates in these groups were 0.7 percent for female applicants, 0.9 percent for women OI1 active duty, and 0.8 percent for women in the reserves (Cowan et al., 1989; Horton, Alexander, and Brundage, 19891. In each group, the seroprevalence rates among teenage females were comparable to those for teenage males. Gender parity in infection rates for this population is a very different pattern than that seen 2930hnson et al. (1987); Chandrasekar, Matthews, and Chandrasekar (1988); Brown et al. (1989); Schoenbaum et al. (1989); and Wolfe et al. (1989). 30For further discussion of methodological problems related to convenience samples, see Turner, Miller, and Moses (1989). Comparisons across these studies are also problematic because samples were recruited from different organizations and locations, and each sample displayed considerable heterogeneity with respect to age, race, and risk behaviors among the women enrolled in the studies. Variable participation rates and potentially significant differences between participants and nonpartic- ipants make these results extremely difficult to interpret. For example, in a study of 7,400 women who delivered babies at a Miami hospital between July and December 1988, 93 percent consented to HIV testing, and 2.3 percent of those who consented were found to be infected (O'Sullivan et al., 1989). Of the 274 women who refused testing, blinded blood specimens were assayed, and the infection rate for this group was found to be 4.9 percent. This substantial difference in prevalence rates between consenters and nonconsenters indicates an important source of bias and calls into question the accu- racy of estimates that have relied exclusively on volunteer samples. It should be noted however, that the direction of the bias has not always resulted in higher rates in blinded samples. See, for example, Chiasson and colleagues, 1990. 31 It should be understood, however, that this population is not representative of the national female population.

57 Cq ._ - - U) ~5 C) ._ Ct c o 3 C) Cal A: ._ I: V o Cal CONS ~2 ~:5 ._ so C) 3 04 o Cal en PA sol C) - o 1, Cal V) 1 - Cal — 5 :> ~ C`: _ ~ . . O O Z Z; c: + C) _ _ ~ := ._ - - CD V: ~:5 o . - C) Cal - V, Ct C~ s~ o ; - ._ C~ V) "o ~ ~_ ~ oo , _ (S~ X ~ ^ ~ C :'G _ ~ X ~ — X ~ ~ ~ X — {L) ' C~ — C; ~ ~ ~ ._ ° '~ ° %, ' ' ' ~ _1 _ [_ ~ C~ ~ oo oo ~ ~ ~ 0 \0 oo t— ~ ~1 ~ ~ C~ _ _ ~ ~ C~ C~ V~ oo ~ O ¢~ <~~ c~ ~ ~- O ~ _ ~ ~ ~ ~1 O V~ ~ ~ ~ — O ~ ~ ~ V~ ~ O O O O O O O O — — ~ — C~ O C~ z V) _ ~ >< ~ ~ _ _ E ~ C ~ V ~ = ~ ~ ~ ~ E _ = _ oo Co oo (~~ ~ oo ~ o~o _ _ ~ — — o o o — o o o o _ _ X — oo C~~ oo oo 04 Ct~ ~ ~ C= ~ ~ ° C~ C~ ~ ~ ce ~ ~ ~ ~ ~ ~ ~ == = = D 0 = =

58 C) Cat :C ~ o o z z - ~ + ~ _ ~ _ .= - 1 - C' of _ _ oo _ Cal ~ ~ oo X ~ (I\ — X X of ~ _ . C) _ _ oo · ~ — — 2- C; — i=` — — ~ '-,' ,= D == ~ O V .,,, ~ on X — (,) ~ 53 X V) in, ~ To — ~— oo =. O ~ O O ~ ~ ~ ~ O ~ O ~ ~ Hi ." ~ .Q .O _ ~ ,, a = ~ ~ ~ D a ~ D o X t_ ~ of to X ~ ~ C~ ~ C°~°, Z ~ ~ ~ Mo ~o _ Ct ~: o ._ C~ V) ~ ~ C) C~ ~ s 3 ,~e _ ~ ,; o ~ o>` ~ ~ VF ~ =

s9 - x x c,' - =\ — ~ — ~ ce — it G) ~ C;) _ 3 _ ~ ~ o ~ ~ to ~ ~ ~ ~ o ~ ~ cat ~ ~ 6 a: oo ~ ~ ~ ~ oo — C—} ~ — _ Z ;5 Z ~ ~ - = o o o ~ C) Cal to Go of Go ~ To _ ~ ~ oo o ~ ~ o — o o o oo ~ oo To To oo (~\ _ oo ~ ~ ao Z ; - C¢ X Y ~ ~ Uw ~ ~ ' == _ =, Z ~ ~ ~ o ~ _ ~o ~ _ C) £ 3 o . . 3 ~ O . _ , =) c: ~ _ C,) ~— et e~ S~ ) ) e ~ c~ ~ ~ :- - v ~ o ~ 'v ~ ·- (v c:: ~ ~ 3 ~ ~ o =` oe o ~ ~ ~ - a ._ .o ~ C~ ~ ._ ._ X ~ O_ X _ ~ 00 X ~ .. ~ ~ O ^ V:} ~ . C) _ ._ ~ U: ~ C~ ._ C) O — CO O ._ o C) ~ ~ C) C) C~ ._ O ·_ ~ ~ C¢ _ V: ~ ~ ~ E~ Ct D .. _ _ ~ 3 ~ £ o ~ ~ C) ~V Se >, D _ · _ , _ c: ~: C~: o O ~ - - C) Z; C~, 3 ~ ~ e>e ~ _ _ C,) ~ _ c5 0 C) ~.= ~ ~ E.= — Cl: Ct ._ _ ~ ~ . - ~ ~3 ) c~ ~ - ~ - '= D o o C~ 3 U: 04 ~: cn s £ o 3 - x C) o C) - s C~ \~ C~ ~ ~ ~ . _ ~ _ C C) ~ C s E_ C, . ~ _ ~D ~ ~ _ ~ ° 3° ~ o ~ 3 c~ ~ O CO ~ -eS ~ ' - ~ =0 'C~ D '_ _) O g ~ ,,= .w o 3 ~ ~ ~ ~ o C) £ ~ £ 3 04 _ ~, — :;: ~ C%t X C) o C~ 3 ct ~— =0 ._ ~) - 3`, ~ ._ C) — ~I~ — _ ~ ~ ~ _ C,) ~ Ca ~o ~ ~ E_, C~ ~ o ._ ~ ,_ o — =, ~ C,: U: ~ _ ~ _ ~ ~ ~ =° 3 ''' ~ ~ o ~ ~o ,~ c~ ~ ~ ~ ~ ~C ~ ~ ~ £ ._ ~ ~ ~ ;,= ~CS o, tC CO ~ ;> o o ~ ~ 3 ~ c5 ~o ° '= .— o ~ ~ ~ o o ~ D _.— ~ oo X cr ° Ct _ ~ - ._ - o C`2 .o - Ct - C~ o o ._ - CC~) 3 C~ ~5 e~ o 3 u: cc C) — ~ E_ ~ ~ ~ C) ~ ~ _ .O ._ D ~ ~ ~ c5 ._ - C~ C) Ct o C) C~ C) o C~ C) - - ~C o o ; U~ ~: o 4> C~ C) ~¢ C) £ · o 3 o - - C~ c C) - c

60 ~ AIDS: THE SECOND DECADE with AIDS case data32 and indicates that, for some populations, women will be bearing a larger share of the AIDS burden in the future. For all age groups in the military surveillance data, the rates of HIV infection were highest for black women. These epidemiological data outline a picture of increasing risk of HIV infection for subsets of women, In particular, m~nonty women and women who are directly (through their own IV drug use) or indirectly (through their sexual partner's drug use) at risk because of drug-related exposure to the virus. Yet, the data on AIDS case distributions and HIV infection give only a profile of risk in this population. A more complete picture, one that provides a sense of depth and breadth, must include a sense of who these women are, how they live, and how they have come to be at risk for this disease. Unfortunately, as in many other areas of understanding related to the epidemic, information about these groups and the patterns of behavior that spread the virus information crucial to the design of efficacious interventions is limited, although the base of knowledge is growing. The sections below present what is now known about several subsets of the population of women at risk for HIV infection. Women Who Inject Drugs Of the women at risk for HIV infection, those who are exposed Erectly and indirectly (see the next section) to IV drug use have been shown to be particularly vulnerable. A small body of research provides some information about the women who inject drugs and the social env~ron- ments in which they live. For example, a subset of women injectors report a history of physical and sexual abuse (Weiner, Kaltenbach, and Finnegan, 1989; Worth et al., 1989~. hndeed, as Worth and her coworkers reported in art ethnographic study of 96 female intravenous drug users from the South Bronx, drug-related violence and illegal activities often characterize the relationships of these women with their sexual parmers. Sowder, Weissman, and Young analyzed preliminary data from ongoing NIDA demonstration projects33 targeting women at risk for HIV infection 32The ratio of male-to-female AIDS cases diagnosed among adolescents and adults was 14.7:1 in 1982 and S.6:1 in 1988. 33The National AIDS Demonstration Research Project, sponsored by the National Institute on Drug Abuse (NIDA), is providing interventions at 63 sites throughout the United States and Puerto Rico, and on the U.S.-Mexican border. These interventions are designed to change risk-associated behaviors and sustain healthier lifestyles, as well as to provide referral to drug treatment and other needed health and social services. The intervention efforts also seek to collect data to evaluate the effectiveness of interventions and to understand the sociodemographic charactenstics, lifestyles, and risk behaviors of targeted women (Sowder, Weissman, and Young, 1989).

STATE OF THE EPIDEMIC | 61 and found that many women who inject drugs are poorly educated and depend on illegal sources of income or government programs for their livelihood (Sowder, Weissman' and Young, 1989~.34 In addition, several small ethnographic studies of female injectors have found high rates of unemployment, unstable housing, and homelessness (Watters, 1988; Weiner, Kaltenbach, and Finnegan, 1989~. It is no suspense, therefore, that many women injectors report that they fee! powerless and lack control over their lives. Many women report that they are introduced to Mug injection by a male sexual partner or family member (Chambers, Hinesley, and Mold- estad, 1970; Chambers and Moffitt, 1970; Eldred and Washington, 1976; Gerstein, Judd, and Rovner, 1979; Worth and Rodriguez, 19871. Un- fortunately, our understanding of the process by which women become injectors is incomplete. Regardless of how the process starts, once women begin to inject, they may engage in the behavior that puts other injectors at risk- namely, the sharing of injection equipment with drug users who may be infected with the AIDS Virus.35 Sharing may occur in a variety of contexts: with friends or family members, within a couple, with "running buddies," in "shooting galleries," or using rented "works" from the drug dealer (Turner, Miller, and Moses, l989:Ch. 3~.36 Preliminary data from the NIDA-sponsored multisite study of female {V drug users indicate that the majority of women who report injecting drugs have shared injec- tion equipment and rented or borrowed equipment in the past (Sowder, 340f the women (N = 1,229) participating in a NIDA demonstration projects on which Sowder, Weiss- man, and Young ( 1989) reported, fewer than half graduated from high school. Among intravenous drug users in the projects (N = 736), 31 percent relied on illegal sources of income, and 28 percent received their major income from government programs. Among women who did not inject drugs but who were sexual partners of IV drug users (N = 493), 9 percent relied on illegal sources of income, and 41 percent relied on government programs (Sowder, Weissman, and Young, 1989). 35The degree of risk is also associated with the duration of drug use and the frequency of injection; the drug used (some drugs, such as cocaine, have a relatively short-lived effect and are therefore injected more frequently); the number of other people who have shared the injection equipment; seroprevalence rates in the local intravenous drug-using community; and the method of injection. For example, sharing injection equipment that has been used to "boot" drugs (i.e., to draw blood back into the syringe while the needle is still in the vein lo pulse the amount of drug being delivered and to extract the maximum amount of drug from the syringe) is riskier than using separate, sterile injection equipment but sharing a cooker. 36"Running buddies" are colleagues with whom drug users may pass recreational time and share drugs and injection equipment; running buddies may also be sexual partners. "Shooting galleries" are com- munal injection sites that are often found in large cities. "Works" are the IV drug user's injection paraphernalia, at least five elements of which carry the potential for contamination. For a more com- plete discussion, see Turner, Miller, and Moses (1989:189-195).

62 ~ AIDS: THE SECOND DECADE Weissman, and Young, 1989~.37 Newcomers to intravenous drug use are unlikely to have their own equipment; therefore, they often employ the works of a more experienced user for their initial injection.38 Women may also seek assistance with injections, depending on male partners, a female friend, or a "doe" (someone who specializes in injecting others and is often paid with part of the drug being injected) to inject them (Rosenbaum, 19811. Sharing may be encouraged by drug withdrawal. Often as addicted injectors find themselves experiencing withdrawal symptoms, the desire to inject becomes stronger. Reducing the time between the purchase of the drug and its injection may assume great urgency; as a result, addicts in withdrawal may be unusually likely to rent a dealer's works or use a nearby shooting gallery (Mandell, VIahov, and Cohn, 1989~. Although it is difficult to estimate precisely the number of female IV drug users arid the distribution of risk-associated behaviors, there appears to be sufficient evidence to state that the injection practices of most female IV drug users leave them vulnerable to HIV infection. HIV transmission through sexual behaviors is also a risk for women who inject drugs. Small surveys that have included female injectors have found that significant proportions report risk-associated behaviors, including multiple sex partners and unprotected intercourse (Snyder and Myers, 1989; Tross et al., 1989; Whittaker et al., 19891. In interviews with 736 female injectors who are participating in the NIDA demonstra- tion project, 38 percent reported two or more sex partners who injected drugs, and 83 percent stated that they never or rarely used condoms (Sow- der, Weissman, and Young, 19891. A separate survey of IV drug users recruited from the streets of Long Beach, California (110 of whom were women), found that approximately one-third of the sample (38 percent) reported that they had engaged in prostitution or had exchanged sex for drugs (Corby, Rhodes, and Wolitski, 19891. As noted earlier, the use of crack may increase the risk for women of acquiring HIV infection through sexual behaviors; indeed, crack use may be an especially important risk 370f the 736 women with a history of IV drug use who are participating in the National AIDS Demon- stration Research Project, 93 percent have shared needles, 71 percent have shared injection equipment with two or more partners, and 76 percent have used borrowed or rented needles (Sowder, Weissman, and Young, 1989). Other studies confirm the risk these behaviors pose to women. In a small survey of 220 female IV drug users recruited from methadone and detoxification programs in the San Francisco area, sharing needles with more than 10 partners over the past year was associated with HIV infection (Wolfe et al., 1989). 38Des Jarlais, Friedman, and Smug (1986) note that almost all injectors share needles at some point in their drug use career.

STATE OF THE EPIDEMIC l 63 factor among women who inject drugs.39 Several studies have shown an association between crack use in women (injectors and noninjectors) and higher levels of both unprotected sex and seropositivity.40 These findings, as well as the documented practice of women exchanging sex for crack, point to the reed for further study of this relatively new drug-related risk pattern. As is true for any woman of childbearing age who is at risk of becoming infected, women who inject drugs have the potential to transmit HIV vertically to their offspring. Stereotypical depictions of IV drug users do not usually include a family and children. Yet most of the female drug users who have been recruited for studies through treatment programs have children (Brown et al., 1989; Wolfe et al., 19891.4~ IV Mug users are reputed to be ineffective or inconsistent users of contraception (Ralph and Spigner, 19861. However, recent data from interviews with female IV drug users indicate that women who were uncertain about wanting children tended to employ safer sex practices (Tross et al., 19891. These data suggest the possibility that behaviors usually labeled as risk taking (i.e., unprotected vaginal intercourse) may not necessarily reflect relapse from condom use in some segments of the IV drug-using population but rather a desire for children. Vertical transmission of HIV would also appear to be an issue for the female partners of male IV Mug users. Yet the parameters of the problem in this subset of women are unknown because there are few available data on many aspects of these women's lives. The next section discusses 39Intravenous drug use does not preclude the use of noninjectable substances, such as cocaine or crack. In fact, the use of multiple drugs (so-called polydrug use) is the norm rather than the exception for some samples of female injectors (Weiner, Kaltenbach, and Finnegan, 1989). In Sowder, Weissman, and Yolmg's ( 1989) sample of 736 female IV drug users, almost half (46 percent) smoked or inhaled cocaine on less than a daily basis; an additional quarter (24 percent) used noninjectable cocaine daily. Sixteen percent reported daily use of marijuana, and 31 percent reported daily use of alcohol. Abramowitz and coworkers (1989) report a dramatic and statistically significant increase in crack use since 1986 among drug users seeking treatment in San Francisco. 400f 303 female IV drug users recruited from drug treatment programs in San Francisco, 84 (27.7 percent) reported that they had used crack in the past 30 days, and crack use displayed a significant association with seropositivity (Wolfe et al., 1989). A separate study of New York City women seeking care for pelvic inflammatory disease (of whom approximately 8 percent injected chugs) found that more than half (51 percent) reported crack use (Hoegsherg et al., 1989). Moreover, crack users had twice as many sexual partners as nonusers and were more likely lo be infected with HIV. 41In a study of the medical records of 593 women participating in a methadone maintenance program in Brooklyn and Manhattan in 1987, Brown and coworkers (1989) found that these women reported a total of 2,298 pregnancies, which resulted in 2.51 live births and 1.37 abortions per woman. In each age group that was studied, black and Hispanic women on average reported more pregnancies and live births and fewer abortions than white women.

64 ~ AIDS: THE SECOND DECADE the limited information available on this group and delineates, as far as is possible, the risk profile of these women. Female Sexual Partners of IV Drug Users The population of female sex partners of IV drug users has been a difficult population to study and reach with prevention programs because these women are not readily identifiable, they do not belong to any unifying, so- cial group, and they may have very unstable living conditions. Although there has been only limited research on this group, service providers, law enforcement officials, and the drug users themselves indicate that the population is quite diverse with respect to race, ethnicity, education, and income (Arguelles et al., 19891. Some researchers have suggested that many female sexual partners of men who inject drugs do not themselves use drugs intravenously (Des JarIais et al., 1984; Murphy, 1987~42 and therefore are not at direct risk of acquiring HIV from this mode of trans- mission. They may, however, use noninjectable drugs, such as alcohol or marijuana, which may, like crack, increase their risk of acquiring HIV through sexual transmission. Preliminary ethnographic data on this group of women raise another disturbing possibility in relation to their increased risk of acquiring HIV through sexual transmission. Some women who do not use drugs are ignorant of their partner's injection practices;43 others may not be able or willing to acknowledge this behavior (Weissman, 19881. Indeed, women may fear the confrontation that might ensue if such practices were openly recognized. Alternatively, male drug users may take pains to prevent their female sex partners from knowing about their Hug use. Without knowledge of their partners' risky practices, these women may perceive them risk as unrealistically low, and there will be little likelihood Hat they will consider protective action. A survey of 62 female sexual partners of IV drug users found a positive association between the women's perception of risk for HIV infection and an awareness of the needIe- shanng practices of their partners (Werrnuth and Ham, 1989~. In this small study, the perception of risk did not differentiate women who used condoms from those who did not (preventive action in the group was predicated instead on knowing the partner's seropositive status). Nevertheless, theories of health behavior indicate that such a perception 42Almost half (42 percent) of 325 IV drug users interviewed in Long Beach, California reported having sexual partners who did not use IV drugs (Corby, Rhodes, and Wolitski, 1989). 43 In a survey of 325 IV drug users recruited from the sweets of Long Beach, California, 18 percent indicated that their partner did not know of their drug use (Corby, Rhodes, and Wolitski, 1989).

STATE OF THE EPIDEMIC ~ 65 is important in beginning to motivate the behavioral change process.44 The findings of this study emphasize the crucial need for data on these populations to design effective interventions to fight the spread of the . . . epluemlc. Female Sexual Partners of Male Hemophiliacs and Transfusion Recipients Women who engage in unprotected intercourse with men who have become infected through exposure to contaminated blood are themselves at risk of acquiring HIV infection, although the exact extent of the risk is not known. As of December 1989, 48 current or foyer heterosexual partners of hemophilic men had been diagnosed with AIDS; an additional 59 reported cases among women were ascribed to sexual contact with a transfusion recipient (CDC, l990a).45 The total number of women who have acquired infection in this manner is not known with any certainty; a number of small surveys have generated widely ranging estimates of HIV prevalence in this population.46 Methodological limitations of these studies make it difficult to interpret the findings: most have relied on small samples of volunteers, and much of the data has been gathered from limited geographic areas. Women participants in these studies have varied with respect to symptoms, infection, contraceptive methods, frequency and duration of sexual contact with an infected partner, and other risk-associated behaviors. The factors that contribute to transmission of the AIDS virus in discordant heterosexual couples (i.e., couples in which only one part- ner iS known to be infected) are uncertain.47 fit iS clear, however, that 1 44 For further information on motivating health behavior changes, see Chapter 4 in Turner, Miller, and Moses (1989). 45It is not possible to calculate prevalence rates for this population because the total number of women who have had sexual contact with hemophilic men is not known. Jackson and colleagues (1988) have estimated that 15,000 hemophiliacs in the United States received large-pool, non-heat-treated factor VIII or factor IX concentrates. This number includes individuals of all ages, but virtually all are males. 46CDC estimates that between 10 and 60 percent of the spouses and sexual partners of hemophiliacs may have become infected through unprotected sexual intercourse (CDC, 1987). Some researchers have speculated that women who are symptomatic may be more inclined to volunteer for research programs, thus leading to high seroprevalence rates in some studies (Lawrence et al., 1989, 1990). A national telephone survey of physicians and hemophilia treatment centers found that, among the 34 percent of female heterosexual partners of hemophiliacs who were tested for HIV infection, 10 percent were positive. A review of 11 other studies of heterosexual partners of hemophiliacs conducted in the United States found seroprevalence rates between 0 and 21 percent (Lawrence et al., 1990). Between January 1988 and September 1989, 667 partners of hemophiliacs sought testing and counseling through programs supported by CDC; 34 (5.1 percent) were seropositive (CDC, 1990b). 47Transmission may require active viral replication that is, the infection must have progressed to the

66 ~ AIDS: THE SECOND DECADE nsk-associated behaviors are reported by these couples to varying de- grees. Few female partners of infected hemophilic men appear to be celibate,48 and, despite the risk of sexually transmitted HIV infection, many couples do not use condoms consistently.49 It is unclear why such risk taking has not led to a greater level of infection In this population. When the only reported risk factor is unprotected vaginal intercourse, it appears that seroconversion rates among the female sexual parmers of infected hemophiliacs are generally Tow (Brettler et al., 1988; Kim et al., 1988; Ragni et al., 1988; Jackson et al., 19891. Yet despite the incomplete understanding of the association between risk and infection in this population, the committee would emphasize that repeated exposure through unprotected intercourse can and does lead to infection among female sex partners of infected men (Lawrence et al., 1990~. With con- sistent protection, however, transmission among discordant couples can be avoided.50 point at which HIV can be isolated from serum (Laurian, Peynet, and Verroust, 1989 - or immuno- logic deficiencies in the infected partner (Smiley et al., 1988). The frequency of sexual intercourse and the appropriate and consistent use of condoms may also affect the likelihood of transmission to the uninfected partner, but there are no data supporting their contributions, if any, at least at present. In fact, the available data show no difference in either frequency of coitus or condom use between se- roconverters and nonconverters (Ragni et al., 1988), nor does the rate of other genital infections differ between these two groups (Smiley et al, 1988). The lack of hard evidence on HIV prevention through condom use by such couples may also reflect their prior inconsistent use of condoms (especially during early infection), the uncertain quality of self-reported data, the small samples used in many studies, or perhaps the failure of condoms to prevent transmission (Lawrence et al., 1990). Furthennore, the evidence has been weakened by posing questions to couples that are too general to be valid or reliable or that have been worded inconsistently. The contribution of other factors is also uncertain. Anal in- tercourse is not required for HIV transmission in discordant heterosexual couples (Kim et al., 1988; Ragni et al., 1988; Smiley et al., 1988), nor is the co-occurrence of other sexually transmitted diseases or sex during menstrual periods (Goedert et al., 1987). However, regardless of the factors that con- tribute to transmission, Mason, Olson, and Parish (1988) report that the risk of infection among these couples does appear to increase with time. 48 In two separate studies of approximately 50 female sexual partners of hemophilic men, all of the women except one individual in each study reported engaging in vaginal intercourse (anal intercourse is rarely reported in studies of hemophiliac couples) (Lawrence et al., 1990). 49In a study by Wilson and Wasserman (1989), for example, fewer than one-fourth (24.3 percent) of sexually active hemophilic men reported consistent condom use. A survey of 56 female sex partners of infected hemophilic men confirms this statistic. An alanning 72 percent of women whose partners were infected and symptomatic and 54 percent of women whose infected partners were asymptomatic reported that they never or rarely used condoms (Lawrence et al., 1989). Parish and colleagues (1989) also report that of 351 hemophilia men in their study, only a third reported consistent condom use. A more detailed discussion of condom use (or lack of use) appears in Chapter 2. 50A small cohort of women remained uninfected after five years of protected sexual contact with an infected hemophiliac partner (Goedert et al., 1987). Other studies have also demonstrated that consistent condom use can prevent infection in this population (Forsberg et al., 1989; Laurian, Peynet, and Verroust, 1989). There is no evidence that transmission occurs from intimate but nonsexual contact (Brettler et al., 1988).

STATE OF THE EPIDEMIC ~ 67 The epidemiological data presented above indicate that the profile of the epidemic among women is changing. It is apparent that actual infection as well as the threat of infection is spreading to encompass wider segments of the female population. AIDS prevention efforts must now address this more diverse group in addition to the unique role of women in vertical transmission of the virus. Some of the issues related to interventions for women are discussed in the next chapter. The epidemiological data presented earlier also indicate the dynamic, fluid quality of the epidemic among other populations and emphasize the fact that the future path of AIDS in this country is still uncertain. Despite the knowledge gained over the past decade, there are persisting gaps in understanding of the dynamics of HIV infection among all at- nsk populations. The next section reviews some of the data needs for improved surveillance efforts. TRACKING THE EPIDEMIC: DATA NEEDS Substantial progress has been made during the first decade of the epidemic in monitoring AIDS prevalence throughout venous subpopulations. A case reporting system is in place, and AIDS case data are available to researchers from CDC. As noted earlier, however, our understanding of the spread of HIV infection and the behaviors that transmit it is far from complete. The committee finds that several technical issues impinge on that understanding, including sampling techniques, completeness of reporting, data access, and standardization of items that appear in different surveys. In general, attempts to ascertain the prevalence of HIV infection in this country have relied on small cross-sectional studies, which can provide only snapshot views of the problem in selected groups in specific locations at one point in time. Seroprevalence estimates for IV drug users, for example, often come from samples of convenience recruited through treatment programs or criminal justice agencies. Seroprevalence data for the gay male population come from a variety of sources, including smaller studies of patient populations recruited through STD clinics and HIV testing and counseling facilities. Improved data collection efforts are needed to understand the spread of HIV infection in these and other at-risk populations, including women. Such efforts should include information on the behaviors that transmit the virus as well as the prevalence of . ~ . ntectlon. There are, however, some important exceptions to these smaller surveys. Large-scale serosurveys of military and Job Corps applicants,

68 ~ AIDS: THE SECOND DECADE which are described in the chapter on adolescents, have provided in- formation on the prevalence of infection in these special populations. In addition, large prospective studies of gay men, such as the National Institute of Allergy and Infectious Disease's Multicenter AIDS Cohort Studies (MACS), have gathered valuable data on the natural history of the disease, the prevalence of infection, and patterns of r~sk-associated behavior in that population. (Both the military and the MACS surveys can provide additional information on the incidence of infection in their samples.) As discussed earlier, CDC's population-based serologic survey of newborn infants has generated some new and very valuable data on infection rates among childbearing women. This effort is only one of several systematic inqu~nes~DC's "family" of surveys intended to measure the prevalence of infection in potentially high-risk locations and subpopulations (e.g., clients at STD clinics, patients at certain "sentinel hospitals," college students). Unlike the neonatal survey, however, these other elements of the family of surveys were originally designed to gather data from non-population-based samples. As a result, in its earlier re- port, the committee recommended that the design of these surveys be augmented to use probability sampling and thus permit researchers and epidemiologists to make generalizations beyond the particular group be- ing studied. The committee is gratified to learn that CDC has already taken steps to convert the college survey to a population-based sample and to explore similar strategies for other elements of the family of surveys. Understanding the course of the epidemic, however, and measuring it as it moves through different populations requires attention to issues broader than sampling. Monitoring the spread of HIV and AIDS requires careful surveillance for changes both in the behaviors that transmit infec- tion and in the manifestations of disease. As AIDS cases are diagnosed among varied subpopulations, epidemiologists have come to recognize a wide spectrum of clinical signs and symptoms that may be associated win the disease. This diversity has resulted in changes in the case definition used for surveillance purposes, which in turn has led to changes in the 5 ~ MACS cohorts of gay men were established in Chicago, Pittsburgh, Baltimore, and Los Angeles in the early years of the epidemic, as was the San Francisco Men's Health Study. Although these cohorts have provided very valuable data, they may not be representative of the larger population of men who have sex with men. For example, as Frutchey and Walsh ( 1989) suggest, MACS participants are volunteers who have made a long-term commitment to a research project; consequently, they may differ from the larger population in teens of motivation. In addition, the cohorts have aged and thus do not adequately represent younger men who may have different patterns of sexual behavior. The men have also been extensively counseled and studied (and as a result may have decreased their risk taking more than men who have not had similar experiences). Moreover, these cohorts provide little information about gay men living in nonurban areas or about men of different ages and cultural backgrounds.

STATE OF THE EPIDEMIC ~ 69 prevalence of AIDS across risk categories. For example, the increase in AIDS cases attributed to IV drug use (noted earlier) is probably a result of an increasing number of IV drug users developing AIDS and the broader case definition established by CDC In September 1987. The original surveillance definition for AIDS was developed primarily through stud- ies of the natural history of the disease among homosexual and bisexual men and reflected the disease as it appeared in those individuals. It is now understood that there may be very different manifestations of the underlying immune defect across risk categones, including a "wasting syndrome" and AIDS-related dementia, and these conditions are now in- cluded as enters in the surveillance definition of AIDS.52 Completeness of reporting is thus a complex issue in these circumstances and is likely to require continued attention as the epidemic evolves over the course of the next decade. Monitonng changes in risk-associated behaviors that can lead to subsequent acquisition and spread of infection is also critically important. Dunng the first decade of the epidemic, investigators who considered the risks associated with drug use focused primarily on injection practices. As the epidemic has matured, however, greater appreciation of the role of other drugs, such as crack, has emerged. At present it is not possible to predict how rapidly HIV will spread among crack users. Factors that affect the rate and extent of spread and that require monitoring include the following: · the number of people who use crack, the frequency with which it is used, and the contexts in which it is used; · the efficiency of heterosexual transmission of HIV from males to females and from females to males; · the number of already-infected persons who may bring the virus to locations where sex is exchanged for crack; · the rates of partner change among different groups of crack users; · the extent to which syphilis and other STDs facilitate HIV 52Several studies from New York City indicate that many of the IV drug users who are dying from HIV- related illnesses have never developed Kaposi's sarcoma or an opportunistic infection, the previous criteria of the case surveillance definition of AIDS (Selwyn et al., 1988; Stoneburner et al., 1988). Recent analyses of national data find that a greater proportion of cases meeting only the new criteria have been reported among heterosexual IV drug users than among homosexual men and among blacks and Hispanics than among whites (Selik et al., 1990). Some researchers estimate that, if all previously unrecognized fatal illnesses associated with HIV infection were reclassified as AIDS, the number of ADS-related deaths among IV drug users might be as much as twice the count derived from official statistics (Stonebumer et al., 1988).

70 ~ AIDS: THE SECOND DECADE transmission and the extent to which the incidence of these diseases can be controlled; and · the extent to which crack users can adopt safer sexual prac- tices. It is clear that crack use and its associated unsafe sexual activity rep- resent a potentially important new wave of HIV transmission in the United States. What is unclear is how large a wave this might be. The commit- tee recommends that the Public Health Service support additional research on crack use, including its epidemiology, its relationship to sexual behavior, strategies to reduce its occurrence (both initiation of use and continuance among low- and high-frequency users), and methods for facilitating change in the sexual behavior of persons who continue to use crack. Understanding the use of crack and its role in the spread of HIV infection will require basic behavioral research, re- search on the social factors associated with initiating and sustaining crack use, biological research to determine effective treatment strategies, and demonstration projects to assess the effectiveness of intervention efforts. Collecting good data regarding AIDS and HIV infection is intnn- sically difficult because of the sensitivity of the subject matter and the difficulties involved in locating and studying some of the subgroups at highest risk. (The last chapter of the report is a more detailed discussion of how the quality of sensitive data can be improved.) Many researchers rely on secondary analyses of data collected by others, including agen- cies of the state and federal governments that are responsible for public health. Timely access to those data is sometimes constrained by policies pertaining to confidentiality as well as by other data release regulations. Moreover, it is often difficult to compare data across data sources. For example, in the CDC serologic survey of newborn infants, the data col- lected on the sociodemographic characteristics of the mother vary from state to state, precluding comparisons. In addition, in those states that only collect limited sociodemographic information, efforts by state and local public health agencies to design and implement programs and ser- vices may be constrained by a lack of data on the subgroups who need attention. Mathematical models hold some promise for improving our understanding of the course of the epidemic and thus for planning the delivery of programs and services. However, shortcomings in existing data on the prevalence and distribution of risk-associated behaviors and

STATE OF THE EPIDEMIC ~ 71 HIV infection, on the natural history of the disease,53 and on other factors relevant to the spread of infection currently limit the use of such models. Improving data quality for ~] populations will require a wide range of additional efforts. In some instances (e.g., studies of adult gay men), improvement will demand that new and younger cohorts be recruited as part of ongoing efforts to understand the disease's natural history and changing patterns of risk. For drug users, improvements in data qual- ity will require innovative sampling strategies, modification of existing surveillance approaches, and attention to the appearance of new drugs whose use may pose additional risk to this population. Progress is needed in several other areas as well, including improving the quality of data from sites currently collecting information on the prevalence of infection and risk behaviors (e.g., STY clinics), resolving interjurisdictional prob- lems of confidentiality and data release, and establishing mechanisms for determining what information to obtain in each survey (e.g., how much detail to elicit concerning mothers in the neonatal surveys or patients in surveys conducted in STD clinics). Improving data quality is vital to improving our capacity to monitor the movement of the epidemic. It is also essential for evaluating the impact of intervention efforts to halt the spread of HIV infection. As the agency within the Public Health Service with primary responsibility for collecting surveillance data, CDC instituted data collection programs in the first decade of the epidemic, targeting diverse at-nsk populations. The dynamic nature of the epidemic, however, mandates continuing attention to these programs to ensure that relevant subgroups are being monitored. Thus, to facilitate the Centers for Disease Control's (CDC) ongoing efforts to improve its AIDS-related data collection systems, the com- mittee recommends that the agency initiate a systematic review of current programs. This effort should draw on the expertise of both CDC staff and outside experts. In summary, sound epiclemiologica] data are important in clevelop- ing accurate an] (3etallecl pictures of at-Hsk populations. The finer the grain of these pictures, the more useful they become. Depictions that Carl distinguish subgroups at Varying levels of risk and that Carl detect changing risk patterns are vital in cl~ectiDg intervention resources to those in greatest need; they ale also essential to the development of relevant intervention programs and to the evaluation of these programs.54 Having 53As noted earlier, prospective studies of large cohorts of gay men (e.g., the MACS) have provided valuable information on the natural history of the disease- for this population. Unfortunately, less is known about the natural history of the disease in other populations, such as adolescents and women. 54For more information on program evaluation, see the recent report of the committee's Panel on the

72 ~ AIDS: THE SECOND DECADE identified and charactenzed the problem, insofar as the data permit, the committee tutus to the question of what should be done. The next chapter reviews intervention strategies for populations at risk arid discusses the barriers that may impede their implementation. REFERENCES Abramowitz, A., Guydish, J., Woods, W., and Clark, W. (1989) Increasing crack use among drug users in an AIDS epicenter: San Francisco. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Altman, R., Grant, C. M., Brandon, D., Shahied, S., Rappaport, E., et al. (1989) Statewide HIV-1 serologic survey of newborns with resultant changes in screening and delivery system policy. Presented at the Fifth International Conference on AIDS, Montreal, June ~9. Araneta, M. R. G., Thomas, P. A., Ramirez, L. L., Weisfuse, I., and Schultz, S. (1989) No change in HIV-1 seroprevalence among parturients and women having induced abortions in New York City, 1987-1988. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Arguelles, L., Rivero, A. M., Reback, C. J., and Corby, N. H. (1989) Female sex partners of IV drug users: A study of socio-psychological characteristics and needs. Presented at the Fifth International AIDS Conference, Montreal, June =9. Armson, B. A., Mennuti, M. T., and Talbot, G. H. (1988) Seroprevalence of human immunodeficiency virus (HIV) in an obstetric population. Presented at the Fourth International Conference on AIDS, Stockholm, June 12-16. Atkinson, W., Troxler, S., Dal Corso, M., and McFarland, L. (1989) Intravenous drug use as a risk factor for AIDS and HIV infection in Louisiana, 1981-1988. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Barbacci, M., Chaisson, R., Anderson, J., and Horn, J. (1989) Knowledge of HIV serostatus and pregnancy decisions. Presented at the Fifth International AIDS Conference, Montreal, June ~9. Battjes, R. J., Pickens, R. W., and Amsel, Z. (1989) Introduction of HIV infection among intravenous drugabusers in low prevalence areas. Journal of Acquired Immune Deficiency Syndromes 2:533-539. Berkelman, R., Karon, J., Thomas, P., Ker7ldt, P., Rutherford, G., and Stehr-Green, J. (1989) Are AIDS cases among homosexual males leveling? Presented at the Fifth International Conferenceon AIDS, Montreal, Juneau. Brettler, D. B., Forsberg, A. D., Levine, P. H., Andrews, C. A., Baker, S., and Sullivan, J. L. (1988) Human immllnodeficiency virus isolation studies and antibody testing. Archives of Internal Medicine 148:1299-1301. Brondum, J., Debuono, B., Dondero, L., Hodge, J., and Johnson, A. (1987) The alternative test site (ATS) in Rhode Island. Presented at the Third International Conference on AIDS, Washington, D.C., June 1-5. Evaluation of AIDS Interventions (Coyle, Boruch, and Turner, 1990). In this report, the panel lays out specific strategies for assessing the effectiveness of three of CDC's major AIDS prevention programs: the national media campaign, testing and counseling, and the health education projects sponsored by its group of funded community-based organizations.

STATE OF THE EPIDEMIC ~ 73 Brown, L. S., Mitchell, J. L., DeVore, S. L., and Primm, B. J. (1989) Female intravenous drug users and pennatal HIV transmission. New England Journal of Medicine 320: 1493-1494. Castro, K. G., Lieb, S., Calisher, C., Witte, J., and Jaffe, H. W. (1987) AIDS and HIV infection in Belle Glade, Florida. Presented at the Third International Conference on AIDS, Washington, D.C., June 1-5. Castro, K. G., Lifson, A. R., White, C. R., Bush, T. J., Chamberland, M. E., et al. (1988) Investigations of AIDS patients with no previously identified risk factors. Journal of the American Medical Association 259:1338-1342. Centers for Disease Control (CDC). (1987) Human immunodeficiency virus infection in the United States. Morbidity and Mortality Weekly Report 36:801-804. Centers for Disease Control (CDC). (1989) HIV/AIDS Projections Workshop October 31-November 1, 1989: Draft Working Group Reports and Summary (December 3, 1989~. Atlanta, Gal: Centers for Disease Control. Centers for Disease Control (CDC). (1990a) HIVIAIDS Surveillance: U.S. AIDS Cases Reported Through December 1989. Atlanta, Gal: Centers for Disease Control. Centers for Disease Control (CDC). (1990b) Publicly funded HIV counseling and testing-United States, 1985-1989. Morbidity and Mortality Weekly Report 39:137- 140. Charnberland, M. E., Allen, J. R., Monroe, J. M., Garcia, N., Morgan, C., et al. (1985) Acquired immunodeficiency syndrome in New York City: Evaluation of an active surveillance system. Journal of the American Medical Association 254:383-387. Chambers, C. D., and Moffitt, A. (1970) Negro opiate addiction. In J. C. Ball and C. D. Chambers, eds., The Epidemiology of Opiate Addiction in the United States. Springfield, Ill.: Thomas. Chambers, C. D., Hinesley, R. K., and Moldestad, M. (1970) Narcotics addiction in females: A race comparison. International Journal of the Addictions 5:257-278. Chandrasekar, P. H., Matthews, M., and Chandrasekar, M. C. (1988) Risk factors for HIV infection among parenteral drug abusers (PDA) in a low-prevalence area. Presented at the Fourth International Conference on AIDS, Stockholm, June 12-16. Chiasson, M. A., Stoneburner, R. L., Telzak, E., Hildebrandt, D, Schultz, S., et al. (1989) Risk factors for HIV-1 infection in STD clinic patients: Evidence for crack-related heterosexual transmission. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Chiasson, M. A., Stoneburner, R. L., Lifson, A. R., Hildebrandt, D. S., Ewing, W. E., et al. (1990) Risk factors for human immunodeficiency virus type 1 (HIV-1) infection in patients at a sexually transmitted disease clinic in New York City. American Journal of Epidemiology 131:208-220. Cohn, D., Douglas, J., Koleis, J., Feeney, F., and Judson, F. (1989) Companson of prevalence of HIV infection in IV drug users (IVDU) from four different test~ng and treatment programs. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Comerford, M., Chitwood, D. D., McCoy, C. B., and Trapido, E. J. (1989) Association between fo~mer place of residence and serostatus of IVDUs in south Florida. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Connor, E, Denny, T., Goode, L., Niven, P., Oxtoby, M., et al. (1989) Seroprevalence of HIV-1 and HTLV-1 among pregnant women in Newark, N.J. Presented at the Fifth International Conference on AIDS, Montreal, June 09.

74 ~ AIDS: THE SECOND DECADE Conway, G. A., Colley-Niemeyer, B., Pursley, C., Cruz, C., Burt, S., and Heath, C. W. (1989) Underreporting of AIDS cases in South Carolina, 1986 and 1987. Journal of the American Medical Association 262:2859-2863. Corby, N. H., Rhodes, F., and Wolitski, R. J. (1989) HIV serostatus and risk behaviors of street IVOUs. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Cowan, D. N., Brundage, J., Miller, R., Goldenbaum, M., Pomerantz, R., and Wann, F. (1989) Prevalence of HIV infection among U.S. army reserve component personnel. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Coyle, S. C., Boruch, R. B., and Turner, C. F., eds. (1990) Evaluating AIDS Prevention Programs, Expanded Edition. Washington, D.C.: National Academy Press. Cozen, W., Mascola, L., Giles, M., Bauch, S., Finn, M., and Heneman, C. (1989) Routine HIV antibody screening in Los Angeles county prenatal clinics: A demonstration project. Presented at the Fifth International Conference on AIDS, Montreal, June 4-9. Curran, J. W., Jaffe, H. W., Hardy, A. M., Morgan, W. M., Selik, R. M., and Dondero, T. J. (1988) Epidemiology of HIV infection and AIDS in the United States. Science 239:610 616. Danila, R. N., Shultz, J. M., Osterholm, M. T., MacDonald, K. L., Henry, K., and Simpson, M. (1987) Minnesota counseling and testing sites: Analysis of trends over time. Presented at the Third International Conference on AIDS, Washington, D.C., June 1-5. Del Tempelis, C. D., Shell, G., Hoffman, M., Benjamin, R. A., Chandler, A., and Francis, D. (1987) lIuman immunodeficiency virus infection in women in the San Francisco Bay area. Journal of the American Medical Association 258:474 475. Des Jarlais, D. C., Friedman, S. R., and Strug, D. (1986) AIDS and needle sharing within IV-drug use subculture. In P. A. Feldman and T. M. Johnson, eds., The Social Dimensions of AIDS: Method and Theory. New York: Praeger Press. Des Jarlais, D. C., Chamberland, M. E., Yancovitz, S. R., Weinberg, P., and Fned- man, S. R. (1984) Heterosexual partners: A large risk group for AIDS. Lancet 2(8415):134~1347. Des Jarlais, D. C., Friedman, S. R., Novick, D. M., Sotheran, J. L., Thomas, P., et al. (1989) HIV-1 infection among intravenous drug users in Manhattan, New York City, from 1977 through 1987. Journal of the American Medical Association 261 :1008-1012. Eldred, C. A., and Washington, M. N. (1976) Interpersonal relationships in heroin use by men and women and their role in treatment outcome. International Journal of the Addictions 17:117-130. Ernst, J. A., Bauer, S., Amaral, L., St. Louis, M., and Falco, I. (1989) HIV seropreva- lence at the Bronx Lebanon Hospital Center A CDC sentinel hospital. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Fehrs, L., Hill, D., Kerndt, P., Rose, T., and Henneman, C. (1989) HIV screening program at a Los Angeles prenatal/family planning center. Presented at the Fifth International Conference on AIDS, Montreal, June =9.

STATE OF THE EPIDEMIC ~ 75 Forsberg, A. D., Sullivan, J. L., Willitts, D. L., Kraus, E., and Brettler, D. B. (1989) Results of HIV antibody testing in sexual partners of seropositive hemophiliacs over a 5-year penod. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Friedman, S. R., Dozier, C., Sterk, C., Williams, T., Sotheran, J. L., et al. (1988) Crack use puts women at risk for heterosexual transmission of HIV from intravenous drug users. Presented at the Fourth International Conference on AIDS, Stockholm, June 12-16. Frutchey, C., and Walsh, K. (1989) Marginalization of gay men in AIDS funding and programs. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Fullilove, R. E., Fullilove, M. T., Bowser, B. P., and Gross, S. A. (1989) Crack use and risk for AIDS among black adolescents. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Gail, M. H., Rosenberg, P. S., and Goedert, J. L. (1990) Therapy may explain recent deficits in AIDS incidence. Journal of Acquired Immune Deficiency Syndromes 4:296-306. Gardner, L. I., Brundage, J. F., Burke, D. S., McNeil, J. G., Visintine, R., and Miller, R. N. (1989) Evidence for spread of the human immunodeficiency virus epidemic into low-prevalence areas of the United States. Journal of Acquired Immune Deficiency Syndromes 2:521-532. Gerstein, D. R., Judd, L. L., and Rovner, S. A. (1979) Career dynamics of female heroin addicts. American Journal of Drug and Alcohol Abuse 6:1-23. Goedert, J. J., Eyster, M. E., Biggar, R. J., and Blattner, W. A. (1987) Heterosexual transmission of human immunodeficiency virus: Association with severe depletion of T-helper lymphocytes in men with hemophilia. AIDS Research and Human Retroviruses 3:355-361. Guinan, M. E., and Hardy, A. (1987) Epidemiology of AIDS in women in the United States: 1981 through 1986. Journal of the American Medical Association 257:2039-2042. Hardy, A. M., Starcher, E. T., Morgan, W. M., Druker, J., Kristal, A., et al. (1987) Review of death certificates to assess completeness of AIDS case reporting. Public Health Reports 102:38~391. Harrison, W. O., and Moore, T. A. (1988) Prenatal HIV screening in a low-risk population. Presented at the Fourth International Conference on AIDS, Stockholm, June 13-16. Hill, D., Kerndt, P., Frenkel, L. M., Settlage, R., Lee, M., et al. (1989) HIV seroprevalence among parturients in Los Angeles County, 1988. Presented at the Fifth Intemational Conference on AIDS, Montreal, June =9. Hoegsberg, B., Dotson, T., Abulafia, O., Tross, S., Des Jarlais, D., et al. (1989) Social, sexual and drug use profile of HIV-positive and HIV-negative women with PID. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Horman, J., and Hamidi, C. (1989) The epidemiology of AIDS in Maryland, 1981-1987. Presented at the Fifth International Conference on AIDS, Montreal, June ~9. Horton, J., Alexander, L., and Brundage, J. (1989) HIV prevalence among military women: An examination of military applicant, active duty, and reserve testing data. Presented at the Fifth International Conference on AIDS, Montreal, June =9.

76 ~ AIDS: THE SECOND DECADE Hull, H. F., Bettinger, C. J., Gallaher, M. M., Keller, N. M., Wilson, J., and Mertz, G. J. (1988) Comparison of HIV-antibody prevalence in patients consenting to and declining HIV-antibody testing in an STD clinic. Journal of the American Medical Association 260:935-938. Jackson, J. B., Sannerud, K. J., Hopsicker, J. S., Kwok, S. Y., Edson, J. R., and Balfour, H. H. (1988) Hemophiliacs with HIV antibody are actively infected. Journal of the American Medical Association 260:223~2239. Jackson, J. B., Kwok, S. Y., Hopsicker, J. S., Sannerud, K. J., Sninsky, J. J., et al. (1989) Absence of HIV-1 infection in antibody-negative sexual partners of HIV-1 infected hemophiliacs. Transfusion 29:265-267. Johnson, J. P., Alger, L., Nair, P., Watkins, S., 3ett, K., and Alexander, S. (1987) HIV screening in the high-risk obstetric population and infant serologic analysis. Presented at the Third International Conference on AIDS, Washington, D.C., June 1-5. Kaunitz, A. M., Brewer, J. L., Paryani, S. G., de Sausure, L., Sanchez-Ramos, L., et al. (1987) Prenatal care and HIV screening. Journal of the American Medical Association 258:2693. Kim, H. C., Raska, K. III, Clemow, L., Eisele, J., Marts, L., et al. (1988) Human immunodeficiency virus infection in sexually active wives of infected hemophiliac men. American Journal of Medicine 85:472~76. Krasinski, K., Burkowsky, W., Beber~oth, D., and Moore, T. (1988) Failure of voluntary testing for HIV to identify infected parturient women in a high-risk population. New England Journal of Medicine 318:185. Lafferty, W. E., Hopkins, S. G., Honey, J., Harwell, J. D., Shoemaker, P. C., and Kobayashi, J. M. (1988) Hospital charges for people win AIDS in Washington State: Utilization of a statewide hospital discharge data base. American Journal of Public Health 78:949-957. Landesman, S. H., Minkoff, H., Holman, S., McCalla, S., and Sijin, O. (1987) Serosurvey of human irnmunodeficiency virus infection in parturients: Implications for human immunodeficiency virus testing programs of pregnant women. Journal of the American Medical Association 258:2701-2703. Landis, S., Schoenbach, V., Weber, D., Mittal, M., Koch, G., and Levine, P. (1989) HIV-1 seroprevalence in sexually transmitted disease (STD) clinic patients in central North Carolina. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Laurian, Y., Peynet, J., and Verroust, F. (1989) HIV infection in sexual parmers of HIV seropositive patients with hemophilia. New England Journal of Medicine 320:183. Lawrence, D. N., Jason, J. M., Holman, R. C., Heine, P., Evatt, B. L., and the Hemophilia Study Group. (1989) Sex practice correlates of human immunodeficiency virus transmission and acquired immunodeficiency syndrome incidence in heterosexual parmers and offspring of U.S. hemophilic men. American Journal of Hematology 30:68-76. Lawrence, D. N., Jason, J. M., Holman, R. C., and Murphy, J. J. (1990) Human iITlmunodeficiency virus transmission from hemophilic men to their heterosexual parmers. In N. J. Alexander, H. L. Gabelnick, and J. M. Spieler, eds., The Heterosexual Transmission of AIDS. New York: Alan R. Liss.

STATE OF THE EPIDEMIC 77 Lindsay, M. K., Peterson, H. B., Mundy, D. C., Slade, B. A., Feng, T., et al. (1989) Seroprevalence of human immunodeficiency virus infection in a prenatal population at high risk for HIV infection. Southern Medical Journal 82:825-828. Mandell, W., Vlahov, D., and Cohn, S. (1989) IVDU characteristics associated with needle sharing. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Mason, P. J., Olson, R. A., and Parish, K. L. (1988) AIDS, hemophilia, and prevention efforts within a comprehensive care program. American Psychologist 43:971-976. McCoy, C. B., Chitwood, D. D., and Page, J. B. (1989) Mobility, risk cities, risk behavior, and HIV status of IV drug users. Presented at the Fifth International Conference on AIDS, Montreal, June =9. McFarland, L., Dean, H., Trahan, B., and Muirhead, L. (1989) HIV infection in pregnant women at a public hospital in New Orleans, Louisiana. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Modesitt, S.K., Hulman, S., and Fleming, D. (1990) Evaluation of active versus passive AIDS surveillance in Oregon. American Journal of Public Health 80:463~64. Moss, A. R., Bachetti, P., Osmond, D., Meakin, R., Keffelew, A., and Gorter, R. (1989) Seroconversion for HIV in intravenous drug users in San Francisco. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Murphy, D. L. (1987) Heterosexual contacts of intravenous drug users: Implications for the next spread of the AIDS epidemic. Advances in Alcohol and Substance Abuse 7:89-97. National Hemophilia Foundation. (1988) What you should know about hemophilia (brochure). National Hemophilia Foundation, New York. New York State Department of Health. (1989) AIDS in New York State through 1988. Albany: New York State Department of Health. Novick, L. F., Berns, D., Stricof, R., Stevens, R., Pass, K., and Wethers, J. (1989) HIV seroprevalence in newborns in New York State. Journal of the American Medical Association 261: 1745-1750. Ognjan, A., Markowitz, N., Pohlod, D., Lee, H., Belian, B., and Saravolatz, L. D. (1989) HIV-1 and HTLV-1 infections in intravenous drug users (IVDUs) in Detroit, 1985-1989. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Olin, R., and Kall, K. (1989) HIV status and changes in risk behavior among arrested and detained intravenous drug abusers in Stockholm, 1987-88. Presented at the Fifth International Conference on AIDS, Montreal, June =9. O'SulIivan, M. J., Fajardo, A., Ferron, P., Efantis, J., Senk, C., and Duthely, M. (1989) Seroprevalence in a pregnant multiethnic population. Presented at the Fifth Internatior~al Conference on AIDS, Montreal, June =9. Padian, N., Moreno, A., Glass, S., Shiboski, S., and Maisonet, G. (1989) Ethnic differences in the heterosexual transmission of HIV in California. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Pappaioanou, M., George, R., Hannon, H., Hoff, R., Willoughby, A., et al. (1989) National surveys of HIV seroprevalence in women delivering live children in the United States. Presented at the Fifth International Conference on AIDS, Montreal, June =9.

78 ~ AIDS: THE SECOND DECADE Pansh, K. L., Mandel, J., Thomas, J., and Gomperts, E. (1989) Prediction of safer sex practice and psychosocial distress in adults with hemophilia at risk for AIDS. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Prendergast, T. J., Maxwell, R., Greenwood, J. R., Burrell, P., and Swatzel, C. (1989) Incidence and prevalence of HIV infection during 44 months of testing prostitutes/IVDUs in the women's prison, Orange County, California. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Public Health Service (PHS). (1988) Report of the Second Public Health Service AIDS Prevention and Control Conference. Public Health Reports 103 (Supplement No. 11:1-98. Quinn, T. C., Glasser, D., Cannon, R. O., Matuszak, D. L., Dunning, R. W., et al. (1988) Human immunodeficiency virus infection among patients attending clinics for sexually transmitted diseases. New England Journal of Medicine 318:197-203. Ragni, M. V., Gupta, P., Rinaldo, C. R., Kingsley, L. A., Spero, J. A., and Lewis, J. H. (1988) HIV transmission to female sexual partners of HIV antibody-positive hemophiliacs. Public Health Reports 103:5~58. Ralph, N., and Spigner, C. (1986) Contraceptive practices among female heroin addicts. American Journal of Public Health 76:101~1017. Rosenbaum, M. (1981) When drugs come into the picture, love flies out the win- dow: Women addicts' love relationships. International Journal of the Addictions 16:1197-1206. Schoenbaum, E. E., Hartel, D., Selwyn, P. A., Klein, R. S., Davenny, K., et al. (1989) Risk factors for human immunodeficiency virus infection in intravenous drug users. New England Journal of Medicine 321:87~879. Selik, R., and Petersen, L. (1989) Epidemiology of AIDS associated with intravenous drug use, United States, 1979-1988. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Selwyn, P. A., Schoenbaum, E. E., Hartel, D., Klein, R. S., Davenny, K., et al. (1988) AIDS and HIV-related mortality ire intravenous drug users (IVDUs). Presented at the Fourth International Conference on AIDS, Stockholm, June 12-16. Smiley, M. L., White, G. C. IL Becherer, P., Macik, G., Matthews, T. J., et al. (1988) Transmission of human immunodeficiency virus to sexual partners of hemophiliacs. American Journal of Hematology 28:27-32. Snyder, F., and Myers, M. (1989) Risk-taking behaviors of intravenous drug abusers. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Sowder, B., Weissman, G., and Young, P. (1989) Working with women at risk in a national AIDS prevention program. Photocopied materials distributed at the Fifth International Conference on AIDS, Montreal, June ~9. Sperling, R. S., Sacks, H. S., Mayer, L., Joyr~er, M., and Berkowitz, R. L. (1989) Umbilical cord blood serosurvey for human immunodeficiency virus in parturient women in a voluntary hospital in New York City. Obstetrics and Gynecology 73:179-181. Stoneburner, R. L., Des Jarlais, D. C., Benezra, D., Gorelkin, L., Sotheran, J. L., et al. (1988) A larger spectrum of severe HIV-1 related disease in intravenous drug users in New York City. Science 242:91~919. Stoneburner, R. L., Chiasson, M. A., Weisfuse, I. B., and Thomas, P. A. (1990) The epidemic of AIDS and HIV-1 infection among heterosexuals in New York City. AIDS 4:99-106.

STATE OF THE EPIDEMIC ~ 79 Thompson, E. G., Suarez, L., Reed, C. M., Buchanan, B. K., and Therrell, B. L. (1989) Serosurvey of women of childbearing age in Texas, U.S.A. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Tross, S., Abdul-Quader, A. S., Des Jarlais, D. C., Kouzi, A. C., and Friedman, S. R. (1989) Determinants of sexual risk reduction in street-recruited female IV drug users. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Turner, C. F. (1989) Research on sexual behaviors that transmit HIV: Progress and problems. AIDS 3:S63-S71. Turner, C. F., Miller, H. G., and Barker, L. (1989) AIDS research and the behavioral and social sciences. In R. Kulstad, ea., AIDS, 1988: AAAS Symposium Papers. Washington, D.C.: American Association for the Advancement of Science. Tumer, C. F., Miller, H. G., and Moses, L. E., eds. (1989) AIDS, Sexual Behavior, and Intravenous Drug Use. Washington, D.C.: National Academy Press. U.S. Bureau of the Census. (1987) Statistical Abstract of the United States: 1988. 108th ed. Washington, D.C.: U.S. Government Printing Office. Valdiserri, R. O., Bonati, F. A., Proctor, D., and Glaser, D. A. (1988) HIV antibody testing in a family planning clinic setting. New York State Journal of Medicine 88:623-625. Valdiserri, R. O., Arena, V. C., Proctor, D., and Bonati, F. A. (1989) The relationship between women's attitudes about condoms and their use: Implications for condom promotion programs. American Journal of Public Health 79:499-501. van Haastrecht, H. J. A., van den Hock, J. A. R., and Coutinho, R. A. (1989) No trend in yearly HIV seroprevalence rates among IVDU in Amsterdam: 198~1988. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Watters, J. K. (1988) Meaning and context: The social facts of intravenous drug use and HIV transmission in the inner city. Journal of Psychoactive Drugs 20:173-177. Weiner, S. M., Kaltenbach, K. A., and Finnegan, L. P. (1989) Drug abuse and pregnancy: Concomitant risk factors for HIV infection. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Weissman, G. (1988) Community outreach and prevention: A national demonstration project. Presented at the 116th annual meeting of the American Public Health Association, Boston, November. Wermuth, L., and Ham, J. (1989) Perception of AIDS risk among women sexual partners of intravenous drug users in San Francisco. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Whittaker, S., Calsyn, D., Saxon, A., and Freeman, G. (1989) Sexual behaviors of intravenous drug users. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Wilson, P. A., and Wasserman, K. (1989) Psychosocial responses to the threat of HIV exposure among people with bleeding disorders. Health and Social Work August: 17~183. Winkelstein, W., Samuel, M., Padian, N. S., Wiley, 3. A., Lang, W., et al. (1987) The San Francisco Men's Health Study: IlI. Reduction in human immunodeficiency virus transmission among homosexual/bisexual men' 1982-1986. American Journal of Public Health 77(91:685~89.

80 ~ AIDS: THE SECOND DECADE Winkelstein, W., Wiley, J. A., Padian, N. S., Samuel, M., Shiboski, S., et al. (1988) The San Francisco Men's Health Study: Continued decline in HIV seroconversion rates among homosexual/bisexual men. American Journal of Public Heals 78:1472- 1474. Withum, D. G., LaLota, M., Holtzman, D., Bum, E. E., Chan, M. S., et al. (1989) Prevalence of HIV antibodies in childbearing women in Florida. Presented at the Fifth International Conference on AIDS, Montreal, June =9. Wolfe, H., Keffelew, A., Bacchetti, P., Meakin, R., Brodie, B., and Moss, A. R. (1989) HIV infection in female intravenous drug users in San Francisco. Photocopied materials distributed at the Fifth International Conference on AIDS, Montreal, June =9. Worth, D., and Rodriguez, R. (1987) Latina women and AIDS. Siecus Report January/February: 5-7. Worth, D., Drucker, E., Eric, K., Chabon, B., Pivnick, A., and Cochrane, K. (1989) An ethnographic study of high-nsk sexual behavior in 96 women using IV heroin, cocaine, and crack in the South Bronx. Presented at the Fifth International Conference on AIDS, Montreal, June =9.