previous chapter
Medical Research on AIDS in Africa: A Historical Perspective
next chapter


Medical Research on AIDS in Africa:
A Historical Perspective

Randall M. packard
Paul Epstein

The history of Western medical research on AIDS in Africa closely resembles earlier attempts by Western-trained medical researchers to understand the epidemiology of infectious diseases, such as tuberculosis and syphilis, that were known in the West but that appeared to exhibit different epidemiological patterns in Africa. Like research into TB and syphilis, early inquiries into AIDS in Africa attempted to understand why African experience with the disease differed from Western experience. All of these efforts were handicapped by the limited state of Western knowledge about these diseases, an absence of adequate epidemiological data for Africa, and a lack of knowledge about the African societies and cultures within which these diseases occurred. Despite these shortcomings, early medical researchers quickly constructed theories to explain the peculiarities of the African disease experience. These theories were strongly influenced by cultural assumptions about Africa and Africans and tended to focus on the peculiarities of African behavior.[1] Once these theories were constructed, they shaped the course of subsequent research, privileging certain lines of inquiry while largely excluding or marginalizing other potentially important areas of research.

In this essay we compare the development of AIDS research in Africa with the history of earlier efforts by Western medical professionals to understand the epidemiology of TB and syphilis. By drawing these parallels, we hope to contribute to a clearer understanding of how Western medical ideas about AIDS in Africa developed and how these ideas have

A different version of this essay is forthcoming in Social Science and Medicine and is reprinted with the editor's permission.


shaped the direction and boundaries of African AIDS research and, ultimately, our understanding of the epidemiology of AIDS in Africa.

Tuberculosis And The "Dressed Native"

Early discussions about the causes of black susceptibility to TB centered on the problem of explaining why Africans had higher rates of morbidity and mortality than Europeans. At the time little was known about the nature of host resistance to the disease or about the role of cofactors in the transmission of infection and in the progression of infection to active disease. Research on TB was in fact in transition from the hereditary arguments of the late nineteenth century to Koch's germ theory. In addition, knowledge of African social and economic life was limited and was infused with racial and cultural stereotypes. Predictably, the explanations of European medical authorities came to reflect the perceptions about Africans that were current in European colonial society.

Central to these perceptions was the image of the "primitive native" making a difficult adjustment to conditions of a "civilized" industrial world. This image, embedded in European discussions of African morality, political participation, and labor skills, came to influence early explanations of TB in Africa. Africans were viewed as more susceptible to TB because they had not adjusted to the conditions of a civilized industrial society; their incomplete adoption of Western clothing and their failure to observe "proper" dietary and sanitary laws symbolized this lack of adjustment.

At the same time, it was argued that Africans who remained in their customary rural environment, working in the open air and wearing traditional attire, were generally healthy. These explanations for African sickness and health, focusing attention on the Africans' maladjustment to civilization, placed responsibility for the adverse living conditions of Africans squarely on the shoulders of Africans themselves and deflected attention from the low wages and inadequate housing policies of employers and government officials. More importantly, these explanations shaped the development of TB control measures, which came to focus naturally on education rather than on social and economic reform. Blacks had to be taught about the dangers of living in overcrowded housing and eating nutritionally inadequate diets, as if they chose to do so out of perversity rather than out of economic necessity.


Later on, discussions about African susceptibility to TB became infused with biological arguments that focused on the Africans' lack of experience with the disease and their consequent lack of physiological resistance to it. Like earlier behavioral arguments, physiological models defined the African as essentially different from the European, as the "other," and at the same time placed responsibility for the disease on the victim.[2] Not until the middle of this century did health officials come to see that the adverse environmental conditions under which Africans lived were not of their own making. Even then, environmental reform efforts continued to be hampered by behavioral explanations that emphasized the Africans' difficult adjustment to the conditions of Western industrial civilization. Typical of this discourse is the following statement by the director of Kenyan medical services in 1963:

The African in his rural setting is strictly bound by tribal patterns of behavior, beliefs and customs. He is an integral part of his community and his thinking tends to be communal. … With the transposition to the town he forsakes the communal life for an individualistic life, unsupported by tribal rules and regulations. While forsaking these supports, he is not yet ready to adopt the codes and rules which have brought social stability to western civilizations. Furthermore, he is abandoning ingrained centuries of agricultural and pastoral tradition and learning the technical skills of an industrial world quite strange to him.[3]

Even today TB control programs in Africa continue to view TB as a behavioral problem, attributing treatment failures to "patient default" rather than to the government's failure or inability to cope with environmental factors that continue to generate new cases of this disease.

Syphilis And African Sexuality

The recent work of the historian Marc Dawson on the history of syphilis in East Africa provides another example of how earlier Western medical researchers came to construct a behavioral paradigm to explain the peculiarities of African disease experience and how these models shaped medical responses to this experience.[4] The epidemiology of syphilis, like that of tuberculosis, was not well understood by Western medical researchers during the first decades of this century. Specifically, the epidemiological and pathological differences between yaws, venereal syphilis, and endemic or nonvenereal syphilis had yet to be sorted out. As a result, there was considerable confusion among early medical


personnel working in Africa. This confusion led early European observers to regard the African experience with the disease as different from the European and to look for the reasons for this difference. As with TB, early theories about syphilis in Africa focused on behavioral theories that were infused with racial stereotypes.

Early medical researchers in East Africa concluded that between 50 and 90 percent of the African population in parts of Kenya and Uganda were infected with venereal syphilis. Col. F. J. Lambkin, a leading British expert on syphilis, who was seconded to Uganda to study the problem, concluded in 1906: "As things are at present, the entire population is in danger of being exterminated by syphilis in a very few years, or of being left a degenerate race fit for nothing."[5] In explaining this extraordinary situation, Lambkin concluded that the major cause of the epidemic was a breakdown of various Ganda social institutions. In this respect he echoed early medical opinions about the spread of TB, as well as later theories about AIDS. Specifically, Lambkin argued that Christianity had broken down customs that restricted the social movement of women. At the same time, sanctions against adultery had been eliminated at the behest of the British colonial government. These changes, he argued, had permitted Ganda women to engage in "promiscuous sexual intercourse and immorality ," resulting from "their natural immoral proclivities " (emphasis added). Lambkin further indicted the Bahima of Ankole as primary disseminators of the disease because of their practice of allowing a man's age-mates and visitors to have sex with his wife.[6]

Similar claims were made by observers in western Kenya. G. L. Gilks, discovering what he believed to be a major epidemic of venereal syphilis in Kavirondo, concluded: "The whole attitude of the native toward sexual matters renders it certain that venereal disease, once introduced, is bound to spread among old and young."[7]

On the basis of subsequent studies and a careful reexamination of the medical evidence, Dawson suggests that what Lambkin and Gilks were observing was not an epidemic of venereal syphilis but nonvenereal or endemic syphilis, which is caused by the same Treponema pallidum spirochete that causes venereal syphilis. Endemic syphilis, however, spreads through bodily contact in warm climates and in the absence of adequate sanitation. According to Dawson, syphilis was clearly being spread sexually into various parts of East Africa as a result of the development of migrant labor, commercial centers, military movements,


and a growing population of African prostitutes; however, its subsequent spread among large numbers of men, women, and children in rural and urban areas was via bodily contact.

The point of this episode is not simply that the disease was misdiagnosed. After all, the differences between yaws, venereal syphilis, and endemic syphilis were difficult to sort out and in fact were not clearly understood until the 1930s. The importance of this episode lies instead in the way these early medical observers constructed the medical evidence they were observing to fit preexisting assumptions about African sexuality and disease. Seeing a disease, which they assumed to be venereal syphilis, Lambkin, Gilks, and others readily constructed a theory to explain its extraordinary rate of spread. That theory was based on assumptions about the extreme sexuality of Africans—assumptions for which they had virtually no empirical evidence. This is not surprising. As Sander Gilman notes, the association of Africans with sexuality and the tendency to link African sexuality with disease have a long history in Western thought.[8] By the end of the nineteenth century, when European powers began carving out African colonies, the association could be found in many works of literature and art in continental Europe and held a central position in the constellation of ideas that made up European perceptions about Africans. As a result, early medical authorities, missionaries, and colonial administrators came to Africa with certain strong assumptions about African sexuality. There can be little doubt that these presumptions colored both their epidemiological findings and their control efforts. Following this behavioral explanation, these authorities advocated public health policies that centered largely on the development of measures, often draconian in nature, to control the behavior of prostitutes. At the same time, problems associated with living conditions and sanitation, which were in fact centrally important to the spread of endemic syphilis, were ignored.

Aids And The "Sexual Life Of The Natives"

Early discussions of AIDS in Africa developed in an intellectual environment similar to that in which early inquiries into TB and syphilis were conducted. When medical researchers first began studying AIDS in Africa, they quickly realized that the epidemiology of the disease was different from that in the West. The ratio of male to female cases was 13:1 in the West, whereas the ratio in Africa was nearly 1:1. This fact, combined with an apparent absence of known risk groups in the form


of either IV drug users or homosexuals, led early researchers to conclude that AIDS transmission in Africa was different from that in the West. They therefore tried to determine what it was about Africa and Africans that accounted for its peculiar pattern of transmission.

In trying to explain transmission patterns in Africa, AIDS researchers were handicapped by limited knowledge about the etiology of the disease. Thus, when African AIDS cases (or what seemed to be AIDS cases) first began appearing in Belgium in 1983, the infectious agent causing AIDS had not yet been identified. In addition, early discussions of AIDS in Africa occurred, and in fact continue to occur, in the absence of any clear understanding about the role of various cofactors in either the transmission or the progression of HIV infection. Finally, early AIDS researchers had only limited experience or knowledge of the societies and cultures within which AIDS was occurring in Africa.

This lack of social and medical knowledge, combined with the suspicion that the key to understanding AIDS in the West might lie in Africa, contributed to a great deal of speculation about the epidemiology of AIDS in Africa and encouraged researchers to construct hypotheses that often were based on extremely limited data. It is therefore not surprising that stereotypic images of Africa and Africans entered into the discourse on the epidemiology of AIDS in Africa.

Early reports on AIDS in Africa took a somewhat eclectic approach to the question of why African populations exhibited an epidemiological pattern different from that of populations in the West.[9] However, a number of influential Western AIDS researchers concluded early on that the apparently equal sex ratio of AIDS cases in Africa was most easily explained by a pattern of heterosexual transmission, a phenomenon relatively rare in the West at that time. This conclusion was supported by early prevalence studies, which seemed to indicate that both cases and HIV seropositivity were most frequent among sexually active adults. But why, then, was HIV occurring through heterosexual transmission in Africa and not to any great degree in Europe or America?

This question quickly led to two theories. The first argued that AIDS had existed in Africa for a longer period of time than in the West and therefore had reached a different stage in its epidemiological history. This theory—combined with the virological research of Essex, Gallo, and others on Simian T-Lymphocyte Retrovirus III in African green monkeys—led to arguments that AIDS originated in Africa.[10] This hypothesis was hotly debated on scientific grounds. More important, it ignited a political firestorm among African political leaders, who regarded


the theory as imperialist scapegoating.[11] This produced a political environment in which the cooperation of African governments in further AIDS research appeared to be in jeopardy. As a result, the Western medical research community appears to have put aside the question of African origins as well as investigations into the possibility that HTLVIII may have achieved a different epidemiological stage in Africa.[12]

The second theory put forth to explain the heterosexual transmission of HIV in Africa focused on African sexual behavior. In brief, it was argued as early as 1985 that the heterosexual transmission of HIV in Africa was the result of higher levels of sexual promiscuity among Africans, or, in the current language of social science research on AIDS, "poly-partner sexual activities." The middle-class businessman or bureaucrat with a string of lovers; the truck driver with sexual contacts all across the African map; and, above all, the pervasive female prostitute, who was said to have literally hundreds of contacts each year—these people were identified as the main vectors of HIV transmission in Africa. Although the association of AIDS in Africa with sexual promiscuity was challenged by both African observers and others with broad knowledge of African societies and cultures, it nonetheless persisted and, like earlier stereotypes concerning black susceptibility to TB and syphilis, became the central focus of medical inquiries into the problem of AIDS.

Why, given all the social and economic factors that distinguish African populations from those in the West, did researchers choose to focus on sexual promiscuity? Officially, the conclusion was said to be based on medical evidence, including studies which indicated that Africans who had multiple sexual partners and other STDs were statistically at higher risk of being infected with HIV than those who did not. Yet evidence for this conclusion, as well as for the association of HIV infection with the years of peak sexual activity, was extremely limited prior to 1986. Therefore, other factors probably contributed to the development of this explanation.

One such factor appears to have been the association of AIDS in the West with the alleged sexual promiscuity of homosexuals. In essence, AIDS was defined as a sexually transmitted disease that was spread in the West within a population defined as sexually promiscuous. Its heterosexual spread in Africa therefore implied similar levels of promiscuity. More than one Western AIDS researcher, in fact, suggested that African heterosexuals had a pattern of promiscuity similar to that of


promiscuous gay men in the United States and Europe (a conclusion that incorporated two discriminatory stereotypes).[13] More broadly, Western research on AIDS had already defined AIDS as a behavioral problem associated with aberrant life-styles.[14] There was thus a predisposition to look for "deviance" in an African setting.

Yet the role of sexual promiscuity in the spread of AIDS in Africa appears to have evolved in part out of prior assumptions about the sexuality of Africans. Thus, at a 1985 conference on AIDS held in the Central African Republic, Dr. Fakhri Assaad of the World Health Organization referred to the widespread pattern of "polygamy without wedding rings" in Africa. According to Assaad, this pattern resulted from the tendency of men to consider it a status symbol to have several wives or mistresses. He noted that, in reference to sexual and other cultural practices that might facilitate the transmission of HIV, that many of the practices had deep religious and traditional significance. In a similar fashion, researchers describing the sexual practices of patients suffering from "Slim Disease" in Uganda in 1985 observed that, although their subjects denied overt promiscuous behavior, they were "by western standards heterosexually promiscuous."[15] These researchers based their conclusion on the testimony of ten HIV-positive truck drivers who admitted to engaging in homosexual behavior. Labeling homosexual behavior as promiscuous by Western standards may have been consistent with the dominant view of Western society. Yet surely the issue in this case was not promiscuity but the fact that these subjects engaged in a known risk behavior. Nonetheless, it was not the homosexual behavior per se that was presented as the risk behavior, but the subjects' promiscuity.

Yet another example of this tendency to view Africans as sexually promiscuous can be found in a study that compared HIV infection among prostitutes and female controls in Rwanda. Commenting on their sampling procedures, the authors of this study noted, "Matching of the female controls was not extended to marital status since, for the age group studied, celibacy [single status] in women is unusual in Central Africa and commonly associated with prostitution " (emphasis added), a conclusion for which the authors presented not a shred of evidence.[16]

The parallel between these attitudes and those of early medical personnel faced with an epidemic of syphilis in East Africa, described above, along with the absence of any reliable data supporting these conclusions, forces us to ask whether the attitudes of medical researchers toward


AIDS in Africa did not also have their basis in a deeply imbedded image or trope which continues to shape Western medical and popular thought about African sexuality.

In a similar vein, it has been suggested that other cultural practices—such as scarification, the therapeutic use of razors, and female circumcision—might also play a role in the spread of HIV. Like sexual promiscuity, these risk behaviors were seen as culturally determined.

The Role Of Anthropologists

Having constructed AIDS as a behavioral problem resulting from particular culturally sanctioned practices, AIDS researchers turned to anthropologists for information on these practices. One might expect that social scientists with extensive African experience would have challenged the sexual stereotypes developed by medical researchers. Such challenges did not, however, occur to any great degree. Although recent social science research has questioned some of the behavioral assumptions underlying AIDS research in Africa, the activities of those social scientists who were most closely linked to the AIDS inquiry in its early stages tended to reinforce these assumptions.

Probably the main reason why anthropologists failed to challenge the dominant paradigm in AIDS research had to do with the conditions under which social scientists were brought into the AIDS inquiry. Instead of engaging in an open-ended dialogue with social scientists, assessing available data, and discussing how research methods and agendas might be modified to fit more closely with the contours of African experience, the medical research community expected the social scientist to adhere to the dominant behavioral model. Specifically, they asked anthropologists and other social scientists to provide information about the "risk behaviors" that might facilitate the transmission of AIDS. In other words, what were the practices, customs, or patterns of social intercourse that provided opportunities for HIV transmission? Constructed in this way, the question immediately narrowed the range of sociological data relevant for the discussion. The question became not "What is the social context within which HIV transmission occurs in Africa?" but, rather, "What are the patterns of behavior which are placing Africans at risk of infection?" While the first construction would have allowed for open-ended discussion of a wide range of social, political, and economic conditions that might be affecting health levels in Africa, the latter formulation quickly narrowed discussion to an inquiry


into the "customs of the natives." At the same time, it placed responsibility for transmission on the actors themselves in a not too subtle form of victim blaming.

The result of this emphasis on "risk behaviors" was that anthropologists found themselves being asked to dig through the ethnographic record on African cultures in order to identify possible patterns of behavior that might facilitate HIV transmission. This exercise resulted in conferences, workshops, and seminars in which the medical community was presented with an array of information that was often excised from its social or economic context and presented in much the same way as ethnographic artifacts are presented in natural history museums. Descriptions and pictures of scarification taken among the Nubia in the 1940s and 1950s were displayed as examples of possible "risk behaviors" involving blood transfers. From the broad array of data on African sexual practices imbedded in the ethnographic record of Africa, data revealing a wide range of patterns and extreme variation with regard to sexual permissiveness anthropologists presented only those cases that constituted possible "risk behaviors." Take, for example, the following description of "risk behavior" reported in a survey on "Social Factors in the Transmission and Control of AIDS in Africa" commissioned by the United States Agency for International Development (USAID):

There is a widespread fear of impotence [in Africa]. Our readings mention instances where an older man might ask a younger man to impregnate his wife. The Gwembe Tonga of Zambia use euphemistic invitation in these circumstances—"go and cut wood for me, my friend." … This illuminates our understanding of the perception of sexuality in certain traditional African settings but also indicates another—though limited—instance of a possible route for spreading AIDS through increasing the number of sexual partners.[17]

The same report contains a lengthy description of ritual sexual intercourse involving the widows of deceased men among the Giriama of Kenya. The authors follow the description with the observation "Clearly if the widow's deceased spouse was an AIDS victim then this custom will contribute to the spread of the disease."

In a similar vein, Daniel Hrdy, who is trained in both medicine and anthropology, wrote in an article on cultural practices relating to HIV transmission in Africa, "Although generalizations are difficult, most traditional African societies are promiscuous by Western standards. Promiscuity occurs both premaritally and postmaritally. For instance, in the Lese of Zaire, there is a period following puberty and before


marriage when sexual relations between young men and a number of women is virtually sanctioned by the society."[18] This article was circulated in advance to the 300 or so participants who attended a conference on Anthropological Perspectives on AIDS (sponsored by the U.S. Agency for International Development and the National Institute of Allergy and Infectious Diseases), presumably because the organizers viewed it as a model of the types of data they hoped would be presented at the conference.

These behaviors may very well occur and may be potential avenues for HIV transmission. Reports of this type—by concentrating on "risk behaviors"—exclude from discussion broader patterns of everyday sexual activity, which in many cases are both less exotic and more monogamous. Moreover, they reinforce, perhaps unintentionally, the impression that sexual promiscuity is culturally determined. For example, Edward Green, in a recent article on the role of behavioral scientists in African AIDS research, noted, "Changes in behavior which promote the spread of AIDS will go against social and cultural norms and values in Africa and against deeply ingrained behavioral patterns." Similarly, Francis Conant, writing in the same volume, concluded, "In dealing with AIDS we are not just dealing with sex; we are dealing with life-ways and complex cultural patterns."[19]

Some of the contributions by anthropologists have been of such questionable relevance to the issue of HIV transmission as to border on being salacious. Take, for example, Hrdy's extensive account of female circumcision.[20] After spending a page describing various patterns of female circumcision in detail, he concludes that there is hardly any correlation between areas in which it is practiced and the distribution of HIV infection, leaving the reader to question the need for the descriptive detail. This pattern of selective reporting only reinforced the popular image of African promiscuity and at the same time strengthened the assumption that the heterosexual epidemic of AIDS in Africa was simply a product of the peculiarities of African behavior.

Remarkably, some of these same anthropologists cautioned us not to make generalizations about African sexual behavior and suggested that the problem was not generalized sexual promiscuity but "urbanization": "Away from the social constraints imposed by commitments and obligations to a network of kin, there is the opportunity to engage in behaviors, including poly-partner sexual activities, that would be difficult to undertake in the home village due to social constraints."[21] In a similar fashion, Hrdy concluded: "As people leave rural villages and


migrate to urban areas, the general level of promiscuity usually increases. This increase may be attributable in part to the relaxation of traditional village values but appears to be due primarily to the destitution of poor migrant women, who may become prostitutes, and to the greater mobility and rootlessness of young male migrants and soldiers."[22] In short, urban promiscuity was the product of the loss of "traditional restraints." The image of the "detribalized" African, the bane of colonial urban authorities, was a central image in earlier discussions of black susceptibility to TB and syphilis. This image, which was fairly well excised from social science discussions in the 1970s, was being resurrected to explain the frequency of heterosexual transmission of HIV and Africans in the 1980s. This explanation both distinguished Africans from the West and placed responsibility for AIDS on the African. Moreover, given the behavioral thrust of the explanation, the recommended response was finding ways to modify urban sexual behavior. "An understanding of the patterns of population movement will help us to identify high-risk mobile populations and to focus educational resources before the virus is established in those populations."[23] While this stress on behavioral modification may not have been as manifestly self-serving as the above-described efforts of medical authorities working with TB to see overcrowding and malnutrition as the result of African ignorance, it shared a similar disregard for the root causes of African sexual patterns within an urban environment.

The Political Economy Of African Promiscuity

None of this is to deny that AIDS is transmitted heterosexually or that multiple sexual partners may in fact be a common pattern within the rapidly growing urban centers of Africa. However, explanations that viewed this pattern as either a cultural phenomenon or as a product of declining social constraints ignored the context within which urbanization is occurring in Africa. At the same time, by focusing attention on sexual promiscuity and other cultural behaviors, these explanations have deflected attention from other cofactors that may be as important for the heterosexual transmission of AIDS in Africa as frequency of sexual contacts. In addition, concerns about African sexual behavior limit efforts to explore other avenues of HIV transmission.

There is every reason to believe that, whatever cultural attitudes shape African sexuality, the tendency to have multiple sexual partners has


been encouraged by the separation of households. This separation has resulted from patterns of labor migration involving rural households, which often must send one or more of their members to seek wage employment in urban or industrial centers, including plantations and other large-scale agricultural projects. Labor migration, in turn, is the result of specific historical patterns of development in many parts of Africa—especially parts of Eastern Zaire, Tanzania, Rwanda, Uganda, Zambia, and Kenya. These areas all have large populations of impoverished rural households without access to land or labor and with few opportunities for acquiring income within the rural economy. As Hrdy rightfully notes, the numbers of men and women seeking employment in urban and industrial centers have increased dramatically since the early 1970s. What he does not indicate is why they have increased. Clearly, a major reason is that declining commodity prices and the increasing cost of agricultural inputs have made small-scale agricultural production unprofitable. The African small holder in many areas of East and Central Africa cannot make a living on the land other than on outgrower schemes, which are often highly exploitive in their treatment of growers. For these impoverished households, survival depends on access to some form of nonagricultural income, primarily wages. Yet employment opportunities are limited, in part because of the capital-intensive nature of many industries in Africa, and wages are often low, a product of lobbying efforts on the part of employers to ensure profits. All this has led to the creation of a class of semiproletarianized men and women, who work in urban and industrial settings but cannot afford to support their families there, and thus to the almost continual separation of rural households. For both men and women this existence fosters the development of "multiple sexual partners": Women who cannot find other employment often must work full or part time as prostitutes in urban and industrial settings; women left alone at home for long periods of time may take on "lovers"; and men may take on second "wives" near their place of employment.

This pattern of multiple sexual partners has undoubtedly also resulted from the political disruption of family life generated by warfare in places such as Mozambique, Angola, Burundi, and Uganda. Not only are families torn apart by these experiences, but the rape of rural women by marauding guerrilla armies must represent a particularly brutal form of "sexual poly-partnerism," which has little to do with cultural norms. Again, as with the impoverishment and disruption of rural households described above, it is important to understand the forces that have generateed


these wars, including the foreign governments that continue to support one side or the other to serve their own political ends.

It is important to understand that the pattern of multiple sexual partners is shaped by strong social, economic, and political pressures, and not simply by cultural norms. The presence of such forces will limit the success of efforts to control the spread of HIV through sex education, just as these factors have limited efforts to control population growth in Africa.

The parallel between AIDS prevention and birth control is in fact highly relevant. One of the principal obstacles to the success of population control programs in Africa and elsewhere in the developing world has been the economic pressures on parents to produce large families. These pressures, like those leading to the separation of households, have resulted from particular patterns of development that have created high demands for family labor. As long as this demand continues, unprotected sexual activity is going to occur with considerable frequency. For sexual activity is not simply about pleasure. It is also about social reproduction. If efforts to control the spread of HIV infection do not include policies that deal with the underlying causes of both family separation and the high demand for family labor, we may be fighting an uphill battle in trying to reduce the heterosexual transmission of AIDS in Africa through behavioral modification and condom use.

The Impact Of Medical Paradigms On Aids Research In Africa

The early contributions of social scientists to our understanding of the epidemiology of AIDS in Africa, therefore, were not very helpful. Instead of providing information that might have encouraged medical researchers to develop a broader perspective on the social and economic factors shaping the AIDS epidemic, social scientists contributed to a narrowing of research and to the development of a medical model centered on the problem of African sexuality. This paradigm has prevented researchers from exploring factors that may be of equal or greater importance in the transmission and progression of AIDS but which are not suggested by the paradigm. This, we believe, has resulted in a premature closure of African AIDS research. Two of these understudied areas are the role of high levels of background infection and malnutrition, and unsterilized needle use in the transmission and progression of HIV.


The Role Of Background Infections And Malnutrition In Aids

A number of studies have indicated a positive correlation between HIV seropositivity and the presence of other immunosuppressant conditions, such as malnutrition, tuberculosis, malaria, or trypanosomiasis.[24] Most of these studies assumed that these other conditions have followed on immune suppression caused by HIV infection. Yet a few studies suggest that these immune-suppressant conditions may have preceded HIV infection and facilitated its transmission. Lamoureaux and his colleagues, for example, noted the high incidence of TB in association with HIV infection among Haitians and Africans and evidence that in many cases infection with Micobacterium tuberculosis appears to have preceded HIV infection. They concluded:

We feel that the prevalence and persistence of M. tuberculosis infection in Africans and Haitians, along with the concomitant increase, due to the infection, in CD4+ lymphocytes and macrophages, which are the target cells of HIV, as well as the frequent provocation of an immunosuppressed state in such TB-bacillus-infected individuals, probably represents a common factor predisposing these two populations to infection with HIV when exposed to the virus.[25]

Similarly, Thomas Quinn and his colleagues reported that "the immune systems of African heterosexuals, similar to those of US homosexual men, are in a chronically activated state associated with chronic viral and parasitic antigenic exposure, which may cause them to be particularly susceptible to HIV infection or disease progression " (emphasis added). The authors noted in addition:

Our serological studies, as well as others, demonstrate that Africans are frequently exposed, due to hygienic conditions and other factors, to a wide variety of viruses, including CMV, EBV, hepatitis B virus and HSV, all of which are known to modulate the immune system. … Furthermore Africans in the present study are at additional risk for immunological alterations since they are frequently afflicted with a wide variety of diseases, such as malaria, trypanosomiasis, and filariasis, that are known to have a major effect on the immune system. The frequent exposure to these multiple microbial agents could act collectively or individually to result in immunological modulations rendering a host more susceptible to HIV infection or by influencing disease progression by increased viral replication and cytolysis of T4-positive cells.[26]


Quinn and his associates concluded: "Prospective studies are warranted in different population groups to examine the specific impact of these viral, bacterial, and parasitic infections and other antigenic stimuli on the susceptibility and development of HIV disease."

Unfortunately, such studies are extremely difficult to conduct, and to date none have been reported, although in NIAID's new International Collaboration in AIDS Research program, such studies will be encouraged.

Other studies have suggested that the infectivity of HIV-infected individuals may increase with disease progression. If declining immune function is in turn accelerated by the presence of concurrent infections or malnutrition, then the risk of transmitting infection, through heterosexual contact or otherwise, may be higher in Africans infected with HIV than in other HIV-infected persons who are not subject to the same levels of background infection.[27]

The one predisposing medical condition that has been examined in considerable detail is the role of genital ulcer diseases (GUDs), which are seen as disrupting genital epitheliums and thereby facilitating the sexual transmission of HIV.[28] Unfortunately, these studies have not controlled for the possibility that HIV is being transmitted by infected needles in the STD clinics where men and women with GUD go for treatment, rather than directly through sexual intercourse. Thus, the relationship between HIV transmission and genital ulcers remains clouded (see below for further discussion of the problem of needle transmission). We should note here that the high degree of attention given to GUDs, in contrast to the relative lack of attention to other infections diseases that might facilitate HIV transmission, is consistent with our argument that AIDS research in Africa has been narrowly focused on the problem of sexual behavior. In this regard it is interesting to note a recent New York Times article reporting the conclusions of AIDS researchers about the spread of AIDS. These researchers have concluded that there is considerable variation in the risk of HIV infection as the result of heterosexual contact. This risk does not appear to be related to frequency of unguarded contacts. Instead, it is evidently related to one of a number of possible cofactors. With the exception of levels of HIV infection, all the cofactors being considered are directly related to sexual contact.[29] Presumably no attention is being given to nonsexual factors such as those suggested here.

In short, the susceptibility of Africans to HIV and the facility with


which it is transmitted heterosexually may be a direct result of the high background levels of infection and malnutrition and other immunosuppressant conditions that exist in most African countries, and not simply a question of frequent sexual contacts or other "cultural practices."

If background infections do facilitate the transmission of HIV, then it seems likely that the population at greatest risk is not the urban middle class, who appear frequently in AIDS statistics, but the urban poor, who possess higher background levels of infection and malnutrition as a result of their impoverishment and lack of access to adequate medical care. Unfortunately, few studies have accounted for class, and those that have are inconclusive.[30]

If it is found that malnutrition and concurrent diseases that suppress the immune system predispose Africans to AIDS infection, and if HIV is transmitted through heterosexual contact, then the potential risk group may be very large indeed, and certainly is not limited to prostitutes, truck drivers, and bureaucrats. The risk group surely will include the rural families from which infected urban workers come and to which they eventually return. In this regard, one would expect to find high prevalence rates in the areas that have traditionally served as labor reserves for the cities which are currently centers for AIDS infection—for example, the Songea area of Tanzania; the Kwango region of Zaire; and areas of northeast Zambia that serve as labor reserves for the copper belt. These areas experience high rates of migration and are thus more likely to be exposed to AIDS infection emanating in the cities than other regions. They are also by definition impoverished regions and thus contain a population that may be particularly vulnerable to AIDS transmission.

Northeast Zambia is an interesting case. For it is not only a labor reserve area but also part of the so-called matrilineal belt of Africa. Hrdy, Brokensha, and Good have suggested that the rising incidence of AIDS in this area may be related to patterns of sexual promiscuity associated with matrilineal descent.[31] For example, Hrdy notes that "in the so-called 'matrilineal belt' centered in south-central Africa, there is an especially high degree of adolescent promiscuity and uncertainty about paternity." He goes on to describe inheritance patterns and concludes that "matrilineal inheritance … may reduce societal pressure to prevent promiscuity; matrilineal societies are often promiscuous societies." Whether or not this assertion is correct, linking matrilineality with HIV transmission represents a clear case of decontextualization, which ignores the wider social and economic conditions associated with participation


in a labor reserve economy. It is thus another example of how, in their quest for "risk behaviors," certain anthropologists have contributed to the decontextualization of African lives and to the narrowing of AIDS research. Clearly, the two explanations call for very different responses. Knowing which one is correct is therefore of considerable importance.

Needle Use: An Alternative Model

The second area of inquiry that appears to have been obscured as a result of the medical research community's fixation on African sexuality is the role of injections and transfusions in the spread of HIV. High levels of background infection are associated in Africa with high levels of needle use in a range of therapeutic settings, which may or may not observe adequate sterilization procedures. Where sterilization is not employed, there is a risk of HIV transmission. Although this area of risk has been recognized, it has been given inadequate attention.

In two control studies of infants and children in Kinshasa, Mann and his colleagues correlated HIV seropositivity with a history of frequent previous injections: "The greater number of injections previously received by seropositive children of seronegative mothers than by seropositive children of seropositive mothers (who presumably have similar HIV burdens) strengthens the argument that these injections represent an important route of exposure to HIV, rather than reflecting medical needs for HIV-associated illness."[32] The authors noted that all children and infants in Kinshasa are commonly given intramuscular injections and that there was a high expectation among mothers that injections are an essential part of any cure. "Injections are often administered in dispensaries which reuse needles and syringes yet may not adequately sterilize injection equipment. Furthermore injections are frequently given by untrained personnel or traditional healers." The seropositive infants and children in these studies also had a history of transfusions, which are commonly given to children who suffer from anemia due to malaria.

Children and infants are not the only segments of the African population who are at risk of transmission through injections and transfusions. The view that injections represent the most effective form of medical therapy is clearly widespread in Africa. And adults are just as vulnerable to this means of transmission as children. Of 500 cases studied by Quinn and his colleagues in Kinshasa, 80 percent had a history of prior injections. This was in fact the greatest risk associated with


AIDS in the study, though the authors noted that it was impossible to assess the significance of risk activities without information on control populations. Similarly, in a study of over 2,000 hospital workers, the only identifiable risk factors were hospitalization during the previous ten years, transfusion within the last ten years, and medical injection within the last three years.[33]

In this context it is worth noting that STD clinic attendees, who are reported to have high rates of HIV infection, also have higher than normal histories of injections. The number of such injections may in fact be remarkably high. It has been calculated that among prostitutes in Nairobi the mean time to reinfection after treatment was only twelve days.[34] Prostitutes and their customers may thus be repeatedly exposed to needles.

Melbye and his colleagues rejected the hypothesis that needles accounted for the high rates of HIV infection among the STD clinic attendees they studied in Lusaka, arguing that in the University Teaching Hospital's STD clinic, where they carried out their studies, "clinic needles are not commonly reused" (emphasis added).[35] In addition, they noted that the prevalence figures were as high for those STD clinic attendees who received injections as for those who did not. These arguments are not very convincing, in part because the researchers made no attempt to elicit information about treatment that these patients may have received from other clinics as well as from various indigenous healers, who may have been somewhat less careful about needle reuse. One might also imagine that clinic staff would have been more careful about sterilization in the presence of the AIDS researchers.

The association of hepatitis B infection with STD attendance also points toward a connection between injections and HIV transmission. Thus, Van de Perre and his associates, in a study of prostitutes drawn from an STD clinic in Butare, Rwanda, found that twenty-nine of thirty-three prostitutes were seropositive for HIV, as compared to four of thirty-three female controls; thirty-one of the prostitutes also had hepatitis B virus markers, as compared to only eighteen of the controls; and thirty-one of the subjects and thirteen of the controls tested positive for Chlamydia trachoma . While high prevalences of hepatitis B have also been reported among homosexual populations in the United States, its transmission among heterosexual populations has been more frequently associated with unsterilized needle use. This finding suggests that the subject population in this study may have been exposed to unsterilized needles more frequently than the controls and that this exposure may have accounted


for their higher rates of HIV infection. Therefore, it is impossible to state with any certainty whether frequent sexual activity results directly in HIV transmission or only puts one at risk of exposure to other STDs and thereby at risk of being infected with HIV through exposure to unsterilized needles. The authors of the study, in fact, concluded: "This study suggests that HTLV-III has to be considered as an infectious agent transmitted among promiscuous Central African heterosexuals by sexual contact and/or parenteral contact with unsterile needles used for STD treatments " (emphasis added).[36]

The possible role of needle transmission among adults in STD clinic settings provides an alternative explanation for the commonly reported correlation between frequent prostitute contacts and HIV infection among African men, as well as the association between genital ulcers and HIV seropositivity. The studies that have shown a correlation between prostitute contact, genital ulcers, and HIV seropositivity in Africa have drawn their subject population from STD clinics and their control populations from among blood donors or medical staff populations.[37] As noted above, these studies have not controlled for the possibility that the chain of causation may involve contact with prostitutes, leading to STDs and genital ulcers, leading to clinic treatment and HIV infection through exposure to unclean needles. Prostitutes may be a major risk group for AIDS, but they need not be the primary vectors for HIV transmission that they are frequently made out to be. They may simply be a source for the transmission of STDs, which are a risk factor for HIV transmission through unsterile needle use or genital ulcers.[38] Clearly, studies need to be done which control for the possibility that needles may be a primary route of transmission.

If needle use is a significant means of HIV transmission, we need to examine why sterilization does not occur. Clearly, one of the reasons is that injections are performed by indigenous therapists who are unaware of the risks involved in reusing needles without sterilization. Yet lack of sterilization also occurs in government-run clinics. We should not conclude that the failure to sterilize needles is necessarily a product of laziness or ignorance on the part of African medical personnel, lest we construct another cultural stereotype to explain the spread of AIDS in Africa. Given limited medical budgets (averaging five dollars per capita) and shortages in foreign exchange in most African countries (a product of the declining value of African exports on world commodity markets and currency devaluations instigated by the International Monetary Fund), disposable needles, sterilizers, and even the chemicals needed for


sterilization are often in short supply. The energy costs involved in running sterilizers, even where they exist, may also limit the sterilization of needles.

Warfare is also contributing to the reuse of needles in many countries, such as Mozambique, Uganda, and Angola, by disrupting electrical power supplies and urban and rural health services. In Mozambique, for example, Renamo "terrorists" frequently destroy power lines to the city of Beira. Until the recent acquisition of generators by the city, these actions disrupted needle sterilization within the city hospital. The destruction of rural clinics and medical equipment also increases pressure to reuse needles.

In short, as the World Health Organization has advocated, we need to examine the political economy of health care, and not simply the incidence of improper needle use.[39] In a similar vein, the role of transfusions in transmitting HIV needs to be viewed within the context of a lack of financial resources to properly screen donated blood and an extremely high incidence of trauma injuries requiring blood transfusions associated with warfare and automobile accidents.

Age Distribution Studies: A Critique

Data on the age distribution of both AIDS cases and seropositivity have been used to support the argument that injections do not contribute materially to the transmission of HIV in Africa. Evidence collected on the distribution of 500 AIDS cases in Kinshasa is presented in figures 6 and 7.

Commenting on this distribution, Quinn and his associates conclude, "The sex and age distribution of 500 AIDS cases reflect patterns seen in other sexually transmitted diseases both in developed and developing countries in which incidence and morbidity rates are higher among younger women."[40] On the last point, Peter Piot recently suggested at an AIDS conference in Washington that older men may be having sex with younger women.

These statements imply that the age-sex distribution of AIDS cases supports the assumption that AIDS is being transmitted through heterosexual contact. Yet if one looks at the age-sex distribution of tuberculosis, one sees a similar pattern. Surely no one would argue that TB is a sexually transmitted disease. As is now recognized, women in their childbearing years have a higher risk of contracting TB because of a generalized lowered resistance.[41] Conversely, children in the age range


6. HIV Seropositive Rates, Kinshasa, 1984–1985
SOURCE : T. Quinn et al., "AIDS in Africa: An Epidemiological Paradigm,"
Science 234 (1984): 955–63.

from four to fourteen, for reasons that are not altogether clear, have a higher level of resistance to a number of diseases, including TB. This phenomenon may in fact occur in AIDS. Thus, several studies recently reported in the New York Times noted that children in their teens infected with HIV show fewer signs of a declining immune system than similarly infected adults. Dr. James Goedert of the National Cancer Institute followed up eighty-nine patients from a hemophilia center, all of whom were infected with HIV. After seven years 35 percent of the adults had symptoms of AIDS, whereas only 10 percent of those infected as children and teenagers had AIDS symptoms. Two other studies reported similar findings. Dr. Goedert concluded that "AIDS could resemble chicken pox, measles or other diseases that are more severe in adults than in children and adolescents."[42] If this is the case, then the low incidence of AIDS in the four-to-fourteen age range may reflect resistance to viral replication and AIDS rather than an absence of infection. In other words, the age distribution of AIDS cases in Africa need not support the hypothesis that AIDS is transmitted primarily through heterosexual contacts. Why, then, did Quinn and his associates choose


7. Distribution of 500 AIDS Cases, Kinshasa, August 1985—December 1985
SOURCE : T. Quinn et al., "AIDS in Africa: An Epidemiological Paradigm," Science 234 (1984): 955–63.

"other sexually transmitted diseases," rather than other infectious diseases, as their comparison?

The possibility that young children are resistant to HIV progression raises an additional issue. That is, in many parts of Africa, all children under five are vaccinated at the same time—often with unsterilized needles that are reused. Since this practice is widely recognized as a potential avenue for the transmission of HIV, one would expect to find more AIDS cases in young children. But few such cases have been found. This lack of cases may, as suggested above, be related to the children's resistance to disease progression rather than to an absence of infection. As a result, the cases acquired through this route would show up in the young-adult population.

Medical researchers will immediately object to this explanation on the ground that seropositivity data appear to reveal the same pattern of age and sex distribution. In other words, it is not just that fewer cases appear among children in the four-to-fourteen age range; they also appear to be infected less. This finding appears to support the heterosexual contact model and to call into question the importance of other


forms of transmission, and particularly injections. Thus, Robert Biggar in a recent review of epidemiological data on AIDS concludes: "The results of serosurveys using accurate tests have shown that HIV infection in central Africa is largely heterosexually transmitted. Evidence supporting this includes … the great concentration of HIV-infected persons in the years of greatest sexual activity (15–65 years old) with the peak age-specific prevalence being during the years of peak sexual activity (25–29 years for women; 30–34 years old for men)."[43]

How reasonable is this interpretation? First of all, it must be noted that there is little reliable cross-sectional data on the age-sex distribution of HIV seropositivity or AIDS cases in Africa. In fact, until recently statements about the age distribution of seropositivity were based on a small number of prevalence studies conducted in hospital or clinic settings involving the use of what have often been questionable sampling methods. The first study to present age-specific seroprevalence data using reliable screening methods, and the one that is cited by Biggar and others, was conducted by Melbye and his associates at the University Teaching Hospital in Lusaka, Zambia, in August 1985.[44] The 1,078 subjects in the study included inpatients and outpatients, blood donors, and medical staff. The study found that high levels of HIV seropositivity occurred in women and men between twenty and sixty years of age, with none of subjects under fifteen or over seventy being seropositive. The investigators concluded: "Our findings strongly suggest that heterosexual transmission is an important route of HIV infection in Africa. Seropositivity was restricted to subjects who were in their sexually most active years of life."[45]

Yet even a cursory examination of the data used in this study quickly reveals a significant problem in the sampling method employed. While the total sample consisted of 1,078 subjects, the "under fifteen years" subgroup included only 12 subjects, all of whom were surgery patients between ten and fifteen years of age. The "over sixty" group consisted of 47 subjects, and only 21 subjects were over seventy years of age. In effect, 97 percent of the sample fell between fifteen and seventy years of age; 94 percent, between fifteen and sixty. Similarly, if one looks just at the blood donors and hospital staff, 94 percent of the sample fell between fifteen and fifty years of age. The small sample size for the populations on either side of this age group, and particularly in the "under fifteen" age group, makes any conclusions about the distribution of seropositivity among the wider population from which this sample was drawn meaningless—especially when one realizes that more than 50


percent of the populations of most African countries are under fifteen years of age! In other words, the sample was not representative of the population from which it was drawn.

There is the additional problem created by the use of a hospital population who may have a range of risk factors, including injection and transfusion histories. Without knowing how these risk factors are distributed through the different age groups, one cannot know whether age or the frequency of particular risk factors is being reflected in the age distribution pattern.

It is unclear whether the same problems apply to later sources of prevalence data that are said to support the sexual transmission paradigm. A study of over 5,000 "healthy persons" living in Kinshasa who were tested between 1984 and 1985 revealed a similar distribution of HIV infection. The results were reported by Quinn and his associates in 1986.[46] The report, however, provided no details on how the sample was collected and made no reference to earlier reports that describe the methods involved. A more recent study, conducted in the Central African Republic and employing random sampling, reported an absence of antibodies to HIV in children four to fifteen years of age. The investigators warned, however, that more surveys were needed to confirm this finding because the sample for this age group was nonrepresentative.[47]

Even assuming that the distribution of HIV infection in the sample population in these studies accurately reflects its distribution in the underlying population, what does this finding indicate? Or, put another way, could the high prevalence of HIV among sexually active subjects and the relatively low prevalence among children reflect the presence of factors other than levels of sexual activity? For example, if the transmission of HIV is facilitated by the presence of preexisting immunosuppresant infections, and if children between the ages of four and fifteen are more resistant to such infections than either infants or adults, the four-to-fifteen age group would have a lower prevalence of HIV than other age groups regardless of how the infection is transmitted. Alternatively, their increased resistance to other infectious diseases may reduce their exposure to risk of infection through needles and/or transfusions. In short, age distribution data, even if correct, need not indicate that sex is the only avenue of HIV transmission in Africa.

These possibilities are admittedly speculative. Moreover, they run counter to the often-quoted instruction given to all medical students: "When you hear hoofbeats, think of horses, not zebras." AIDS, however, is a relatively new disease, about which there is a great deal more


to be learned. Thinking of horses, while reasonable in a diagnostic setting, may prevent medical researchers in Africa from seeing the entire range of factors at play in the epidemiology of AIDS in Africa. In any case, in Africa one can just as easily come across a zebra as a horse.


The point in all this is that assumptions about the importance of sexual promiscuity in the transmission of HIV in Africa were initially based on limited, and in some cases methodologically questionable, data. These assumptions, nonetheless, shaped both the questions that AIDS researchers asked and the way that they interpreted data. This narrowing of research, in turn, discouraged serious consideration of the role of alternative avenues of transmission, such as injections, or of the role of possible cofactors, such as high background levels of infection and malnutrition and associated problems of poverty and maldevelopment, which may be as important in the heterosexual transmission of HIV as the frequency of sexual contacts.

We are, in fact, much further from understanding the epidemiology of AIDS in Africa than some medical researchers, development officers, and social scientists would have us believe. It is clear that heterosexual transmission of HIV occurs in Africa, but how or why it occurs has not been demonstrated. AIDS, like Burkett's lymphoma, may yet prove to have a complex etiology involving a combination of political and economic forces associated with underdevelopment in Africa. These forces have brought together particularly susceptible populations—populations subject to high background levels of viral, parasitic, and bacterial infections—in a social setting marked by high levels of familial separation and multiple sexual partners. These conditions, in turn, have contributed to the spread of other STDs, which in turn have created a high risk of exposure to HIV via genital ulcers and/or infected needles.

The medical research community and the social science community working with it must develop research agendas that will illuminate these complex interactions, instead of obscuring them through a precipitous move to find quick answers that can be easily translated into AIDS containment programs. The early history of Western research on TB and syphilis should serve as a warning that, if we continue to look for easy solutions, those solutions may have a limited impact and, at the same time, our understanding of the wider epidemiology of AIDS may be


diminished. Before we spend millions on the type of behavior modification model of intervention now being developed, we must have a higher degree of certainty about how HIV is being transmitted and what the real risk factors are, as well as knowledge of the social and economic context within which risk behaviors are set. If primary risk factors are poverty and unemployment, our proposed interventions must address the causes of these conditions. We must not allow AIDS to become one more symptom for which the West finds a cure without addressing the underlying causes of this and many other health problems. At the very least, studies that explore other cofactors and avenues of transmission need to parallel efforts to understand and prevent the sexual transmission of HIV.

Any attempt to initiate a more comprehensive approach to AIDS research in Africa requires that medical researchers and social scientists develop a more productive working relationship. Such a relationship will necessitate adjustments on both sides. Social scientists need to resist attempts to limit their input to collecting and presenting cultural artifacts; they need to be critical of colleagues who continue to accept this limited role. If the conferences and workshops at which social scientists and medical researchers are brought together are to be productive, their agendas must be expanded to address broader issues of African social and economic life and not simply the "sexual life of the natives" and other forms of risk behaviors. At the same time, any attempt to open up the agenda must be sensitive to the medical research community's need for questions and hypotheses that take account of the existing epidemiological data and that can be empirically tested.

For their part, medical researchers need to take a more open view of how the social science community can contribute to epidemiological research on AIDS in Africa. They need to recognize that AIDS research in Africa can be illuminated by a more fundamental knowledge and understanding of the contours of African social and economic life, which involves more than a cataloging of risk behaviors. Medical researchers, like social scientists, also need to be more openly critical of their own colleagues. Research based on unrepresentative samples or faulty sampling methods need to be challenged. Peer review procedures need to be uniform, so that conclusions presented on AIDS in Lusaka are viewed as critically as conclusions made about AIDS in New York City. The medical community also needs to be more forthright and open about the limits of our present knowledge about the epidemiology of AIDS in Africa. Too many studies have taken an authoritatitve tone that is not


warranted by the data available and, in doing so, have encouraged premature closure of African AIDS research.

previous chapter
Medical Research on AIDS in Africa: A Historical Perspective
next chapter