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.
3 The Future Course of the Epidemic and Available National Resources Short- and long-term estimates of the magnitude, pattern, and trends of AIDS and other HIV-related conditions are crucial to health care planning efforts and to the design of prevention and treatment strategies. One of the major problems in planning these efforts has been that, because AIDS is a relatively new disease and HIV is an unusual virus, there is little previous experience on which to base predictions about the epidemic's behavior. This chapter describes what can be projected on the basis of present knowledge and the resources that can be brought to bear on current and anticipated problems. (Chapter 5 discusses the implications of current projections for the provision and financing of health care and psychosocial support for those with HIV-related conditions. Chapter 6 identifies epidemiologic and other areas of research that must be pursued so that better predictions can be made.) PROJECTIONS BY THE PUBLIC HEALTH SERVICE Following a June 1986 planning conference at Coolfont, Berkeley Springs, West Virginia, the Public Health Service (PHS) issued updated projections of the incidence and prevalence of AIDS by 1991 (see Appendix G). Following is a summary of the major projections made by the PHS: · There are 1 million to 1.5 million Americans currently infected with HIV. Of these, 20 to 30 percent are expected to develop AIDS by 1991. 85
86 CONFRONTING AIDS · By the end of 1991 there will have been a cumulative total of more than 270,000 cases of AIDS in the United States, with more than 74,000 of those occurring in 1991 alone. · By the end of 1991 there will have been a cumulative total of more than 179,000 deaths from AIDS in the United States, with 54,000 of those occurring in 1991 alone. · Because the typical time between infection with HIV and the de- velopment of clinical AIDS is four or more years, most of the persons who will develop AIDS between now and 1991 already are infected. · The vast majority of AIDS cases will continue to come from the currently recognized high-risk groups. · New AIDS cases in men and women acquired through heterosexual contact will increase from 1,100 in 1986 to almost 7,000 in 1991. This figure includes those heterosexuals reporting contact with people known to be infected or with people in known high-risk groups and heterosexuals who are presumed to have acquired the disease from contact with individuals not known to be in such groups. · Pediatric AIDS cases will increase almost 10-fold in the next five years, to more than 3,000 cumulative cases by the end of 1991. PROBLEMS IN MAKING PROJECTIONS There are substantial uncertainties about such factors as the prevalence of HIV infection, the rate of transmission of the virus among various population groups, and the risks of disease among those infected. Accordingly, any projection of the future incidence and prevalence of AIDS (whether by the PHS or by others) will be subject to considerable uncertainty. Nevertheless, empirical projections of the incidence, prevalence, and cost of AIDS, however crude or uncertain, are essential for planning a response to the epidemic. The critical issue is to identify the value and limitations of such projections and their policy implications so that improved projections of the burden of disease can be developed. Also, by assessing the limits of such models, the data that need to be collected can be identified. The PHS estimates of the incidence of AIDS were derived from an empirical model based on a statistical trend analysis of AIDS cases reported to the CDC through May 1986 (Morgan and Curran, 1986~. A very similar statistical model was used in earlier projections based on reported cases through mid-1985 (Curran et al., 1985; W. M. Morgan, Centers for Disease Control, personal communication, 19861. Such mod- els depend on the assumption that observed trends in a disease, such as the distribution of cases by age, sex, geographic location, and risk group,
THE FUTURE COURSE OF THE EPIDEMIC 87 will not change with time. They are adequate for reasonable short-term projections but are of far more limited use for long-term projections (e.g., rr.ore than five years). Obviously, more complex models that incorporate known information on the sizes of populations at risk, viral transmission and infectivity, and the natural history of HIV infection and its associated diseases would be expected to yield more accurate, and thus more valuable, predictions. However, the data in those areas necessary to construct such models were considered by the committee's Epidemiology Working Group to be limited in the following ways: 1. There are no survey data that can be considered to accurately represent the general population. Surveys to date include those of blood donors seen at blood banks after voluntary deferral of donation by high-risk individuals was requested; applicants for military service, who are a disproportionately young, minority, and economically disadvan- taged population; and members of high-risk groups. These groups are almost certainly not representative of the general population, and how to analyze data obtained from them to deduce what is happening in the general population is not known. 2. In high-risk groups (for example, homosexual men) there is wide variation among communities in the prevalence of disease and sero- positivity, based on location, age, and possibly frequency of high-risk behaviors (such as anal intercourse). This makes difficult the estimation of national prevalence in high-risk groups, or even estimates of the likely spread within these groups. 3. There are major differences as to the time when the virus was introduced into communities in various parts of the country, even among the same high-risk groups. Given the long and uncertain time lag between lIIV infection and symptoms of that infection, this variation makes extrapolation from the number of AIDS cases meeting the CDC definition to the likely number of infected persons at a given time nationwide very tenuous. 4. The natural history of the disease is not yet fully defined. The proportion of infected persons who will develop AIDS or ARC is not yet known, nor is the time frame for the occurrence of these conditions. Thus, estimates useful in health care planning, such as hospital days required for treatment or days of work lost, are very difficult to derive. Empirical models do not have to take into account these poorly understood factors to enable projections of the epidemic's future. How- ever, they suffer from their own set of uncertainties. Though case reports to the CDC constitute the most reliable source of analysis of AIDS trends, such data have important limitations. First, the CDC criteria (Appendix
88 CONFRONTING AIDS E) are undoubtedly too restrictive to include all serious manifestations of HIV infection. If the mix of manifestations of HIV infection changes over time, the predictions from empirical models will be inaccurate. Second, cases of AIDS that meet the CDC criteria may be underre- ported, although the extent of underreporting is not known reliably. As long as underreporting rates have not varied over time, empirical analysis of time trends in reported AIDS cases would remain unbiased. However, increasing awareness of the main modes of acquiring AIDS and its decreasing novelty make it plausible that underreporting has increased as the disease has become more common, a phenomenon that has occurred with other diseases. If so, purely empirical models of AIDS trends will show a spurious deceleration of the epidemic. Third, there are delays in reporting cases to the CDC. Accordingly, the time series of cumulatively reported cases understates the actual number diagnosed, especially for more recent months. In an empirical analysis of trends in the incidence of AIDS, the observed data on diagnosed cases need to be corrected for reporting lags. Because such corrections will mostly affect recent data, apparently minor changes in the correction method can significantly affect distant projections from time trend mod- els. The method that CDC uses to correct for such delays assumes that the distribution of delays between the diagnosis and the report of the case to the CDC remains constant over time (Curran et al., 19851. Such an assumption needs careful and regular scrutiny. Fourth, statistical confidence intervals (see Appendix G) surrounding future projections from empirical models are mathematically and biolog- ically problematic because there is little basis for estimating the distribu- tion of errors. Fifth, projections of the prevalence of AIDS are based not only on projections of the incidence of AIDS but on estimates of the life expectancy of future AIDS victims. If changes in the natural history of the epidemic or improvements in medical care result in prolongation of life for AIDS patients, the prevalence of the disease would rise even faster than the incidence. The prevalence of disease is an indicator of the number of AIDS patients that will be alive and in need of health care. Cases of AIDS that are diagnosed in the near future will reflect the consequences of past infection. Therefore, despite the uncertainties discussed above, the PHS projections (and those of other purely empirical models) are likely to be highly accurate in the short term. Consequently, despite the fact that current projection methods are crude, it is reasonable to assume that the rising incidence of AIDS will not soon reverse itself. Disease and death resulting from HIV infection are likely to be increasing 5 to 10 years from now and probably into the next century. The committee believes that the PHS estimates are reasonable at this
THE FUTURE COURSE OF THE EPIDEMIC 89 time and supports their use for planning purposes. However, its accep- tance of these projections does not imply that they are precise, nor does it obviate the need to continue to acquire data that will permit the construction of more sophisticated models. THE EPIDEMIC WITHIN AND BEYOND HIGH-RISK GROUPS The populations at highest risk for HIV infection in the near future will continue to be homosexual men and IV drug users, but no accurate data exist on the size of these groups. It is also not known with what frequency homosexual men practice the behaviors (primarily receptive anal inter- course) that put them at high risk of HIV infection. Thus, the distribution of risk within the total group of men at risk because of homosexual activity is not known, nor can the likely rate of spread be calculated. Estimates of the overall percentage of the male homosexual population infected with HIV must take into account the definition of the population under consideration. Lower estimates of the prevalence of seropositivity are usually associated with larger estimates of the total homosexual population (encompassing individuals who presumably have had fewer homosexual encounters). Given this consideration, the committee's Epi- demiology Working Group estimated that seropositivity among male homosexuals ranges from over 50 percent in some areas for men who have had a large number of partners to under 20 percent for a population including any individual who has participated in homosexual activity. By far the largest number of persons now seropositive in the United States presumably acquired their infection through homosexual activity. However, as discussed in Chapter 4, there is evidence that the spread of HIV through homosexual activity has slowed. The trend is attributed to behavioral change in response to the AIDS epidemic, but it may also be that many persons in the highest-risk subgroup (those with the most partners) have already been infected. HIV infection will probably con- tinue to spread in homosexual males, although possibly at a slower rate because of the use by some of "safer sex" practices (e.g., avoidance of intercourse with infected persons, increased use of condoms, and avoid- ance of anal intercourse). The numbers infected and at risk among IV drug users are even more difficult to estimate. The total number of IV drug users in the United States is not known, and persons move in and out of the group rather frequently. Although evidence indicates that there has been some modi- fication of behavior in response to the AIDS epidemic, behavioral modification is much less pervasive in this high-risk group than among male homosexuals (see Chapter 4~. In locales such as New York City where needles and syringes are extensively shared, many IV drug users
90 CONFRONTING AIDS may have already been exposed, but this may not be true for other urban centers. Continuing spread of HIV in IV drug users throughout the United States is expected in the future. There is a broad spectrum of opinion about the extent of the likely spread in the United States of HIV infection in the heterosexual popula- tion, but there is strong agreement that the present surveillance systems have only limited capacity to detect such spread. Because of the much larger size of this population as compared with the recognized high-risk groups, there is potential for wide-ranging estimates. Opinions provided to the committee by members of the Epidemiology Working Group ranged from the estimate of HIV infection as a minor problem among heterosexuals to an estimate that perhaps millions of heterosexuals who have multiple sex partners or who patronize prostitutes will ultimately be affected. In central Africa, bidirectional heterosexual transmission is believed to be the dominant mode of transmission (Mann, 19861. Interpretations of the data from Africa are complicated, however, by large numbers of sexual partners and by frequent prostitute contact among heterosexual African AIDS patients and by reports of repeated use of unsterile needles and syringes in many medical care settings. Whatever the efficiency of heterosexual transmission, it is clear that the infection will continue to be amplified among populations in countries or regions with a high prevalence of infection by frequent transfusion of blood (unless screening of the blood supply begins), by the vertical transmission of infection to mother and child, and by the continued medical use of unsterile needles. Thus, the disease will continue to increase dramatically in those areas. In the United States, where such amplification will generally not be present, it is presumed that heterosexual spread will be slower. However, IV drug use in some communities or groups may amplify sexual trans- mission. Much of the male-to-female spread of HIV infection in the United States has been associated with IV drug use and has been confounded by the possibility that the women are also IV drug users. The relatively high seropositivity in some prostitute groups has also been attributed to IV drug use. A small amount of data is beginning to appear on the proportion of male homosexuals who also have heterosexual contact. The figure may be as high as 10 to 20 percent, and these individuals represent a large reservoir for potential infection of women and their offspring and for further heterosexual spread of infection. In this regard, it should be noted that the PHS projections of the future number of heterosexually acquired cases are based on the observations of the heterosexual spread that has occurred thus far in the epidemic predom- inantly heterosexual transmission from individuals who became infected through IV drug use or homosexual activity (as in the case of bisexuals). .
THE FUTURE COURSE OF THE EPIDEMIC 91 It is not known how the infection will behave when spread in that segment of the heterosexual population which has no other risk factors (Winkelstein et al., 19861. Overall, the committee concludes that over the next 5 to 10 years there will be substantially more HIV infections in the heterosexual population and that these cases will occur predominantly in those subgroups of the population at risk for other sexually transmitted diseases. These cases are expected initially to occur mainly in the geographic areas and among the demographic groups that already have a high frequency of AIDS or IV drug abuse. In addition, increased HIV infection in infants is expected as more women in their childbearing years become infected, but this may be moderated by screening, as discussed in Chapter 4. (See Chapter 6 for recommendations of studies that would track the course of the epidemic through heterosexual contact and permit interventions to be appropriately targeted.) THE PROPORTION OF SEROPOSITIVE INDIVIDUALS WHO WILL DEVELOP AIDS Opinions in the Epidemiology Working Group varied widely regarding the proportion of seropositive persons who will eventually die of HIV- related causes. At this time, there are only five years of observations on which to base such predictions. The data now available show that the proportion of a cohort of seropositive individuals that have progressed to AIDS is still rising five years after infection. Furthermore, once infected, a person may well remain at risk of clinical disease for life. With some of the less common clinical manifestations, particularly those that are necrologic, there may be a very long delay after infection. The estimate provided to the committee by the Epidemiology Working Group was that 25 to 50 percent of seropositive persons will develop AIDS as defined by the CDC within 5 to 10 years of seroconversion, and that a higher percentage cannot be ruled out on the basis of present studies. This estimate is consistent with but goes beyond that of the PHS, which projected that 20 to 30 percent of currently seropositive individuals will be diagnosed with AIDS within 5 years (Appendix G). In addition, there is an increasing number of reports of manifestations of HIV infection that fall outside the CDC definition of AIDS, which therefore modify projections upward. LONG-TERM PROSPECTS HIV infection is likely to continue to spread among those individuals who engage in behavior known to transmit the virus. HIV infection behaves somewhat like hepatitis B. but a number of factors make it
92 CONFRONTING AIDS difficult to predict whether it will reach or exceed the prevalence of that disease. These factors include the asymptomatic period after HIV infec- tion, which facilitates "silent" spread of the virus, the presumed lifelong infectivity of all infected individuals, the lack of data on the efficiency of transmission, and the difficulty of predicting changes in behavior and transmission that education may generate. There is also insufficient knowledge at this time to predict how the virus will evolve in its apparently new (human) host. Therefore, it is impossible, whether by model or by analogy, to predict the long-term course of the epidemic with any degree of certainty. It is clear, however, that reducing transmission is a difficult proposi- tion. Because no vaccine is likely to become available in the near future (see Chapter 6) and because of the seriousness of the disease, the only prudent course of action is an immediate, major effort to stop the further spread of infection through public health measures, particularly educa- tion. Any delay will bequeath to future policymakers a problem of potentially catastrophic proportions and will condemn many thousands of individuals to infection and disease. NATIONAL RESOURCES FOR DEALING WITH AIDS AND HIV There are many organizations, groups, and individuals that could be drawn upon to address aspects of the public health measures (Chapter 4), health care (Chapter 5), or research (Chapter 6) related to AIDS and HIV infection. Additionally, the epidemic has prompted the development of new groups to address certain problems and the extension of existing groups into new areas. This is particularly true of male homosexuals, who have developed community support groups. Former drug users have also conducted educational efforts. There are many resources spread across both the public and private sectors; thus, many of the needed actions can and perhaps should be undertaken by groups at both levels. A complete inventory of the national resources available for dealing with AIDS and HIV infection would include information on existing and potential activities or areas of concern (e.g., research, health care, education), on the nature of each resource (e.g., pharmaceutical com- pany, community group), and on the level of activities (e.g., national, state, local, risk group). Appendix D lists various groups and organiza- tions that are already active or that could be enlisted to work against AIDS and HIV infection. Impediments to Involvement Many of the groups and organizations listed in Appendix D have al- ready been engaged in problems related to AIDS and HIV infection that
THE FUTURE COURSE OF THE EPIDEMIC 93 are appropriate to their capabilities, often with considerable success. But, in the judgment of the committee, owner important resour`;~s In uoc~ `~c public and private sector have not yet become appropriately engaged or have addressed relevant problems inadequately. The reasons for noninvolvement vary, but they include the following: ]; ~, . . .. .. · lack of awareness about the magnitude of the problem and the needs · slowness in responding to obvious needs · apparent reluctance or inability to pursue particular programs be- cause of political, ideological, moral, or religious considerations, includ- ing the social stigma associated with some of the risk groups · lack of appropriate recruitment efforts · lack of funds or other resources (e.g., facilities) for pursuing prom- ising opportunities · lack of appropriate inducements to enter the field (e.g., stability of funding) · perceptions regarding the availability of reagents or other resources necessary for productive research · specific commercial disincentives (e.g., liability for vaccine-related injuries, uncertainty over market size and public health policy) · insufficient development of the basic research data base upon which further commercial development might proceed · uncertainty with regard to federal agency responsibilities Some of these impediments may be removed by relatively simple actions recommended in other parts of this report. Others are more complex and may require new mechanisms or more time to be reduced. Mechanisms for Coordinating Activities Mechanisms exist for coordinating certain facets of the overall ap- proach to AIDS and other HIV-related problems. Within the executive branch, the Public Health Service has developed a plan to guide its constituent agencies (Appendix G). The committee concurs in general terms with the overall goals and approaches outlined in this plan. However, the plan focuses on only part of the federal government's activities and potential. Though the Public Health Service is a significant resource, it represents only a portion of the national capacity to address the problems caused by HIV infection. Other groups, such as the military and the Department of Education, are also well situated to conduct certain types of epidemiologic or clinical research. No individual in the Public Health Service currently has primary responsibility for identifying priorities in implementing the PHS plan, but a Public Health Service AIDS Task Force has been established and a coordinator appointed. Certain activities will require contributions from
94 CONFRONTING AIDS various federal agencies outside the PHS, such as the Department of Defense and the Department of Education. Furthermore, many activities outlined in the plan might be better conducted with broad participation of nonfederal groups. No formal mechanisms exist for ensuring efficient collaboration in these areas. Certain task forces under the AIDS coordinator of the Public Health Service have responsibility for monitoring efforts conducted by the PHS agencies with their purview. The committee believes that while these task forces may promote communication within and between federal agencies, they sometimes have not sufficiently engaged or informed other relevant national resources and have not identified priorities and devised or articulated the strategic plans necessary to attain the desired goals in the shortest possible time. In many areas, there are no mechanisms for ensuring concerted action against AIDS and HIV infection, especially where activities involve diverse public and private sector bodies. Five important areas where this is the case are (1) vaccine and drug development, (2) epidemiologic and natural history studies in the United States and abroad, (3) evaluation of models for the appropriate care of HIV-associated conditions, (4) the financing of that care, and (5) the U.S. contribution to international efforts. Each of these subjects is considered in detail in the chapters that follow. There is a need to mobilize all existing resources through more effective interaction between the public and private sectors. To meet this need, and also to inform the American public, Congress, and the executive branch, the committee proposes the establishment of a National Commission on AIDS. Such a body should be advisory to existing administrative entities. (For fuller discussion of this recommendation, see Chapter 1.) REFERENCES Curran, J. W., W. M. Morgan, A. M. Hardy, H. W. Jaffe, W. W. Darrow, and W. R. Dowdle. 1985. The epidemiology of AIDS: Current status and future prospects. Science 229: 1352-1357. Mann, J. M. 1986. The epidemiology of LAV/HTLV-III In Africa. P. 101 in Abstracts of the Second International Conference on AIDS, Paris, June 23-25, 1986. Morgan, W. M., and J. W. Curran. 1986. Acquired immunodeficiency syndrome: Current and future trends. Pub. Health Rep. 101:459-465. Winkelstein, W., J. A. Wiley, N. Padian, and J. Levy. 1986. Potential for transmission of AIDS-associated retrovirus from bisexual men in San Francisco to their female sexual contacts. JAMA 255:901.