|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
LABOR MIGRATION, GOLD MININ\G, AND LOW HIV PREVALENCE INT GUINEA
ADAM DANIEL KIS
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
Adam Daniel Kill
To Kristi Kill, my wife, supporter, biggest fan, and best friend
I thank the chair of my advisory committee, H. Russell Bernard, as well as the rest of the
committee Marianne Schmink, Willie Baber, and Leonardo A. Villal6n for their guidance
and wisdom. I also thank my wife, Kristi KiS, and son, Zachary KiS, for their support during the
entire graduate school enterprise. I thank my parents, Miroslay and Brenda KiS, and my parents-
in-law, Gorden and Cheryl Doss, for their support and encouragement at each stage of the
journey. I thank the ADRA Guinea Country Directors, Sharon Pittman and Patrick Millimono,
for their logistical support to help make this research possible. And I thank all seven of my
research assistants Fodedj an Kei'ta, Koffl Edem Dzotsi, Sekou Thidiane Bangoura, Mohamed
Camara, Elisabeth Haba, Oumou Sidibe, and Saran Kei'ta for their translation services,
questionnaire development, and assistance with all aspects of data collection. And finally, I
thank God for helping me to successfully complete my graduate studies.
TABLE OF CONTENTS
ACKNOWLEDGMENT S .............. ...............4.....
LI ST OF T ABLE S ............ ...... ._ ._ ...............7....
LI ST OF FIGURE S .............. ...............8.....
AB S TRAC T ..... ._ ................. ............_........9
1 INTRODUCTION ................. ...............10.......... ......
The Research Problem ................. ...............10........... ....
Guinea' sHistorical Context ................. ............. .................. .................14
The Dissertation Outline ................. ...............17........... ....
2 ANTHROPOLOGICAL PERSPECTIVES ON HIV/AIDS .............. .....................2
Causes/Transmission .............. ...............20....
Effects ................. ...............25.......... .....
Risk Groups .............. ...............28....
3 AIDS PREVENTION: WHAT HAS WORKED, AND WHAT HASN'T ................... .........33
Individual HIV Prevention Approaches .............. ...............33....
Multi-Faceted HIV Prevention Approaches ................. ...............42................
Conclusion ................ ...............55.................
4 THE EDUCATIONAL MODEL OF SOCIAL CHANGE .............. ..... ............... 5
Theoretical Considerations ................. ........... .. ..... ..... ..... ..........5
The Failure of the Educational Model of Social Change ................. ......... ................61
Successful Education: A Multi-Faceted Paradigm ................ ...............66........... ...
5 THE ANTHROPOLOGY OF DISEASE IN WEST AFRICA .............. .....................7
Background ................. ...............80.................
Cultural Taboos .............. ...............8 1...
Behavioral Prevention Factors ................. ............ .. ...............85.....
The Role of Tangential Practices in Disease Prevention............... ...............9
Conclusion ................ ...............91.................
6 RESULTS AND ANALYSIS OF DATA COLLECTED IN THE BOURE .........................94
Condom Availability .............. ...............96....
Condom Knowledge ................ ...............104................
Condom Enforceability ................. ...............113................
Actual Condom Usage ................. ........... ...............118 .....
Foudoukoudouni (Short-Term Marriage) ............ ......__ .....__ ...........12
7 DISCUS SION AND CONCLUSIONS ............ .....__ ...............155.
C or ABC? ............ ...............156.....
Beyond ABC............... ... ............ ................15
The Educational Model of Social Change in the Boure .............. ..... ............... 16
Primary Factors in the Boure's Success .............. ...............165....
Conclusion ........._._ ...... ___ ...............169.....
LIST OF REFERENCES ........._... ...... ..... ...............172...
BIOGRAPHICAL SKETCH ........._... ...... .___ ...............187....
LIST OF TABLES
6-1 Influence of education/knowledge upon practice of ABC behaviors .............. .............150
6-2 Marital status and practice of ABC behaviors ..........._ ..... .__ ......__.........5
6-3 Sex and practice of ABC behaviors ..........._.....___ ...............150
6-4 Place of origin and practice of ABC behaviors ..........._ ..... .__ ........__........150
6-5 Influence of predictors upon condom use at last sex ................ ......... ................15 1
LIST OF FIGURES
1-1 Adult HIV Rates in Africa for 2003 .............. ...............18....
1-2 HIV Prevalence in Guinea by Region ................. ...............19......_._. .
6-1 Average condom count ...._.._ ................ ...............146 ....
6-2 Breakdown of correct responses ...._.._ ................ ........._.._ ....... 14
6-3 How to use a condom properly ........._.._.. .......... ...............147..
6-4 Number of correct responses .............. .....................148
6-5 Distribution of correct responses ........._.._.._ ...._.._.......... ...........4
6-6 Women's empowerment sorted by gender .............. ...............149....
6-7 Self-reported AIDS avoidance methods .............. ...............149....
6-8 Decision modeling tree for condom use .............. ...............152....
6-9 Decision modeling tree for condom use .............. ...............153....
6-10 Is foudoukoudouni faithful? ............. ...............154....
6-11 Revised faithfulness chart ................. ...............154...............
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
LABOR MIGRATION, GOLD MININ\G, AND LOW HIV PREVALENCE INT GUINEA
Adam Daniel Kill
Chair: H. Russell Bernard
Guinea, West Africa has an estimated HIV prevalence of between 1.5 and 3.2%. This
figure is among the lowest in the region and on the continent. Though Guinea has low HIV
rates, it also has many of the same traits that correlate with high HIV infection rates elsewhere.
For example, typically when male labor migrants are away from home for long periods of time,
they engage in unprotected, multi-partner sex with prostitutes, rapidly facilitating the
transmission of HIV. In Guinea, mining accounts for approximately 70% of exports, and the
mines attract thousands of labor migrants from across the country and from neighboring
countries. Yet these mining regions have the lowest HIV rates in the country. My study offers
cultural and behavioral explanations for this phenomenon. My data shows that faithfulness to
one's sexual partner or partners was the most widely-practiced HIV prevention strategy, and
short-term marriage is one cultural expression of this sexual partner reduction.
The Research Problem
Much research has gone into understanding the cultural causes and effects of AIDS in
places that are hardest-hit by the pandemic worldwide. Anthropological and epidemiological
theories about HIV transmission have sprung out of studies of those nations with the highest
infection and death rates in the world. Billions of dollars a year are poured into academic
research and development proj ects aimed at alleviating the effects of AIDS on individuals,
families, and employers in high prevalence nations.
In contrast, aside from Thailand, Uganda, and Senegal, comparatively little inquiry has
delved into understanding low HIV prevalence countries. Nations with low HIV rates may hold
important theoretical clues that can help explain and reduce HIV transmission in high prevalence
places. The Republic of Guinea, West Africa, a nation of more than 9.5 million people (Central
Intelligence Agency 2006), is one such low HIV prevalence place.
The highest estimates place Guinea' s HIV prevalence at 3.2% (Joint United Nations
Programme on HIV/AIDS 2004; see Figure 1-1), while a recent Demographic and Health Survey
co-conducted by Macro International (Direction Nationale de la Statistique 2005a) concurs with
more conservative estimates (Joint United Nations Programme on HIV/AIDS 2001) placing HIV
prevalence at 1.5%; this suggests either a drop in infection rates, over- or underestimation of the
prevalence of the disease, or statistical variation. In any case, these figures are some of the
lowest in the region and on the continent. In neighboring C8te d'Ivoire, for example, the HIV
rate is 7.0%, while in Botswana it is 37.3% and in Swaziland it is 38.8% (Joint United Nations
Programme on HIV/AIDS 2004).
Some 25 years into an epidemic that has ravaged the continent of Africa, it is remarkable
to find a sub-Saharan African nation that has still not succumbed to the disease in any significant
way. Perhaps because it seems initially counterintuitive to some that a sub-Saharan African
nation can truly have such low AIDS rates, there are those who quibble with these figures. Most
often cited as counterevidence are personal reports from health workers in the field who believe
rates to be much higher (though personal opinion cannot trump scientific results substantiated
through rigorous research) and obj sections relating to the reliability of sentinel surveys of
maternity clinics (though such obj sections are not often used to invalidate reported rates of high
HIV prevalence countries). The flaw in these obj sections, however, is that though sentinel
surveys may be an imperfect method of determining national HIV seroprevalence, nothing
suggests that the method would be any more imperfect in Guinea than elsewhere. Whatever
error is introduced into a reported national prevalence rate through methodological insufficiency
is no more likely to skew Guinea's figures any more than other sub-Saharan African countries
whose rates were determined using exactly the same methods. Therefore, regardless of any
inaccuracies in Guinea's reported absolute rate, the figures indicate lower relative HIV
prevalence than most other sub-Saharan African nations.
In addition, the recent Demographic and Health Survey (which reported a 1.5% prevalence
rate) was among the first to move away from the maternity clinic sentinel survey method, instead
drawing blood directly from a national sample of both men and women. The report describes
their methodology as follows (author' s translation from the French):
With the inclusion of an HIV test in the EDSG-III survey, Guinea is one of the first sub-
Saharan African countries to enrich its HIV database with data from a national survey,
representative of the general population. For the first time, the country has a direct
measure of HIV prevalence instead of an estimation derived from a model based on data
from sentinel sites and adjusted by certain parameters. UNAIDS and WHO recommend
that the results from a nationally representative survey be used to calibrate the results of
regular surveillance. [Direction Nationale de la Statistique 2005a:36]
Others have attempted to discredit the significance of Guinea' s low AIDS rates with the
tautological argument that the rates are so low because AIDS hasn't yet arrived in Guinea full
force. Yet that does not explain why AIDS hasn't yet arrived in full force, especially considering
the higher rates of virtually all of Guinea' s adj acent neighbors. That is precisely the
phenomenon of interest. Despite a constant influx of immigrants, refugees, and temporary
migrant workers from its higher-prevalence neighbors, national AIDS rates in Guinea have
remained low for years.
Guinea is particularly intriguing because, although it has low HIV rates, it also possesses
some of the very traits that correlate with high rates of HIV infection elsewhere. For example, it
is argued that high poverty is associated with high rates of HIV transmission because poverty
produces prostitution, inadequate access to risk-reduction information, and poor access to public
health services (Seeley et al. 1994; Hulton et al. 2000; Wyohannes 1996; Krueger et al. 1990).
Guinea' s annual per capital GDP of $1,960 puts it at 157th of 175 ranked nations of the world
(United Nations Development Programme 2003), yet it lacks the predicted accompanying high
rate of HIV. It has also been observed that early sexual debut (associated with increased
numbers of sexual partners over a lifetime) increases the probability of being infected with HIV,
particularly where early debut involves very young women having unprotected sex with older
men (O'Donnell et al. 2001). Guinean youth, however, make their sexual debut at an average
age of 16.3 for girls and 15.6 for boys that is, at approximately the same age as their
counterparts elsewhere in high HIV prevalence Africa (Goirgen et al. 1998), but there is no parity
of HIV prevalence rates between Guinea and most other parts of Africa.
Particularly perplexing is the fact that Guinea, like many high HIV prevalence countries in
sub-Saharan Africa, has a high rate of male labor migration. Longstanding theory predicts that
when men are away from home for a long time, they are likely to engage in unprotected, multi-
partner sex, particularly with prostitutes, which greatly increases both the likelihood of their
becoming infected with HIV and of their transmitting the virus to others (Hunt 1989). In Guinea,
mining (particularly bauxite, diamonds, and gold) accounted for over 70% of exports in 2004
(Central Intelligence Agency 2006). These mines attract thousands of labor migrants from
across the country and from neighboring countries with higher HIV rates. Yet according to a
USAID study, out of the five regions of Guinea, the two regions containing the largest gold,
diamond, and bauxite deposits in the country (Haute Guinee [gold and diamonds] and Basse
C8te [bauxite]) have the lowest HIV rates (2.1% and 2.7% respectively; Fouta Djallon has 3.9%,
Conakry has 5.0%, and Guinee Forestiere has 7.0%) (Lartigue 2001; see Figure 1-2). In other
words, the mining sites for the three principal minerals in the country which account for some
70% of exports and attract hundreds of thousands of workers from higher HIV prevalence
regions and neighboring countries are located in the regions with the lowest HIV in the country.
In fact, of all the high-risk groups identified in the country, miners in Guinea have the
lowest HIV prevalence at 4.7%. This is 68% higher than the 2.8% prevalence in the general
population (as reported in the same study), but it is significantly lower than that of commercial
sex workers (42%), tuberculosis patients (16.7%), truckers (7.3%), and members of the military
(6.6%) (Lartigue 2001; Direction Nationale de la Statistique 2005b).
What might be going on in Guinea' s mining communities that has a protective effect
against HIV transmission? Understanding how this dominant industry has managed to stave off
widespread infection thus far can fill in a piece of the puzzle that explains how Guinea as a
whole is avoiding the pandemic. It can also contribute important insights into our broader
understanding of HIV transmission in Africa in general.
Guinea's Historical Context
Prior to colonization, the Malinke people had established dominance over Upper Guinea,
and soon thereafter (by the 14th century), the whole of Guinea' s territory was incorporated into
the Empire of Mali. A century later, Portuguese sailors were among the first Europeans to reach
Guinea' s coast, and by 1891, Guinea had become a French colony. There were several
resistance movements throughout the French colonization period, the most famous of which was
led by Samory Toure. By the mid-1950s, a descendent of Samory, Sekou Toure, was again at
the forefront of an independence movement.
With restive independence movements gaining steam across the continent by 1958, the
French tried to stay a step ahead of the inevitable by offering each of their West African colonies
a choice between increased autonomy within the French system or immediate, complete
independence. Guinea was the only country choosing immediate, complete independence, with
Sekou Toure declaring that "we prefer freedom in poverty to prosperity in chains." What
happened next is somewhat of a mystery. Conflicting reports indicate that the French were so
incensed by this choice that they destroyed infrastructure that they couldn't take with them,
ripping out telephone and electric lines, destroying buildings, and damaging roads and bridges as
they pulled out. Such large-scale destruction does not appear to be a hallmark of subsequent
French pullouts from other African colonies, either indicating that the reports from Guinea were
inflated or that the French realized later on that such a method of pullout was not good policy. In
SInformation contained in this section was gleaned from Lonely Planet 2002, Central Intelligence Agency 2006, US
Department of State 2006, R~my 1999, and personal conversations with Guineans in the field.
any case, what is certain is that as the French colonial administration and private citizens fled in
haste, they took their immense amounts of capital with them.
This sudden large-scale pullout created a vacuum into which Sekou Toure stepped as the
country's first president. The ascension of a Guinean to the country's leadership, however,
didn't alleviate the need for outside assistance. The French pullout had been so abrupt that
Guinea's economy risked collapse if there was to be no transitional assistance. Having alienated
the West with its abrupt rej section of France' s offer of autonomy, Guinea looked for assistance
behind the Iron Curtain. But when the Soviet Union proved lukewarm in its support, Sekou
Toured instead modeled his country's economy on the Chinese Maoist system of collectivized
agriculture. Having started out on a bad foot economically, this "cultural revolution" dragged
the country down even further, leaving Guinea so devastated that it still reels from the
consequences today, despite attempts at reform and subsequent outside assistance from the West.
When Sekou Toure died of heart failure in 1984, General Lansana Conte succeeded him in
office, where he remains to this date.
To this day, the Malinke are one of the largest ethnic groups in Guinea, and the Malinke
language is a lingua franca of sorts throughout the country. Malinke are influential in all levels
of society, producing one of francophone Africa' s best known authors (Camara Laye), several of
Guinea' s key political leaders, both past and present, and hosting one of the best universities in
Guinea in Malinke territory (the University of Kankan). Today, the Guinean Malinke heartland
is centered in Upper Guinea, and Siguiri (the site for this research proj ect) is one of this region' s
Because of its particular history, Guinea is less developed than many of its next-door
neighbors in the region, having ranked at or near the bottom of the UN' s quality-of-life index for
the past 15 years or so. Like many sub-Saharan African nations, Guinea is also dealing with the
arrival of AIDS. Its long-term isolation from the outside world, however, may also have
shielded it from developing AIDS as endemically as the rest of Africa. African nations with the
most outside contact with each other and the rest of the world (particularly the West) have the
highest rates of HIV on the continent. Correlation does not necessarily imply causation, but it is
impossible nonetheless to ignore the relationship. Aside from HIV, Guinea's contemporary
challenges include gross underdevelopment, accommodating over half a million refugees from
neighboring Liberia and Sierra Leone, and steering clear of involvement in cross-border strife
stemming from those nations' recent civil wars.
The History of Gold Mining in the BourC. Gold has been mined in the Boure since at
least the time of the Empire of Mali in the 14th century (Boure 1999). The Empire of Mali was
known for its wealth of gold, which was passed on to successive empires that ruled the same
region (Clarke 1964). Industrial gold mining in the modern era centers on Siguiri, with the
largest operation in the country run by SAG2 Since 1985 (Direction Nationale de la Statistique
2005b). Information on the history of traditional mining in Guinea is extremely scarce, but it is
reported that children from the age of 12 onward sometimes accompany their parents to the
mines, providing labor with no recompense save for experiential knowledge enabling them to
become miners themselves. In this way, there is continuity with the past and children continue
the family history of involvement in gold mining (Direction Nationale de la Statistique 2005b).
2 There does not seem to be consensus on what SAG stands for. It has been alternately referred to as Soci~tC
Ashanti de Guinde, Soci~tC Ashanti Goldfields, and Soci~tC Aurifibre de Guinde. It appears, however, that SAG is
owned by AngloGold Ashanti, a merger of South Mfrica's AngloGold and Ghana's Ashanti Goldfields.
The Dissertation Outline
The rest of this document outlines HIV prevention strategies being used in Guinea' s
mining communities, and situates this information in a historical context. Data for this research
proj ect were collected between October 2005 and April 2006 near Siguiri, Haute Guinee (Upper
Guinea), Guinea. Seven rigorously selected and trained field research assistants (four male, three
female) helped with various portions of data collection and translation between French and
Malinke. During the research I was a part-time volunteer with ADRA Guinea (Adventist
Development and Relief Agency). This gave me immediate entry to the area and access to
resources such as office support. The study, however, was conducted independently of ADRA:
the organization had no part in designing or supervising the research.
The next four chapters address the historical and theoretical background of this study.
Chapter 2 discusses the history of anthropological investigation into HIV/AIDS through mid-
2006. Chapter 3 examines the sometimes acrimonious debate between adherents to radically
different paradigms about how best to Eight the epidemic. Chapter 4 investigates that cherished
Western ideal, the educational model of social change, and explains why it has limited impact on
behaviors rooted below the superstructural level. Chapter 5 traces the history of indigenous
disease prevention in West Africa, particularly in the context of non-Western models of disease
Chapter 6 presents the data from this study and analyzes them in reference to specific
hypotheses and cultural narratives. Chapter 7 concludes with a discussion of the research
Endings, how they tie in to the historical and theoretical background presented in earlier
chapters, and what inferences can be drawn from the information presented.
(%l pota Il*4 Se 10@]
03 3" .9
kwa.~~U r:kasR.wwrybr MC~hs
Fgr1-.Adult HIV Rate inArc fr20 (ore oitUiedNtosrgam on
[ rHIV/ Al I~DS 2004)~
1-EY prevdence rate amnag uarba-base prospeant womnw,
by nature reganm irCuinea
Figure 1-2. HIV Prevalence in Guinea by Region (Source: Lartigue 2001)
ANTHROPOLOGICAL PERSPECTIVES ON HIV/AIDS
Almost as soon as AIDS was identified, researchers from nearly every discipline rushed to
describe this new disease. Efforts to understand its causes, manifestations, and transmission
modes revealed more questions about its origins, effects, and preventability. Now, more than 25
years since its emergence, AIDS is still one of the hottest topics in biomedical and sociocultural
research, both because a cure still evades the medical research community, and because there is
sharp disagreement within the behavioral scientist community about the best way to prevent and
control the epidemic.
This chapter traces the history of HIV/AIDS research from an anthropological perspective,
focusing on literature that deals with the social aspects of the disease, rather than its biomedical
aspects. The literature on AIDS is addressed from three different angles: causes, effects, and risk
groups. The section on "causes" addresses cultural factors that assist in the transmission of
AIDS; "effects" looks at cultural impacts of the disease; and "risk groups" analyzes what
constitutes a risk group (and the validity of defining such groups).
This literature review is not exhaustive. A search for "HIV" in Journal Storage: The
Scholarly Journal Archive (JSTOR) alone produces 8,856 hits. Limiting the "HIV" keyword
search to anthropological j ournals produces 244 hits (search conducted on www.j stor.org, April
26, 2005). Expanding the search to other journal search engines would add to these Eigures. In
this chapter, I focus on the maj or themes of social science research on HIV/AIDS.
Where did AIDS come from in the first place? There is strong speculation among
biological scientists that it was a crossover virus that leapt from other primates into humans in
Africa. But not all theories of origins are biological or accidental. Some see a sinister, human
hand in the genesis of the disease. There are those who believe AIDS arose from a batch of
contaminated polio vaccines intentionally administered in central Africa in the 1950s, a la
Tuskegee (Marshall 2004).3 Others, such as the outspoken Kenyan Deputy Environment and
Natural Resources Minister and recent Nobel peace laureate, Wangari Maathai, advance the idea
that the AIDS virus was deliberately created by a Western scientist for biological warfare and
mass extermination. It was, she claims, created to exert control over black people, though no one
knows exactly who created it (Agence France Presse 2004).
Among the early attempts to describe the epidemiology of AIDS, Quinn et al. (1986)
recognized that AIDS is caused by a virus that is spread through both social means (sexually;
through childbirth) and through non-social means (blood transfusions; unsterilized needles).
They were also among the first to recognize that, although the biological manifestation of AIDS
is similar in most patients, "the epidemiology and clinical features of the infection in different
countries may vary, depending on cultural differences, endemic diseases, and other unidentified
risk factors" (Quinn et al. 1986:955-956). Homosexual contact and intravenous drug use were
already established modes of transmission of AIDS in the United States at that time; through a
careful accounting of AIDS cases in Africa, it was acknowledged that AIDS can also be
transmitted through heterosexual contact, and that this latter form of transmission is much more
prevalent in Africa than any other.
Over time, these differences in transmission patterns became more widely recognized, and
it was deemed useful to formalize these differences for the sake of targeting interventions more
appropriately. The epidemiological community initially identified two primary transmission
3 In the infamous Tuskegee study of syphilis, patients were not adequately informed of their diagnosis and treatment
options in order that the disease could run its course and observations could be made about each stage of the illness.
This caused permanent damage or death in many study participants, and though the study advanced modern
medicine's knowledge of the disease significantly, it has been cited as an all-time low in research ethics.
patterns, but the list was eventually expanded to four patterns. Each pattern typically
predominates in a specific region of the globe. Pattern I is typified by male-to-male sexual
transmission and transmission by intravenous drug users (this is found in industrialized countries
of the West). In Pattemn II, male/female transmission predominates, and perinatal transmission is
common (this typifies sub-Saharan Africa and the Caribbean). Pattern III, which exhibits few
cases, involves transmission through infected blood supplies (this is found in Eastern Europe,
North Africa, and the Middle East). Finally, Pattern IV primarily involves transmission from
female sex workers (F SW' s) to their clients, with a higher number of infected men than women
(a few FSW' s are core transmitters to a large proportion of male clients; this typifies Asian
transmission patterns) (Green 2003b).
According to Green, the epidemiological community has now somewhat abandoned these
designations, believing that only one pattern would emerge over time (2003b). This
abandonment may have been motivated by ignorance of how to deal with multiple modes of
transmission. At the outset, there was only one known way to approach HIV reduction (based
upon the gay community in the West), and there may have been reluctance within the ranks of
AIDS experts to acknowledge patterns that they did not know how to deal with conclusively.
But time has demonstrated that pattern differences persist, and unique social and cultural
influences on HIV transmission should not be ignored. Acknowledging different causal patterns
opens the door to multiple prevention solutions, which may lead to more successful outcomes
than are currently being seen in the worldwide fight against AIDS.
Nested within the various patterns themselves, specific factors (behavioral and biological)
are known to cause or facilitate HIV transmission. Body fluid exchange is the primary biological
driver, while the presence of STD's facilitates HIV transmission during sexual contact. High-
risk groups (such as prostitutes) can disseminate the disease rapidly, and HIV-positive mothers
can transmit the infection to their children through childbirth and breastfeeding. Ritual
scarifieation, tattooing, circumcising, and piercing also have the potential to transmit HIV
(Quinn et al. 1986).
Other researchers have noted additional specific factors that cause or facilitate the spread
of HIV. For example, Parker (1987) described how Brazilian cultural understandings of
sexuality permit both male-male and male-female anal intercourse to be practiced by
heterosexuals without necessitating a change in sexual identity. This cultural trait permits HIV
to jump easily between the homosexual and heterosexual populations, and it blurs the lines
between the distinct transmission patterns discussed above. This does not necessarily mean that
a new pattern should be developed for every possible variation. But at the very least, it needs to
be acknowledged that exceptions do exist to the recognized transmission patterns, and these
exceptions should trigger awareness of the specific cultural contexts that influence how HIV is
spread. Rather than discrediting the pattern identification system altogether, exceptions refine it
and make the system more versatile to help develop appropriate responses.
Other well-known HIV infection routes include transmission by truckers and migrant
workers who frequent prostitutes (see more about this under the "Risk Groups" section of this
chapter). Hunt (1996) labeled these activities "sexual networking", and they were accelerated in
Africa as a result of the introduction of colonial enterprises that relied upon trucking and migrant
labor. Colonists deliberately left certain areas of their territories underdeveloped, so that
inhabitants would continue in subsistence agriculture and remain as a labor pool from which they
could draw workers for enterprises in other regions (Hunt 1996). In the post-colonial era,
trucking and labor migration remain vitally important to the economies of many African nations,
and sexual networking of the kind described by Hunt remains widespread.
HIV may also be transmitted via dangerous, unethical research practices. Like the
Tuskegee study alluded to earlier, studies of HIV transmission may neglect to treat and inform
research subj ects about the HIV status of their partners so that they can make an informed
decision about whether or not to continue the relationship at the risk of their own infection
While some researchers look at behavioral factors that directly affect HIV transmission,
others prefer to focus on external factors that influence behaviors themselves. One example is
Farmer' s account of AIDS in Haiti (1992). He focuses on structural inequality and poverty as
catalysts for AIDS transmission; those in "power-down" positions or extreme poverty are less
able to resist unwanted sexual advances; they engage in money-generating sexual behaviors; and
they have the least access of anyone to adequate health care. Another example is Poku (2002)
who argues that Africans are no more hypersexualized than many Westerners who also engage in
unprotected multipartner sex, so a behavioral explanation for differential HIV transmission rates
won't do.4 Though Poku may be right (notwithstanding the difficulty one would face in creating
an unbiased index of hypersexuality with which to compare the two groups), similar levels of
hypersexuality would not necessarily indicate similar expressions of hypersexuality. Cultural
and behavioral differences would still generate different patterns of HIV transmission, even if
levels of sexual activity were similar between the West and Africa.
4 MarShall (GI r 14) takes this a step further by saying that racial mischaracterization of Africans' hypersexuality
causes many Westerners to believe that Africans deserve AIDS, fueling Westerners' apathy towards its mitigation in
Though Poku says that sexual behavior can be an important factor, "alone, however, it
appears totally inadequate in explaining HIV prevalence as high as 30% of the adult population
in some African countries and less than 1% anywhere in the Western world" (2002:533).
Instead, Poku fingers poverty as the culprit driving the epidemic. It constrains the treatment of
STD's cofactorss in HIV transmission), drives people to seek work away from home (increasing
multipartnering), and creates even deeper poverty by killing off productive family members and
driving girls and women into prostitution to survive. In addition, Poku argues that structural
adjustment policies which cut governmental health expenditures to bring budgets more in line
with donor expectations constrain the ability of behavior change strategies to be successfully
Rather than just rely on anecdotal evidence, Seeley et al. (1994) tested the hypothesis that
socioeconomic status is associated with HIV risk, using type of dwelling, available land size,
cattle ownership, and an index of household items as measures of wealth. The study, conducted
in southwest Uganda, found that all four indicators revealed that the poorest heads of household
were most likely to be HIV-positive. The authors speculated that the reason for this is partly due
to the nature of income-generating strategies practiced by the poor in order to survive.
Behavioral and structuralist explanations are not mutually exclusive. Poverty drives
people to engage in risky sexual behaviors to survive, but sexual behaviors that spread AIDS
exacerbate poverty. This means that we should not be asking whether to pour prevention money
into behavior modification or into changing the structures of poverty, since both kinds of
intervention are needed.
In addition to sociocultural causes of AIDS, researchers are also interested in the effects of
AIDS on individuals and societies. Bloom and Carliner (1988) note the economic impacts that
AIDS can have, such as affecting market output if a significant number of workers fall ill or die
from the disease. It also affects family economics, as earnings from a sick individual are
foregone, in addition to any household services they might render if healthy. Paul (1994) reports
that in Thailand, tourism and foreign investment are also impacted, which has serious financial
repercussions for the nation.
Fredland (1998) argues that AIDS has consequences for the national psyche of many
African countries. Independence and nationalism are relatively recent phenomena, but as AIDS
cripples African countries, Fredland believes that their national identity (patriotism, national
pride, optimism for progress) will be undermined. Life expectancy is also affected by AIDS. In
countries hardest hit by the epidemic, life expectancies can drop by more than 20 years in a
relatively short time, and the most highly educated elites may be severely affected (Fredland
1998). This can rapidly destabilize a country that depends on those elites for smooth operation
of government and society.
Knodel et al. (2001) discuss the effects of AIDS on the elderly, including strains of caring
for their sick adult children, providing financial support, raising orphaned grandchildren, dealing
with emotional stress, and losing support for themselves that would have been provided, had
their children not died prematurely of AIDS. The effects are particularly great in cultures that
place a high value on intergenerational care giving and familial support.
In earlier work, Kill reported the effects on funeral culture in Malawi as rural people face
rising death rates from AIDS (2007). Attending funerals cuts into productive work time, puts a
strain on finances (as each attendee is expected to bring a gift to the family of the deceased), and
sometimes interferes with other social obligations. In response, Malawians are attending only
the highest priority funerals (defined as those for people they feel indebted to or in whose
families they want to create a sense of social indebtedness) or attending a small part of each
funeral when several are in progress. In addition, increasing property seizure by relatives is
creating a financial burden for AIDS victims, as such seizure used to occur only after death.
Some effects of AIDS loop back into causes in a vicious cycle that continues unchecked
without deliberate interference. An example is the difficulty Thai AIDS orphans face in
acquiring property rights when their parents die. Without property to sustain them, many
children have no alternatives but petty crime and sex work, which place them at high risk for
HIV infection (Paul 1994).
Stigmatization of AIDS victims removes those infected from contexts where they might
contaminate others. But it' s also a large disincentive to be forthcoming about infection, and
keeping things under wraps makes for good chances that AIDS will be passed on unknowingly.
AIDS neutralizes development gains as resources are poured into prevention at the expense
of other needs. The neglect of other development needs worsens the overall situation, and
circumstances become more favorable for HIV transmission as poverty increases (Fredland
AIDS leads men to seek younger and younger sexual partners who are less likely to be
infected (Fredland 1998). In some parts of Africa, according to popular opinion, sex with a
virgin can cure AIDS, and the youngest possible partners are sought (including infants) under the
assumption that they are more likely than anyone else to be virgins (see Vickers 2006; Swindells
2003; Sandars 2006). In a particularly egregious feedback loop of effects and causes, this
predation on children exposes them to the virus.
Many Africans are reluctant to alter their reproductive strategies after being infected with
HIV. Reasons given include wanting to prove that they can still bear healthy children, wanting
to keep up appearances of normalcy to avoid being ostracized, and wanting to stick with a preset
plan for the number of children desired. In Zambia, however, AIDS does impact childbearing
decisions: families sometimes reduce the number of children that they produce in order to care
for relatives' children whose parents have died of AIDS; and HIV-positive women who know
their status are less likely to continue bearing children so that it doesn't "bring out" the disease in
them (Rutenberg et al. 2000). In these cases, seropositivity engenders childbearing limitations
primarily for self-protection rather than to protect unborn children from the disease.
AIDS also impacts Zimbabweans' reproductive decision making. If parents discover that
they are HIV-positive, the desire to produce children as insurance in old age (a common strategy
throughout Africa) is squelched, as the parents are unlikely to ever reach old age. Even without
a medical diagnosis, Zimbabweans take a child's death under five years of age as a potential sign
of seropositivity. Conversely, if their children successfully reach five years of age, it is
understood as a clean bill of health and a license to produce more children. In addition, having
children as early as possible is seen as a way of producing HIV-free children, under the
assumption that the number of sexual partners increases with age, which increases the probability
of contracting the disease (Grieser et al. 2001).
Effects of AIDS that feed back into causes are an important target for interventions.
Providing more leaves of absence for Malawians to attend more funerals helps people cope with
the devastation of AIDS, but does nothing to prevent the disease. Breaking feedback loops,
however, can lower the incidence and prevalence of AIDS.
As with any disease, AIDS has a congeries of risk factors, including risky behaviors, which
increase the probability of infection, and identifying groups of people associated with risk factors
has been helpful in targeting interventions. Some researchers and activists have been reluctant to
define risk groups, for fear that in doing so, members of those groups would be stigmatized and
that individuals not belonging to any of the risk groups might feel that they were immune. These
are valid concerns, but the overwhelming value of identifying risk groups in order to target
interventions has kept the risk group concept alive in AIDS research.
Walters (1988), for example, was an early supporter of targeting interventions to male
receivers of anal intercourse, IV drug users of both sexes who share needles, and women who
have vaginal intercourse with IV drug users. Walters justified such targeting on two grounds:
intensive coverage of those particularly at risk is likely to be more efficient than general
education messages, and higher risk groups need even deserve more intense warnings of the
risks they are exposed to.
Caldwell and Caldwell also endorse risk group categorization. They contend that "the
temptation to keep governments and individuals vigilant by arguing that the progress of [AIDS]
is inexorable and threatens all parts of society and all societies equally will almost certainly
prove to be self-defeating" (1993:817). Indeed, to insist that all sectors of society (from
sexually-active adolescents to monogamous geriatrics) are equally at risk in order to maintain
vigilance against AIDS can lead to the belief that it apparently doesn't matter what you do:
you're just as likely to get AIDS anyway. This could lead to a laissez-faire attitude toward
proven prevention behaviors on the part of those who need vigilance the most: members of high
Some risk behaviors appear to be determined by group affiliation. A study ofHIV risk
within an Ethiopian/Eritrean community in California noted risky behaviors such as not using
condoms, having multiple sexual partners, consuming alcohol, and avoidance of HIV testing due
to stigma (Beyene 2000). These attitudes and behaviors among immigrants were found to be
very similar to their counterparts in Africa, suggesting that risk behaviors follow cultural lines,
rather than geographical ones, at least in the short term. This underscores all the more the need
to tailor prevention methods to specific cultural contexts.
That risk group definitions may not have cross-cultural validity is suggested by Karnik
(2001), who argues that importing pre-defined risk groupings to India makes little sense in that
cultural context. It also obscures our ability to discern risk groupings of a different sort than are
expected according to Western models. We need to be more open to the idea that there are risk
categories beyond those already identified, but the concept of risk groups has proven effective.
Some groups that have been documented to have greater risk of contracting HIV include
hemophiliacs, HIV-negative partners of infected persons, prisoners, and health care workers
(Walters 1988). In addition, heightened HIV risk has been documented in uncircumcised men
(though this particular factor has been disputed), those receiving blood transfusions, those
practicing dry sex, those practicing mass scarification, and residents of violent war-torn areas
(as invading armies often rape local women) (Caldwell and Caldwell 1993). Prostitutes and
those with other STD's have also been found to be at greater risk of HIV contraction (Walters
1988; Caldwell and Caldwell 1993). In addition, military personnel are at risk for HIV
contraction and dissemination through frequenting prostitutes while separated from their wives
and girlfriends (Walters 1988; Hunt 1996).
A study of American teenagers showed that early sexual debut increases the probability of
other risky sexual behaviors, such as having multiple concurrent sexual partners, forcing a
partner to have sex, and having intercourse while drunk or high (O'Donnell et al. 2001).
5 Dry sex is more than simply avoiding the use of lubricant. It is the practice of inserting desiccating agents into the
vagina prior to intercourse under the premise that drier sex increases friction and pleasure. Unfortunately, increased
friction also increases risks of tearing the vaginal wall, adding the potential of blood-to-blood HIV transmission in
addition to seminal transmission.
Another study of American teenagers and young adults found that alcohol and drug use, as well
early debut, were associated with increased odds of having multiple sexual partners, whereas
marriage was associated with decreased odds of having multiple partners (Santelli et al. 1998).
An interesting twist on the risk group concept suggests that those perceiving themselves to
be at risk from AIDS might also benefit from targeted interventions, even if their actual risk is
low. Prohaska et al. (1990) found that fear of AIDS, concern over one' s health, shame
associated with AIDS, identification as Asian-American, and lack of identification with any
particular religious group were associated with increased perceptions of risk for HIV infection.
One of the most well-known risk factors for HIV infection and transmission is migration.
Rural-urban migration, return migration, refugee migration, tourism, business travel, and travel
associated with the drug trade have all contributed to the spread of HIV. Collectively, the
members of these migrations comprise a potent AIDS risk group. Separated from erstwhile
steady sexual partners, many migrants have sex with new partners and contribute to the spread of
the disease to new locations across societies.
In Africa, it is common (and has been since colonial times) for young men to go abroad in
search of better wages and employment opportunities than are available back home. South
African gold and diamond mines have attracted large numbers of workers for decades. The
mostly male migrants send remittances to their families in Malawi, Zambia, Zimbabwe and other
countries. Long absence from home led many migrants to seek other sexual partners in their new
locale, thus increasing their overall number of lifetime sexual partners. Labor migration has long
been recognized as facilitating the transmission of all sexually-transmitted diseases (Hunt 1989).
There is now enough evidence about HIV transmission, effects, and risk groups to
formulate prevention strategies. Some of those strategies are generated by researchers, far
removed from the situation; others are generated by people on the ground in the worst affected
areas. The next chapter analyzes the most prominent of these prevention strategies to produce a
picture of contemporary AIDS control and mitigation efforts.
AIDS PREVENTION: WHAT HAS WORKED, AND WHAT HASN'T
Strategies for the prevention of HIV/AIDS must grow out of our knowledge of how the
disease is transmitted. Those who favor behaviorist explanations can be expected to focus on
behavioral interventions (such as promoting the lowering of extramarital contacts for truckers).
Those who favor material explanations, on the other hand, focus on changing the structural
conditions (like severe poverty) or infrastructural conditions (like the lack of condoms) known to
be associated with the transmission of HIV.
While this seems straightforward, working in a cross-cultural context inj ects an additional
dimension the superstructure that must be addressed by those working to halt the spread of
the disease. How well has this been done? Wonderfully in some cases, abysmally in others.
Thankfully, much of what has been accomplished and learned through field experience has been
well-documented so that others can eschew interventions with marginal impact and embrace
high-impact interventions that are adapted to local contexts.
Some interventions are promoted alone. Others are introduced as a suite of efforts. This
chapter analyzes each approach in turn.
Individual HIV Prevention Approaches
Though AIDS is a socially and culturally complex disease, some interventions approach it
somewhat simplistically. One benefit of cases where a single intervention has been introduced is
the ability to assess a given approach's impact in a relatively direct, isolated way. Teasing out
interaction effects of simultaneously promoted multiple approaches is far more difficult. But a
maj or drawback to singularly focused prevention efforts is that they tend not to have broader
impacts beyond limited scenarios. Even in their limited target scenarios, single-pronged
interventions may not clearly be effective, indicating that perhaps a multi-pronged approach
might work better.
For example, Robles et al. (1998) studied whether a needle-exchange program (NEP)
would be used by intravenous drug users (IDU's) in San Juan, Puerto Rico, aiming to reduce the
chances that needles would be reused and shared. The NEP reported a 40.3% return rate for
syringes that they had passed out (identified by numerical markings), indicating that some of the
original customers were indeed returning to this reliable source for clean needles, though a
majority still did not. The authors claimed that the program was effective in reducing sharing of
syringes; however, it is not clear whether this was directly measured or simply extrapolated from
the fact that needles were being exchanged (one might share a needle with dozens of people
before returning it). So the results of the study were inconclusive as to whether NEP' s are
effective at reducing sharing of needles. As a single-pronged intervention, the program did not
address other potential sources of HIV infection in IV drug users.
Anti-retroviral drugs (ARV' s) are typically considered to be palliative rather than
preventive interventions that make HIV seropositivity easier to live with. However, they may
have preventive effects through reducing an HIV-positive person's viral load and thus
postponing their chance of developing AIDS. In addition, if certain combinations of ARV' s are
administered within 24 hours of a person's exposure to the virus, they can prevent development
of HIV seropositivity (Centers for Disease Control and Prevention 2001). Relying on ARV' s
alone to protect one from AIDS is neither practical nor recommended in many places (they are
prohibitively expensive and hard to distribute in areas with poor infrastructure). Yet that is
essentially what happens in cases of disinhibition. Because of people' s confidence of being
treated in case of infection, some abandon other prudent prevention measures. A false sense of
security leads some to freely engage in more and more risky sexual behavior, depending upon
drugs to bail them out when things go awry (Green 2003b). And yet, over-reliance on ARV' s
may increase the burden of HIV-infected individuals needing ARV treatment, when there aren't
enough ARV's to go around even now.
In addition, inconsistency in the implementation of treatment regimens can lead to the
predominance of ARV-resistant strains of HIV, which lowers ARV' s effectiveness for
prevention. Chloroquine-resi stant strains of malaria and antibiotic-resi stant strains of pneumonia
are well-documented examples of diseases that have morphed in response to treatment drugs.
This may already be occurring with AIDS. The popular press has picked up on reports of new,
virulent strains of HIV that kill their victims within six months, compared to more common
strains that take up to ten years to produce full-blown AIDS (Brewster 2005; Bartholomew
2005). This, too, demonstrates the need for a variety of HIV prevention methods.
Apart from drugs, HIV prevention is largely a behavioral enterprise. Because it is driven
by social contacts, interventions must address people's social behaviors that might increase or
limit their risk for infection. As Caldwell puts it, "We now know enough about the social
context of the epidemic, and the interventions that would probably succeed, to begin to limit the
epidemic' s impact without waiting for the development of vaccines or depending on
antiretroviral drugs for prolonging life" (2000: 117).
Of all the maj or behavioral HIV prevention methods, none has been studied, promoted,
and implemented alone more than condom use. (Abstinence or fidelity have so rarely been
promoted alone as to make it impossible to analyze their separate impacts). This means that a
large body of evidence has accumulated over the years, allowing one to analyze in-depth the
effectiveness of promoting condom use. For example, in Thailand, most new infections occur
through commercial sex work and a 100% condom policy in Thai brothels has lowered HIV
transmission significantly (Green 2003b). And De Vincenzi (1994) has demonstrated that
among serodiscordant heterosexual couples (who remain sexually active), those who used
condoms consistently (about half of the study participants) had no cases of seroconversion over
the course of some 15 thousand episodes of intercourse, while those who used condoms
inconsistently (the other half) had a seroconversion rate of 4.8 per 100 person-years over the
course of approximately 12 thousand episodes of intercourse.6
Yet condoms have their detractors, too. "AIDS prevention organizations... speak
confidently about 'proven interventions,' which refer to condoms... and treatment of sexually
transmitted infections (STIs)," writes anthropologist Edward C. Green of Harvard University
(2003b:3-4). But Green counters that condoms are not a proven intervention outside tightly
controlled scenarios, such as those described above. Condom interventions are typically
designed based on an understanding of AIDS in America in the mid-1980s, not heterosexual
epidemics in the general population of developing countries today (2003b). The Thai situation
resembles America in that AIDS is concentrated in a high risk group, not the general population.
If condoms are targeted to that risk group, they may impact HIV transmission significantly. But
this does not provide evidence that they will work in an epidemic where the disease is rampant in
the general population, such as in sub-Saharan Africa.
Aside from demonstrating condoms' potential effectiveness in a target group, De
Vincenzi's study also raises an experimental confound. Participating couples were not assigned
to the condom group or the no-condom group; all were counseled about safe sexual practices,
6 Almost half of all study participants ended their sexual relationships before the end of the research project, most
often because of their partner' s illness or death. Those who continued for at least three months after enrollment
were included in the computed statistical results. Thus, the 15 thousand episodes of intercourse (and potential for
infection) are not spread evenly among all participants, and the self deselection of the most ill participants (and their
partners) may skew the data toward representing those with minimal infectivity and viral shedding.
and couples decided for themselves whether they would use a condom or not. Thus, "even
among repeatedly counseled European couples known to be exposed to HIV, nearly half
continued to have unprotected intercourse" (Johnson 1994:391). Among those inconsistent
condom users, there was an estimated 12.7% cumulative incidence after 24 months of exposure.
Furthermore, De Vincenzi's sample was chosen so as to control for the interference of other risk
factors in the study. IV drug users, homosexuals, recipients of unscreened blood transfusions,
those with multiple sexual partners, and those with partners from sub-Saharan Africa were
specifically excluded. The results cannot, then, be extrapolated to a general situation where
some people are monogamous, some are polygamous, some are prostitutes, and some are
abstinent. Johnson concludes:
The difficulties of compliance with condom use, even in a cohort of research subj ects,
should make us pause and think about crusades for the supremacy of the condom in
preventing the transmission of HIV. People need choices in reducing their risk of
infection, but each method requires demonstration of its effectiveness. [1994:392]
This difficulty in achieving consistent condom use is not limited to Europeans in controlled
studies. A study of HIV in the general population of Rakai district, Uganda, found that
consistent condom use significantly reduced HIV incidence. However, only 4.4% of the study
participants reported consistent condom use and 16.5% reported inconsistent use during the past
year (Ahmed et al. 2001).
Indeed, consistency in condom use seems to be an elusive ideal in many parts of the world.
And figures for condoms' effectiveness even when used consistently are hotly debated. In the
course of a lively, non-peer-reviewed debate hosted over several months on the AIDS and
Anthropology Research Group list serve (AARG), Elizabeth Onj oro (a Kenyan who questions
the effectiveness of condoms) confronted Douglas Feldman of SUNY Brockport (a prominent
supporter of condoms) by asking, "While condoms may seem the next best thing in your mind,
the question is how will you convince/influence Africans to use them consistently and correctly
100% of the time in order to achieve 85% effectiveness?" (AIDS and Anthropology Research
How 85% effectiveness was calculated is unclear, but there are other studies similar to De
Vincenzi's yielding very different figures. Hearst and Chen (2004) suggest differences may arise
from random variation, correctness of condom usage, and how correct usage was ascertained,
among other confounders. In addition, if condom effectiveness is less than 100%, then that small
proportion of risk will cumulatively add up over repeated exposures to infected individuals.
Thus, it is meaningless to talk of condom effectiveness in terms of percentages unless number of
exposures to HIV is specified (Mann et al. 2002; Green 2003b; AIDS and Anthropology
Research Group 2003), and this may be an additional source of varying figures.
Several meta-analyses of available data have attempted to combine information from
multiple studies such as De Vincenzi's. Davis and Weller (1999) estimated a level of protection
of approximately 87% and report a level of 80% in a later study (Weller and Davis 2002).
Commissioned by UNAIDS to make sense of all the disparate reports, Hearst and Chen (2004)
ran a search for all reports on condom effectiveness that appeared in peer-reviewed publications,
professional conferences, and public media. The most rigorous studies and those substantiated
by sufficient documentation provided an estimated condom effectiveness of 90% in preventing
HIV transmission. What emerges from these reports is that condoms are highly effective for
preventing transmission of HIV, but they are not fail-proof, even when used consistently and
correctly. When used inconsistently and incorrectly, condom effectiveness plummets.
Aside from technical difficulties (leakage, breakage, and incorrect use), there are other
barriers to condoms' effectiveness. Weintraub pointed out in the AARG discussion that safe sex
condom campaigns have convinced people that risky kinds of sex are safe (anal, multipartner,
etc.), which has minimized motivation to change behaviors (AIDS and Anthropology Research
Group 2003). As with ARV' s, disinhibition causes some people to place undue confidence in
condoms' protective effects and to engage in even riskier sexual behaviors than they would
otherwise, even to the point of canceling out condoms' protective effects (Kajubi et al. 2005).
Additionally, there is marked reticence to accept and promote condoms in many parts of
the developing world. Zambia recently banned condom distribution in schools (starting the day
the U.S. initiated a program to hand them out) under the premise that doing so encourages sex
among youth and goes against Christian values (Shacinda 2004). And Broomhall noted that use
of condoms hasn't significantly increased in Kenya because they haven't been promoted heavily
enough by local health workers, despite repeated training; health workers thought that clients
wouldn't be receptive, and some withheld condoms from groups that they thought would use
them immorally (AIDS and Anthropology Research Group 2003). These barriers raise questions
such as, Should reticent Africans be more aggressively targeted for condom marketing to
overcome their reluctance? Or should Westerners be more willing to work with approaches that
many Africans are already comfortable with, especially those approaches shown to be effective?
Some Westerners are reticent to dictate policy to Africa. Green reports that the founding
chairman of the Kenya National AIDS Control Council, Muhammad Abdullah once told him, "I
am sure you are aware that at times help comes with strings attached. We received US $10
million aid but it had to be only condoms. We had two options: to accept all the US $10 million
worth of condoms or refuse" (AIDS and Anthropology Research Group 2003). Feldman agrees
that dictating policy for Africa from the West is not appropriate but argues that "applied cultural
anthropologists need to take the lead in finding out how to design culturally appropriate
interventions that would make proper condom use routine among all sexually active
multipartnering males in sub-Saharan Africa" (2003:7). Is promoting routine condom use itself a
form of policy dictation? Should condoms be a non-negotiable part of foreign aid in the fight
against AIDS? Or can HIV prevention programs be responsibly tailored to local sensibilities -
even if that precludes condoms in some cases? Because billions of donor dollars ride upon these
policy questions, many African countries are caught in the unfortunate squeeze that Mr.
But addressing challenges such as these makes some uncomfortable. "[Some] have even
questioned whether condoms are safe and effective," Feldman wrote (2003:7). Remes, an
AARG contributor, fretted that Africans are picking up on this argument and promoting it,
saying that condoms can't protect against disease and adding to other rumors already circulating
about condoms (AIDS and Anthropology Research Group 2003). "I am very uneasy when talk
of condoms being ineffective start circulating for fear the message will be twisted and re-
interpreted in harmful ways," added Broomhall (AIDS and Anthropology Research Group 2003).
"Concern that facts might be 'twisted' is no reason to hide the truth from Africans," countered
Weintraub. "That' s an essentially colonialist attitude: Africans are children and if we tell them
the facts about failure rates they'll become hopelessly muddled" (AIDS and Anthropology
Research Group 2003).
Green points out that although there is wide consensus in the academic community that
AIDS prevention (based predominantly on condom promotion) has not been very effective so
far, many journal articles and the scholars who write those articles argue that condoms are the
most effective way to prevent HIV transmission and that the solution to the problem of
transmission is to redouble condom promotion efforts (2003b). For example, Broomhall writes,
"I firmly believe that the cumulative effect of condom promotion, HIV prevention education,
voluntary HIV counseling and testing, and the increasing availability of ARVs, eventually will
help to achieve a reduction in HIV/AIDS throughout [sub-Saharan Africa]" (AIDS and
Anthropology Research Group 2003). Hearst and Chen, however, report that "The public health
benefit of condom promotion in settings with widespread heterosexual transmission... remains
unestablished" (2004:39). Despite a lack of supporting evidence for a condom-only approach
and the presence of evidence questioning it, condom promotion is viewed by some as the only
realistic option to consider.
Feldman acknowledges condom failure rates but argues that "if you don't use condoms at
all the failure rate is 100%!... We need to really try condom use in Africa before we say it
doesn't work... Condoms are the single most important thing Africans can do to reduce HIV
transmission on the continent. The fact that only 4.6 condoms are available per person per year
in Africa tells me that it hasn't actually been tried" (AIDS and Anthropology Research Group
2003; cf. Feldman 2003:7). In fact, while Shelton and Johnston (2001) estimate an average of
just 4.6 condoms per male aged 15-59 per year in sub-Saharan Africa, the rate varies widely
across nations. In places where condoms have been widely distributed and promoted, AIDS
rates are the highest and climbing (Hearst and Chen 2004; Allen and Heald 2004). For example,
between 1993 and 2001, condom sales increased from one million to three million in Botswana,
during which time HIV prevalence rose from 27 to 45%. At the same time in Cameroon,
condom sales increased from 6 million to 15 million and HIV prevalence rose from three percent
to nine percent (Hearst and Chen 2004). "Of course, prevalence might have risen even faster
without increased condom use," write Hearst and Chen, "but no clear examples have emerged
yet of a country that has turned back a generalized epidemic primarily by means of condom
It is curious, then, that condoms are such a popular remedy in Africa, almost to the
exclusion of other strategies. It may partly be due to ease of monitoring; technological fixes are
a tempting way to solve complex behavioral problems, not least because condoms and STD
drugs can easily be counted. This is a great advantage when reporting the results of AIDS
interventions to donors (Green 2003b). But is counting truly a measure of effectiveness? One
might call an intervention successful if 5,000 condoms were distributed and 900 people were
educated on how to use them properly. But if the incidence of AIDS does not go down as a
result, would it still be considered a successful intervention? With donors demanding immediate
reports (and making future funding contingent on those reports), it is tempting to rely on
interventions that can be quickly counted and monitored, regardless of their ultimate impact.
Despite the critiques, condoms have an important role to play in HIV prevention. Lives
have been saved by condoms. But they are not the panacea for AIDS prevention. The evidence
calls for a multi-faceted approach. Moreover, it is unrealistic to imagine condoms as the only
HIV prevention device for a population of some 890 million people, as in the case of Africa.
The collective effect of multiple highly effective interventions has the potential to be greater than
any single method alone. Because no single strategy is likely to be accepted by everyone, multi-
faceted approaches have better potential for adoption and sustainability. The following section
addresses approaches that combine multiple strategies.
Multi-Faceted HIV Prevention Approaches
Kelly (1995) has assessed what successful HIV interventions targeted to widely divergent
risk groups have in common. All had similar conceptual foundations and similar intervention
procedures; all exhibited cultural tailoring of prevention messages and methods; and all were
longer and more intense than less successful interventions. Green (2003b) has found that
successful approaches to prevention worldwide share high-level government commitment,
primary behavior change (PBC) strategies, and a multilevel response in common. Green writes
that "PBC seems to be the natural response to concern over, or fear of, HIV infection"
(2003b:278), indicating that even when PBC is not actively promoted, most people engage in it.
In many of the successful countries, PBC was manifested by a rise in age of sexual debut
(increased abstinence) and a decrease in number of sexual partners. PBC may have other
manifestations, too, but they can all be broken down into two main categories: risk avoidance
and risk reduction. Risk avoidance includes complete abstinence and perfect mutual partner
fidelity. For these two risk avoidance behaviors, exposure to sexually-transmitted HIV is
reduced to zero. Risk reduction, by contrast, minimizes but does not eliminate exposure to
HIV. The use of condoms, for example, is a risk reduction behavior. Risk reduction programs
assume that people will never change their sexual behavior, so the best we can do is to protect
people while they do what they want.
HIV peaked from 1982-84 among male homosexuals in England and Wales. Changes in
sexual behavior (PBC) as early as 1983 brought about a prevalence decline from 1985 onward.
Johnson and Gill write, "As the risk factors for AIDS emerged, the homosexual community
mounted an impressive education campaign with three main messages for risk reduction: avoid
unprotected, penetrative anal intercourse; reduce numbers of partners; and use condoms
However, Johnson and Gill warn that behaviorall risk reduction will only reduce the
incidence of infection if it is of a magnitude sufficient to outweigh increased risk inherent in the
rising prevalence" (1989: 116). In other words, supposing 90% of homosexuals were HIV-
positive, then even selecting only one partner and sticking to him faithfully would be extremely
risky. Exactly where the tipping point between risk reduction behaviors and rising prevalence
lies is unknown, but evidence shows that multi-faceted behavioral modification made a
difference early on within the homosexual community of England and Wales.
Uganda is another country that has implemented a strategy of primary behavior change to
fight AIDS. Unlike England and Wales, however, the effort in Uganda was on a national level,
and not just with one particular community. Uganda is the only African country to see a large
drop in AIDS rates from 20.6% in 1991 to 6.1% in 2000, according to one study (Green
2003b), or from around 15% in 1991 to 5% in 2001, according to another study (Hogle 2002) -
and as such, it has been the center of an important debate among researchers trying to understand
what brought about this dramatic prevalence decline. As a rare success story in Africa, Uganda
has been studied and restudied and analyzed from many angles by scholars from many
disciplines. To this day, the debate continues as to what exactly is responsible for Uganda' s
At the center of the debate is the ABC program, instituted by Uganda' s president Yoweri
Museveni early in the e idemic. ABC stands for: A be Abstinent as the first line of defense
against HIV transmission; B if you can't be abstinent, then Be faithful to your sexual partners)
of choice; C if you simpl cannot be abstinent or faithful, then use a Condom. This pro ram is
a multi-faceted approach, much like the one that developed spontaneously among homosexuals
during the 1980s in England and Wales (Johnson and Gill 1989). But beyond being multi-
faceted, the ABC program is also explicit about which groups should utilize which
componentss. No single component is for everybody, and certain components are more
effective for specific types of people (condom promotion, for example, would have limited
impact upon those who have chosen to be abstinent).
Multi-faceted disease prevention methods are well known outside of AIDS. Malaria in
Africa is fought on several fronts which, curiously, somewhat parallel the components of the
ABC model: prevent mosquitoes from feeding on people (avoidance of the vector altogether,
analogous to sexual abstinence for AIDS prevention; or minimization of exposure to the vector
via physical barriers, analogous to condom use); prevent or reduce mosquito breeding (reduce
chances for exposure to the vector, analogous to partner reduction in AIDS prevention); destroy
adult mosquitoes and eliminate malaria parasites from human hosts (attack the reservoirs of the
disease within the population, analogous to the use of ARV' s) (MacCormack 1984). It has long
made sense to take a multi-pronged approach to the prevention of any infectious disease; AIDS
should be no different.
Thus, Uganda attacked AIDS on several fronts with remarkable success, including not only
ABC promotion, but also high-level government commitment, reducing AIDS-related stigma
(through social marketing programs by the government and NGO's), advancing the status of
women and youth, requiring a true partnership between Ugandan authorities and donors
(enabling a collaborative but home-grown approach to emerge), using multiple media to blanket
the country with AIDS messages, using fear arousal to avert risky behavior, promoting education
about AIDS and sex in schools, involving faith-based organizations in prevention, making
voluntary counseling and testing widely available, involving traditional healers in the fight, and
targeting high-risk groups for special intervention (Green 2003b; Hogle 2002). However, Green
cauti ons :
It must be remembered that many of the elements of Uganda' s response, namely,
decentralized planning and multisectoral responses, do not impact HIV infection rates
directly. Behavior must change for this to happen... Although Eighting stigma or bold
political leadership at the highest levels might be cited as maj or contributing factors in
Uganda's success, it must be remembered that these are indirect factors. We must
understand which behavior changed, and how and why they [sic] changed if a Uganda
model of prevention is to be replicated elsewhere. [2003b:221-222]
This raises an important point, because some argue that it's hard to separate out which
factors had the most impact (see Remes in AIDS and Anthropology Research Group 2003, for
example). Though they may be important, however, contributing factors are indirect factors
only. They cannot prevent AIDS transmission alone; they are useful only insofar as they affect
HIV transmitting behavior itself.
One way to test whether national ABC policies make any difference in HIV prevalence is
to compare Uganda with other countries. A study of the Kagera region of Tanzania (which
borders Uganda) suggests that since the beginning of the AIDS epidemic, "there have been
significant changes in sexual behaviors, norms, values, and customs that are considered high-risk
for HIV transmission" (Lugalla et al. 2004:185), including an increase in condom use,
abstinence, zero grazing (sexual Eidelity), and voluntary HIV testing, with a concurrent decrease
in polygyny, levirate, excessive alcohol consumption, and sexual networking. The result has
been a decline in both HIV prevalence and incidence (Lugalla et al. 2004).
On the other hand, Botswana and Uganda differ significantly not only in HIV
prevalence, but in approaches to HIV prevention. While both countries introduced AIDS
awareness campaigns at the same time, have accepted external assistance, and have been open to
international advice, Uganda has been successful while Botswana failed in HIV prevention.
Allen and Heald (2004) argue that this is because the promotion of condoms at an early stage in
Botswana was counterproductive (perhaps shutting down exploration of other strategies),
whereas the lack of condoms in Uganda early on forced it to develop alternate strategies which
(coupled with condoms later on) provided a successful, multi-faceted attack on AIDS. Likewise,
a comparison of the border region between Uganda and Kenya revealed that all Ugandan sentinel
sites showed a decrease in HIV prevalence over time while all Kenyan sentinel sites showed an
increase in HIV prevalence over time. With no prominent differences in ethnic groups and
practices on both sides of the border, Moore and Hogg (2004:542) conclude that "decreasing
HIV prevalence in Uganda is not due to the natural course of the epidemic but reflects real
success in terms of HIV control policies."
As is the case with condoms, there are skeptics about the effectiveness of the ABC
approach. Feldman claims that the immediate (and short-term) reaction to AIDS was that
Ugandans reduced their number of partners which lowered HIV prevalence and that this was
sustained by condoms brought in en masse in the mid-1990s. Feldman' s concern about
acknowledging and supporting A and B is that those who practice them will denigrate those who
don't. He asserts that abstinence and fidelity should not be parts of national HIV prevention
policies for everyone, and that if it' s true that 95% of Ugandans had zero or one partners in 1995
(while only 6% reported using a condom at last sex), "then we are talking about a rather
pathological condition in Uganda that should certainly not be used as a model anywhere else in
Africa" (AIDS and Anthropology Research Group 2003).
Green counters, "My basic thesis is that we on the Western donor side need to move from
consensus based to evidence-based AIDS prevention... There was a time when I thought our
current approach of promoting the maximum number of condoms... would actually work. Look
at the data in Africa and you will see otherwise" (AIDS and Anthropology Research Group
2003). The data point out that Uganda' s precipitous drop in HIV prevalence began well before
condom social marketing and availability became widespread. Even after condoms began to be
promoted their use remained low for a long time. By 2000, 93% of Ugandans aged 15-49
reported abstinence or partner fidelity, while only 8% reported condom use during last
intercourse (the best available measure of regular use at the time). "It is difficult to conclude",
Green writes, "that a risk reduction behavior reported by 8% of the population is contributing
more to HIV infection aversion than two risk avoidance behaviors reported by 93% of the
population, even in 2000 when reported condom use is highest" (2003b:151). Regarding the
ABC debate on the AARG list serve, Singer wrote:
One of the issues under discussion is: do condoms really work in Uganda? This is an
empirical question not a moral one. So far, the facts are in dispute. When facts are in
dispute... it is easy for those engaged in debate to cite their favorite studies and data
bases... we are generally blind to our own cultural assumptions. The culture in them is
hidden to us, because they seem to make so much damn sense. And so we screw-up.
[AIDS and Anthropology Research Group 2003]
Green says that the bias in AIDS prevention circles against risk avoidance is a consequence
of the American sexual revolution and the emergence of a cultural norm that favors protecting
people from the risks of pregnancy or STD's rather than asking them to change their behavior.
In addition, those who have been promoting condoms for so long may have a hard time
admitting that primary behavior change is an essential factor in Eighting AIDS. "Professional
reputations and egos are at stake," Green writes. "People don't want to admit that they may have
completely missed something vitally important for years" (2003b:80).
Finally, many who work in AIDS prevention have family planning backgrounds and are
predisposed to think of contraceptives for STD prevention. They are wary of approaches that are
religiously condoned, since some high-proHile religious groups oppose contraception. Yet family
planning experts know that the condom is one of the least effective methods of contraception.
"How could the paradox have escaped them that the same method is considered the most
effective method of preventing HIV infection?" Green asks (2003b:78).7 Green also points out
that AIDS prevention is a substantial industry in the United States:
It would be politically naive to expect that those who profit from this industry would not
be inclined to protect their interests. Those who work in condom promotion and STD
treatment... do not want to lose market share, so to speak, and so they may go out of their
way to ignore, disguise, or discredit findings that show something else is working to bring
down HIV infection rates. [Green 2003b~:79]
As a result, even though lip service is paid to balanced approaches, nearly all donor funds
go into condom promotion alone.8 In addition, program indicators almost always focus entirely
on condoms. Green calls such approaches Pattern I solutions for Pattern II problems (2003b).
Those who are wary of approaches that are religiously condoned often link ABC with
religion. Feldman argues that "a commitment to abstinence and faithfulness, and the growth of
fundamentalist religion" was not what turned the AIDS epidemic around in Uganda (2003:6),
and asserts that:
trying to impose a sex-negative morality across all African cultures will not only fail to
reduce HIV seroprevalence, but it will only bolster the rapidly growing danger of
fundamentalist religion on the continent, and take Africa on a downward spiral into sexual
repression and hostility... Foisting American religious fundamentalism and
abstinence/fidelity messages on all of Africa as an excuse to control HIV is not the way to
go. [AIDS and Anthropology Research Group 2003]
Schoepf believes that A and B should be promoted via any method other than religious
NGO's "because they are a) counter-productive with some population sub-groups and b) violate
the US Constitution which mandates separation of churches and the state" (AIDS and
Anthropology Research Group 2003). It is true that religion-based NGO's may be ineffective
SThis is not necessarily a paradox: it could be possible that condoms were the best for HIV prevention, even if they
were not good for contraception. Pills, for example, are considered one of the best methods of contraception, but
they are ineffective for HIV prevention.
SIt is encouraging to note that at least in the case of USAID (one of many organizations committed to fighting
AIDS), the President's Emergency Plan For AIDS Relief (PEPFAR) now aims to earmark funds equally for all three
components of ABC one-third of the funds for each part. This has not been historically so, however, and it is an
encouraging break from the past.
with some population sub-groups, but the same can be said for secular organizations that try to
promote behavior change among highly religious people. Nothing can be expected to appeal to
everybody. Schoepf s church-and-state argument assumes that AIDS control belongs to the
state. In fact, most of the real, on-the-ground work against AIDS in many African countries is
done by NGO's. This is recognized by USAID, which has supported NGO's, including religion-
based NGO's, like ADRA and World Vision, since at least 1991 in the fight against AIDS in
Finally, Green reminds us that Africans are highly religious, and often begin meetings with
prayer. "What should the participating-observing anthropologist do in such a situation? Disrupt
the meeting and tell them they are violating the US constitution [sic]? Inform them that religion
is the opium of the people?" He continues, "I am sorry if [A and B] are not behaviors many of
us approve of because of American struggles with the religious right. But that' s what happened.
Maybe its [sic] time for a little cultural relativism and acceptance that something happened that
we did not anticipate" (AIDS and Anthropology Research Group 2003). He adds:
[Feldman] assumes that all religions found in Africa are fundamentalist and condemners of
HIV+ people... the religious groups that worked in AIDS prevention in Uganda...
Anglicans, Catholics and Muslims... are not terrorists [sic] missionary groups who just
want to spoil fun; they are groups with great influence in Africa whether Doug and I like it
or not... Don't exclude A and B inteventions [sic] as 'missionary terrorism.' [AIDS and
Anthropology Research Group 2003]
Onj oro adds, "I guess our friend Doug is taking the stand that he knows more about
Africans [sic] sex life and behavior than Africans themselves... As an African, I know that
abstinence is built into many cultures as parts of rituals, and was not introduced by conservative
Christians." Onj oro said that her parents not the church taught her abstinence (for the sake of
not getting pregnant) and that Africans don't have hang-ups about religion getting mixed up with
A and B (AIDS and Anthropology Research Group 2003).
Though A and B are not inherently religious, Africans often couch them in religious terms.
"I sometimes put it this way," Green writes. "It doesn't matter if most of my Western colleagues
and I happen to be urban, liberal, and secular; most Africans are rural, conservative, and
religious. When we are designing and implementing programs in Africa, we must always
remember where we are" (2003b:324). Research has shown that countries that have been
successful in HIV prevention have tended to include faith-based organizations (FBO's) in the
Eight against AIDS. Those organizations already have extensive networks in place with which to
disseminate information, and they tend to have a natural inclination toward promoting A and B
(Green 2003b). One way FBO's can get involved in HIV prevention is through collaborative
work where each actor performs according to their particular strengths (Smith et al. 2004). Why
force FBO's to promote condoms if they're really good at promoting Eidelity? Better to utilize
their strengths and pair them up with other agencies that will fi11 in the gaps with their particular
Despite the fact that abstinence and monogamy promotion are not inherently religious,
some still see a dangerous link between the two. Feldman writes:
Fundamentalist churches and mosques have dramatically grown in their social and political
influence during the past decade, casting a pall over sexual freedom and expression across
the African continent. The last thing that Africa needs now is an ineffective, culturally
inappropriate HIV prevention program based upon a misinterpretation of the data that will
further embolden these regressive religious organizations. [2003:6-7]
This excerpt makes clear that sexual freedom is an important part of the debate. Feldman
supports what he calls ACCDGLMT: A (anti-discrimination), C (condoms), C (culturally-
appropriate interventions), D (destigmatization), G (government involvement), L (less risky sex),
M (media campaigns), and T (traditional healers) (AIDS and Anthropology Research Group
2003). Additionally, he supports getting elders to conduct initiations safely, encouraging
masturbation among youth as an alternative to intercourse, and promoting acceptance of same-
sex sexual behavior (Feldman 2003:7). It is interesting to note that this approach expressly omits
abstinence or fidelity, key components in Uganda' s struggle.
In another posting, Bailey writes, "Promoting condoms does not mean, and seldom ever
meant, ignoring messages of faithfulness. Treating STI' s does not mean, and seldom ever meant,
ignoring faithfulness or condoms" (AIDS and Anthropology Research Group 2003). In other
words, there would be no issue about condoms versus abstinence versus fidelity versus STI
treatment (ad infinitum) if each aspect of prevention were receiving all the funding it needed.
Since funding is limited, if one approach is well-funded, then other approaches will face budget
shortfalls. This may account for part of the vigor with which this debate is carried out.
And despite the utility of moving beyond ABC to include contributing factors, Green notes
that his disagreement with Feldman boils down to the role of condoms and partner reduction
(which cannot be reduced to "abstinence" and "shaming"). "Partner reduction has worked. Lets
[sic] not dismiss it as 'abstinence and spreading stigma' because some of us persist in seeing it
though [sic] an American lens" (Green in AIDS and Anthropology Research Group 2003). This
is an interesting point, because some who promote greatly expanded acronyms consider the ABC
approach to not be multi-faceted enough. Because of the intensity of public debate in the United
States over abstinence-only sex education, some mistakenly consider ABC to be an abstinence-
only approach, too. Green recounts:
Many an AIDS expert has likewise told me that abstinence sounds wonderful, but of
course, this is not a realistic option--especially in Africa. The unspoken assumption or
implication here is that Africans are more sexual than the rest of the world and that they
could never control their sexual behavior. If I point out that this is racial stereotyping and,
in fact, at odds with existing data, my colleagues might say something like, Well, you
know, in the tropics, in these little villages, there is little else to do but have sex. It's all
they have for recreation. If I mention the evidence about delay of debut or partner
reduction, many immediately think this is a trick to weaken their critical judgment. Or it is
a Troj an horse that will be used to slip abstinence-only programs through their defenses
and into AIDS prevention. [2003b~:87]
Despite strong opposition from some, ABC has now become part of PEPFAR the
President' s Emergency Plan for AIDS Relief. The worry seems to have been that ABC would
translate into AB, but the document that provides guidance to implementers of PEPFAR states,
"Implementing partners must not promote condoms in a way that implies that it is acceptable to
engage in risky sex... Likewise, abstinence and faithfulness programs and messages must be
medically sound and based on best practices that indicate effectiveness" (Tobias 2005:6). Thus,
all three components are subj ect to full disclosure regarding effectiveness, without favoring any
of the three.
It has been pointed out that abstinence and condom use have strong advocates, but that
fidelity/partner reduction does not. Wilson writes, "As AIDS educators, we often publicly
promote approaches that we would not countenance in our own personal lives, such as the notion
that it is acceptable for our spouses or children to have multiple partners, provided condoms are
used." He continues, "Partner reduction is good epidemiology, not good ideology, and we must
ensure that the ABC approach remains sufficiently scientifically grounded to withstand shifting
ideological sands" (2004:848).
Some fear that findings about ABC may be used as a political tool to advance conservative
causes. "(I am concerned", Pach writes to Green, "whether you have presented and clarified your
reasonable and empirical views for the benefit of a wider, less informed, potentially more
confused, yet powerful audience?" (AIDS and Anthropology Research Group 2003). Schoepf
recommends avoiding situations where misinterpretations might arise, and encourages Green to
clarify that though A and B are effective, they don't work for everyone (AIDS and Anthropology
Research Group 2003).
In response, Green says that just because an extremist may take some of his ideas and run
with them doesn't mean he should not state his centrist position in the highly polarized
"abstinence versus condoms" debate. "I am sorry if conservatives like some of the data I have
been bringing to light. I don't have control over the political use of empirical findings. All I can
do is try to say the right thing myself... and move AIDS prevention in a more evidence-based
direction" (AIDS and Anthropology Research Group 2003).
One issue that challenges the successful adoption of ABC is the empowerment of women
to enforce it. If a woman cannot say "no" to her husband if she suspects he has become infected
through infidelity, or if she cannot resist sexual advances from others, or if she is not in a
position to mandate condom use with her partners, ABC can potentially fall apart. That's why
Uganda's successful example of ABC implementation involved real structural changes that
empowered women (greater education leading to economic sufficiency, and strong enforceable
laws against the exploitation of women, for example).
A study in Kampala, Uganda noted that Baganda women typically adhere to cultural bans
on sex outside of marriage (with a few key exceptions). However, they still fear contracting
AIDS since their partners are not culturally required to avoid sex outside of marriage and they do
not do so voluntarily (McGrath et al. 1992). This is a concern frequently brought up by those
who question ABC, and it is a valid one. The effectiveness of B, for example, can be
experienced even if only one partner is faithful, but it is most effective when practiced by
everyone in a sexual network (either two faithful monogamous partners, or multiple polygamous
partners who are exclusively faithful to each other). That is why ABC messages need to be
widespread and targeted to multiple levels of society to work.
Finally, some question the validity of findings about ABC's effectiveness based upon self-
reported behaviors. For those wont to discredit data about A and B self-reporting, Green notes
that self-reporting about C behaviors is subj ect to just as much unreliability as A and B (2003b).
Consistency demands that anyone who discredits the possibility that self-reported A and B
behavior is accurate would need to discredit the possibility that self-reported C behavior is
accurate, too. This does not guarantee that self-reported A and B behaviors are not skewed, but
they are not any more likely to be skewed than self-reported C behavior.
Despite its turbulent history, ABC's effectiveness on a population-level epidemic as a
multi-faceted, multi-level response is much better supported by available evidence than is
condom use alone. Of course, no single intervention is likely to be adopted by everyone, but a
multi-faceted offering of individually effective strategies allows people to choose which
approach fits their circumstances the best. Still, there have been attempts to document cases
where a condom-only approach might have worked. One example is a recent study from
In early 2005, reports of an unpublished study started leaking to the press. The study
claimed to find that ABC was not responsible for HIV prevalence decline in Rakai district,
Uganda, but that premature death from AIDS and increased condom use were. The unpublished
paper was presented to the 12th Conference on Retroviruses and Opportunistic Infections in
Boston with significant press coverage. According to the study, percentages of individuals
practicing either abstinence or monogamy remained the same or decreased slightly during the
period 1994-2003 during which time AIDS prevalence dropped from 20% to 13%. Yet HIV
incidence slightly increased during that time frame, while condom use increased significantly.
The conclusion was that Uganda's much lauded success with fighting AIDS has little to do with
abstinence and monogamy promotion (Kaiser Family Foundation 2005).
Until the study is published in a peer-reviewed j oumal, one can not endorse nor dismiss its
findings. However, some concerns arise from the news report itself. First, Uganda's prevalence
decline began well before the mid-1990s not from the mid-1990s to 2003 (Green 2003b);
Hogle (2002) indicates that national HIV seroprevalence peaked in 1991, meaning that incidence
peaked even earlier (Hogle estimates that it was some time in the late 1980s). Second, while
rates may have dropped in the Rakai district during the study period, the decline began at the
national level much earlier. Thus the Rakai study does not support the claim that abstinence and
monogamy were not responsible for Uganda' s HIV decline as a whole. Since incidence of HIV
remained low throughout the study without any significant increase, "the real cause of the
success was a large reduction in new infections before the study began" (James 2004). It is quite
possible (when looking at the full timeline) that even though fidelity and abstinence declined
somewhat during the span of time studied, rates of both overall were still significantly higher
than necessary to tilt the balance toward prevalence decline.
If anything, one should be concerned that during a time when the only thing that increased
significantly was condom use, HIV incidence also increased slightly. The increase may still be
low enough not to reverse prevalence decline, but one must not ignore the covariance of
increased condom use and increased HIV incidence. The popular press may seize upon works
and make claims that are not warranted, but this should not be taken as scientific support for the
findings. Traditional scientific checks and balances of peer-review and publication are still
needed to validate study findings.
The British M~edical Journal (BMJ) took the unprecedented step of endorsing the study
before publication, stating emphatically, "Use of condoms and death explain the substantial
decline in the prevalence of HIV in Uganda in the past decade" (Roehr 2005:498). This move by
the BMJ produced immediate reaction. One letter to the editor pointed out that an enormous
amount of peer-reviewed literature has established that all three components of ABC were
important in reducing HIV in Uganda and questioned the BMJ' s position. "Prof Edward Green,"
the letter said, "predicted that many Western donors were determined to unpick the ABC
strategy, rather than model future policy on it... This BMJ news item indicates that Professor
Green is a true prophet" (Stammers 2005). "Unfortunately," wrote another correspondent, "the
BMJ article and others confuse incidence with prevalence. I'm afraid the current batch of news
articles based on Dr. Wawer' s presentation [at the conference] represent more of the pointless A
versus C debate" (Shelton 2005). Yet another wrote, "the only biologically plausible way the
number of deaths could exceed the number of cases in a given year is if the HIV incidence in
earlier years was much higher... and had declined to current levels" (Mosley 2005).
Furthermore, "these findings do not support a conclusion that condom use can result in a
reduction in HIV incidence (and prevalence), but rather that the combination of condoms with
more sexual partners (behavioral disinhibition?) will only maintain incidence at a high and
steady state" (Mosley 2005).
Though the debate rages on, it appears that multi-faceted ABC is gaining a following.
Increasing numbers of researchers and experts are endorsing the ABC approach to AIDS
prevention. In late 2004, a statement entitled "The time has come for common ground on
preventing sexual transmission of HIV" was published in the British medical j ournal, the Lancet.
Authored by Daniel Halperin, Edward Green, Norman Hearst and others, the statement included
a list of those who endorsed it about 150 individuals, including influential AIDS thinkers such
as Paul Farmer, Daniel Low-Beer, Yoweri Museveni, Rand Stoneburner, and Archbishop
Desmond Tutu. The statement said:
Changing or maintaining of behaviors aimed at risk avoidance and risk reduction must
remain the cornerstone of HIV prevention. We call for an end to polarising debate and
urge the international community to unite around an inclusive evidence-based approach to
slow the spread of sexually transmitted HIV. [Halperin et al. 2004: 1913]
The statement acknowledges that the status quo in prevention is no longer acceptable. It
urges that prevention should focus on programs that are locally-endorsed and relevant to the
indigenous context, acknowledges that all three elements of ABC are essential to reducing HIV
incidence, promotes a multisectoral approach to prevention, supports significant expansion of
HIV testing and counseling services, and endorses the pursuit of medical interventions for AIDS
in a way that does not impede adoption of preventive behaviors.
"Common ground" statements such as these are not likely to silence the debate once and
for all. It will continue as long as there are competing interests with limited resources to fight
AIDS worldwide. But in light of those limited resources, it only seems prudent to reserve
limited funds for those approaches shown by evidence to be most effective for AIDS prevention.
ABC may not be a panacea either, but it has repeatedly been shown by evidence to be the best
approach for AIDS prevention so far.
THE EDUCATIONAL MODEL OF SOCIAL CHANGE
Despite the preponderance of data suggesting the superiority of a multi-faceted approach to
AIDS prevention and health promotion in general (see Valente 2002; Backer et al. 1992;
Reardon 1989), mono-faceted approaches are still remarkably popular. This is partly because (as
previously suggested) they are more easily monitored for impact. The most popular mono-
faceted approach is education and information campaigns. Problems with condom use
compliance? Educate people further. Can't get gays to stop having risky sex? Instruct them in
ways to have safer sex. The underlying assumption behind such an approach is that people
aren't avoiding AIDS because they don't know how. Give them enough education and
information and behavior will change automatically (Bernard 2002). This approach often
doesn't consider whether or not an intervention is compatible with the existing values of
potential adopters, an important prerequisite to acceptance of the intervention itself,
notwithstanding whether or not a new behavior is actually adopted (Rogers 2003).
Bernard (2002) calls this the educational model of social change (EMSC). The idea,
according to Bernard, is that since the last thing that happens before an action is a thought, if you
want people to engage in better actions, then give them better thoughts through education
interventions. The EMSC is widespread throughout Western society and is the foundation for a
maj or international industry. This chapter addresses the theoretical underpinnings of the EMSC,
its manifestations in contemporary society, its ubiquity in AIDS prevention, and some of the
weaknesses of relying on the EMSC alone as a mono-faceted AIDS prevention strategy.
The educational model of social change highlights the differences between an idealist and
a materialist paradigm for explaining differences in human behavior across time and space.
Idealism, espoused by thinkers such as Claude Levi-Strauss and Edmund Leach, proceeds from
the premise that reality (and its attendant behaviors) begins in the mind that is, with ideas and
that behavior is influenced primarily by the way one thinks about the world. Thus, the best way
to change behavior is to assess and then intervene in people's thoughts about that behavior.
Educational interventions, information campaigns, and persuasive speeches are the hallmarks of
the idealist approach to behavior change.
The materialist perspective, advanced by theorists such as Karl Marx (1998) and Marvin
Harris (1979), proceeds from the belief that behavior is primarily influenced by structural and
infrastructural conditions. From the materialist perspective, behavior change is best effected by
improving road systems (to make commodities more accessible), by boosting local economies
(to increase purchasing power), and by changing laws and penalties (to provide tangible
incentives/di incentives for certain behaviors).
In fact, idealism and materialism are both required to understand much behavior change,
particularly in the short term; this is what Hornik et al. (2002:219) call a "match between
intervention goals and structural opportunities for behavior change." If all physical and financial
barriers are removed from participation in immunization programs, for example, but no one
knows about the programs and their utility, then no will use them (see Zimicki et al. 2002).
Conversely, if people are educated about the benefits of vaccinations but lack the financial or
physical resources (like transport) to take advantage of them, then they will not. Valente
concludes, "It is unwise to generate demand for products and services that cannot be supplied,
and inappropriate to supply products for which there is no demand" (2002:50).
That the EMSC is ineffective at structural and infrastructural levels is a point that has been
principally argued by Bernard (2002). "The closer a behavior is to the culture (or superstructure)
of society, the easier it is to intervene culturally," Bernard writes. "But if people' s behavior is
rooted in the structure or infrastructure of society, then forget about changing their behavior by
educating them to have better attitudes" (2002:82). For example, unless a supply chain of
vegetable delivery is established (involving transportation, price accessibility, etc.), no amount of
educating Arctic peoples on the merits of vegetable consumption would induce them to adopt
such a diet in a land where none could grow. A child who learns about the dangers of second-
hand smoke in school (a superstructural intervention) is not in a position to move out of the
house (structural constraint) if the parents do not quit smoking at home. Holtgrave et al.
(1995:25) conclude that "at best, communication campaigns will cause people to make
incremental movements toward behavior change."
The Failure of the Educational Model of Social Change
Piotrow et al. (1997:26) write that "if knowledge and attitudes are observed to change
within the period under study, the likelihood that behavior eventually will change, too, is
increased." Maibach and Cotton (1995:44-45) counter, "It takes little more than common sense
to appreciate the fact that knowledge is a necessary precondition for behavior change.
Conversely, although knowledge is necessary, it is not sufficient to motivate or activate behavior
change." Just ask cigarette smokers in America; 80-90% of adults are aware of various health
risks associated with smoking, yet in the 1987 Great American Smoke-Out, only 11.5% of
participants were able to quit for a full day, and only 7.3% maintained that for 1-3 days
(McAlister et al. 1989). This seemingly obvious lesson turns out not to be so obvious. "The
educational model," Bernard writes (2002:33), "is the basis for one of the world' s biggest
industries--social change and development." The results of research and interventions based on
the EMSC are reported in terms of how many people were educated, or how intentions, attitudes,
and beliefs changed as a result of the intervention. Changes in actual behaviors are rarely
reported. Valente (2002) argues that this is an inappropriate method for evaluating the success of
behavior change campaigns. Smith, suggesting best practices for educational interventions,
The focus should be on behavior, not attitudes, knowledge, beliefs, or any other
intermediate variable such as communication products, channel exposure, or public
support. Although these intermediate variables can be indispensable as a means to
influence the behavior, they should not be the focus for measuring success. [2002:334]
Yet all too often, intermediate variables are the measure of success. Raghubir and Menon
(1998), for example, found that educational messages are taken to heart more thoroughly if they
are linked to personal experience. Self-positivity bias (the belief that one is less likely to
contract AIDS than others are) is reduced and attitudes and intentions regarding sexual behavior
are swayed toward safer sex if people recall occasions when they have engaged in risky sex or
know someone who has experienced the consequences of unsafe sex personally. All of this is
good and useful information for designing better social marketing campaigns, but it does not tell
us whether those campaigns will result in behavior change (not just superstructural change).
The industry to which Bernard refers social change and development depends on
successful reports in order to receive future funding of proj ects from donors. This creates a bias
toward reporting things that are easily measured like how widely an educational intervention
was implemented rather than things like how successfully an intervention changed behavior.
Fitzgerald et al. (1999) administered an AIDS educational intervention to Namibian youth
and then evaluated knowledge, attitudes, intentions, and risk behaviors of control versus
intervention youth. In their own words, the study "demonstrates significant changes in
knowledge, attitudes and intentions regarding HIV risk activities and marginally significant
changes in risk behaviors" (Fitzgerald et al. 1999:60; see also McCombie et al. 2002 for an
example of significant changes in knowledge with marginal changes in behavior). All of those
changed attitudes and intentions didn't translate into significantly changed behaviors, but the
educational intervention was considered a success, as evidenced by its publication in a peer-
Apparently, the criterion for success in such proj ects is to produce better-informed HIV-
positive people. And that is exactly what Desgrees du LoG (1999) found out. A group of World
Health Organization (WHO) surveys conducted in 1995 in 15 developing countries found that
public awareness of AIDS is closely linked to media and educational access. Somewhat
paradoxically, sexual behavior was often found to be most risky among the best-informed
individuals, suggesting that sexual behavior may be determined more by circumstantial factors
than by rational health concerns (Desgrees du LoG 1999). It could also be that people who are
better informed think that they are practicing safe sex when they are not.
The failure of education alone has been acknowledged by some preeminent agencies in the
fight against AIDS. The Joint United Nations Programme on HIV/AIDS (UNAIDS) has stated
succinctly, "research has proven numerous times that education alone is not sufficient to induce
behavioral change among most individuals" (1999:5). Fineberg concurs, "educational efforts to
date have succeeded more in raising awareness and knowledge about AIDS than in producing
sufficient changes in behavior" (1988:592). Fineberg cites specific obstacles to effective
education for AIDS prevention, such as the biological basis and social complexity of behaviors
to be changed and "dual messages of reassurance and alarm from responsible officials"
(1988:592). Maibach and Cotton (1995) contend that a lack of skills is one factor that impedes
behavior change. And Agha suggests (based on research in Zambia) that programs that teach
women to negotiate safer sex may not work as well as with men, "as they are likely to work only
when women have control over decision-making" (1998:36). Lack of control is a significant
structural barrier for women that can subvert any potential positive effects of safe sex education
(Hoffman et al. 2004).
Botswana is a sobering example of the limited power of education alone. In 1987, the
national Ministry of Health and WHO j ointly launched a campaign for public education,
awareness, and prevention of AIDS, although a 1984 survey had found no cases of HIV
seropositivity among the rural population. The campaign lost steam by 1989, since few had
experienced AIDS personally; with no circumstantial motivation, education had limited effect.
Despite the fact that 80% of Botswanans had heard of AIDS after the campaign, only 45%
reported that they had changed their sexual behavior because of it.9 And of those 80% who had
heard of AIDS, fully one-fifth knew nothing about the disease beyond its name (Ingstad 1990).
Today, Botswana has one of the highest rates of HIV in the world. If the government' s
educational intervention had coincided with circumstantial motivation of some kind, the
campaign may not have ended prematurely and the outcome may have been very different.
The correlation between AIDS education level and HIV prevalence is not limited to
Botswana. Caldwell (2000) reports that despite evidence of thorough educational campaigns
(98% of Tanzanian men and 99% of Zambian and Kenyan men surveyed knew very simple
basics about AIDS), AIDS is still out of control in sub-Saharan Africa. "A decade ago it was
believed that such knowledge should be sufficient to contain the epidemic," Caldwell writes. "In
this sense the educational approach has failed" (2000:122).
The educational model of social change has failed outside of AIDS prevention, as well.
The ultraviolet (UV) index was developed to describe how likely it is that a person can receive
9 It is hard to say how many more of that 80% would have changed their behavior if they saw people that they knew
dying from AIDS. It is safe to assume, however, that the educational campaign would have had far greater (and
longer-lasting) effect in today's context, provided that structural barriers to adopting interventions were not present.
sun damage to the skin on a given day. In response, it is hoped that individuals will take
protective actions on days when the UV index is high. In Western Australia, the UV index is
reported daily and awareness of the index is heavily promoted. However, Blunden et al. (2004)
found that although 90% of respondents in Perth had heard of the UV index, only 5% had
noticed the UV index forecast for that particular day. The authors concluded that a campaign to
increase awareness of the UV index was unwarranted, but rather a campaign to change behavior
might have more effect. However, Hill et al. (2002) reported that just such a program promoted
throughout Australia produced inconclusive evidence that changes in knowledge, attitudes, and
self-protecting behaviors were attributable to the education of the SunSmart campaign.
This was also found to be the case in a study of Ugandan adolescents and their behaviors,
motivations, and perceptions of risk for pregnancy and disease in the context of sexual behavior.
"The most important findings to emerge", write the authors, "are that knowledge of safe-sex
behavior and reported behavior have little in common and that the fundamental barriers to
behavioral change lie within the economic and sociocultural context that molds the sexual
politics of youth" (Hulton et al. 2000:35). In other words, structural barriers impede the
successful implementation of superstructural interventions (safe-sex knowledge campaigns).
They continue, "Young males' lack of responsibility for the outcomes of their behavior is
identified as an important barrier to improved sexual health" (2000:35). Unless and until
circumstances compel young men to be responsible for the outcomes of their behavior
(pregnancy, for example), it is unlikely that they will change them due to an information
Touchette (1985) found that although 90% of Bangladeshi women in a rehydration therapy
program learned the salient points of using oral rehydration salts (ORS), only 8% of them
actually used ORS in treating diarrhea. Structural barriers to more widespread adoption of ORS
included lack of access to a reliable source of ORS and disincentives such as the vomiting of
ORS shortly after administration. Again, education as a mono-faceted intervention proved
insufficient to engender behavior change. Smith (2002:332) concludes, "Communication works,
but not if it attempts to substitute for needed structural changes."
Successful Education: A Multi-Faceted Paradigm
Education does have its merits, though. There are cases where education has helped
facilitate positive changes. Smith (2002) argues that to discredit education altogether on the
basis of poorly executed campaigns would be analogous to discrediting surgery altogether if
amateurs regularly practiced it for non-surgical illnesses (diabetes, food poisoning, etc.).
"Clearly," he writes, "in the practice of prevention, communication, and behavior change, skilled
practitioners, adequate resources, and appropriate techniques are necessary in order to succeed"
(Smith 2002:333). What are the conditions under which education can make a difference?
Published examples can give us a clue.
Veverka et al. (2003) studied the impact of targeted information (delivered via the Internet)
in encouraging U.S. Air Force enlisted men to adopt positive diet and exercise behaviors.
Outcomes were assessed by fitness and health measures. The authors concluded that health
measures related to diet (weight, body mass index [BMI], percent body fat) indicate that
nutritional behaviors improved, but that exercise behaviors did not. They attribute the failure of
the targeted exercise information to insufficient intensity of resultant workouts.
Before analyzing the results of the study, some methodological issues must be addressed.
Why were outcomes measured by fitness and health changes? The study title ("Use of the stages
of change model in improving nutrition and exercise habits in enlisted Air Force men") indicated
intent to assess behavior change (nutrition and exercise habits), not overall fitness and health.
But outcome measures referred to fitness and health. Perhaps this was considered a more
reliable indicator of actual behavior change, as one can lie about improved exercise habits, but
it' s impossible to hide body fat.
Methodological issues aside, two interesting points emerge from this study. The first is
that specifieally-targeted information has the potential to impact behavior, perhaps to a greater
extent than general information (cf. Rogers 2003; Backer et al. 1992). This is underscored by
Worden and Flynn (2002) who used targeted messages to reduce teen smoking rates by 35%, by
Palmgreen et al. (2002) who did the same for marijuana use, and Hornik et al. (2002) who
analyzed the effectiveness of targeted information to induce behavior change for vaccinations,
use of oral rehydration salts for diarrhea, intake of vitamins, promotion of breastfeeding, and
promotion of birth spacing in eight different countries around the world. Palmgreen et al.
summarize it thus:
It should not be broadly concluded that "televised anti-drug PSAs [public service
announcements] produce behavior change," or that "PSAs alone are sufficient for
prevention purposes." The data do indicate that PSAs can affect drug behavior, but only in
the context of carefully targeted campaigns that achieve high levels of reach and
frequency, and with messages designed specifically for the target audience on the basis of
social scientific theory and formative research. [2002:52; see also Backer et al. 1992]
The second point is that (as Veverka et al. 2003 concede, cf. Rogers 2003) the achievement
of a specific overall outcome (such as fitness level) depends on the extent and complexity of the
required behavior change. As they discovered, dietary health outcomes can be significantly
impacted by relatively small behavior changes (regularly choosing chicken instead of steak for
lunch can help you lose five pounds relatively easily), whereas fitness health outcomes require
significantly more difficult behavior changes in order to be impacted (measurable cardiovascular
improvements require sustained strenuous exercise over a long period of time).
This suggests that the EMSC can more easily impact outcomes involving relatively small
(easy) behavior changes than outcomes requiring large (more difficult) behavior changes (Backer
et al. 1992). Even a behavior with deeper structural implications (such as improvement of
health) can be affected by educational input if the behavior is relatively easy and painless to
implement. An educational campaign to wear sun block on the beach could be expected to be
more successful than a campaign (on the same superstructural level) to promote vigorous
exercise for 30-45 minutes five times a week.
Aside from targeted information and promoting small, easy changes, what other conditions
can contribute to the success of educational interventions? Allen and Heald write, "Human
behavior rarely changes because of health education alone. Change is facilitated when
information is linked to procedures of compliance" (2004: 1152). Snyder and Hamilton (2002)
concur; in their meta-analysis of multiple individual case studies, they found that overall, 9%
more people performed the promoted behavior after the campaign than before. When they
controlled for campaigns using and not using enforcement, however, they found that campaigns
without enforcement induced only a 5% behavior change, whereas campaigns with enforcement
induced a 17% behavior change. Authority decisions (as Rogers  calls them) generate the
fastest rate of adoption of new behaviors.
One concrete example of the effectiveness of information linked with enforcement is a
study of seat belt use in North Carolina. Despite knowledge about seat belt laws in North
Carolina, seat belt use in the state was only 64% at baseline. After a campaign that combined
information on the importance of seat belt use with rigid law enforcement (seat belt checkpoints,
roving seat belt patrols, and high fines for nonuse), seat belt use in the state rose to 81%
(Williams et al. 2002). "Follow up telephone surveys of the observed nonusers in North
Carolina indicated that many said they would respond to driver' s license points, but not higher
fines," report Williams, Wells, and Reinfurt (2002:95). To obtain further results, more intense
procedures of compliance would be needed not further education, as nonusers surveyed by
telephone already knew about the laws. And it can be expected that any reduction in
enforcement would be accompanied by a simultaneous reduction in compliance.
One of the principles Austin (1995) recommends for effective health campaigns with
young audiences is to send consistent messages from a variety of sources for a long period of
time. Backer et al. (1992) add that evaluating a program while it is in progress increases the
likelihood of success. Koblin et al. (2004) found that intense, frequent education with good
oversight and follow-up can produce at least a temporary positive effect in behavior change. Not
only was the health outcome (reduction in HIV acquisition in homosexual men) achieved
through targeted education (18.2% fewer infections in intervention than control group), but
specific targeted sexual behaviors also changed (20.5% lower frequency of anal intercourse with
unknown-status partners in intervention than control group), which likely influenced the ultimate
outcome. Nevertheless, the effects were found to be strongest only during the 12-18 month
period following the intervention. Extended oversight, however, (in the form of accountability
partners or support groups throughout the duration of an intervention) can sustain campaign
effects in people who quit smoking (McAlister et al. 1989).
Thus, even though targeted education campaigns can induce behavior change, the change
does not appear to be permanent. This is confirmed by a study on vasectomy promotion for birth
control in Brazil (Kincaid et al. 2002). During a media blitz promoting vasectomy to Brazilian
men, calls and visits to clinics performing the surgery dramatically increased over pre-campaign
levels. But for many clinics in the six month period following the campaign, call and visit levels
were at or below pre-campaign levels (although others reported higher levels). These mixed
results led the authors to conclude:
This pattern of an initial increase in demand followed by a drop to a plateau higher than the
original level is common in mass media promotions. When postcampaign levels drop
below the original level, researchers usually presume that preexisting demand "bunched
up" during the promotion period, and no net increase in performance will occur over the
long term. [Kincaid et al. 2002: 184]
Another way that the EMSC can change behavior is through "brand theory" (Bernard
2002). This theory says that education (advertising) can influence whether I buy a Toyota Land
Cruiser or a Jeep, provided I was already intending to buy a four-wheel drive vehicle in the first
place. In that sense, education (a superstructural intervention) can influence my brand selection
(a superstructural choice). Hornik (1989:329) agrees that media information alone can influence
behavior if it promotes obj ectives "reflecting the material possibilities for change." But no
amount of educational rhetoric will influence me to buy a car in the first place if I have no
income (a structural constraint) or to give up driving if the nearest affordable housing to my j ob
is 30 miles away and there is no public transportation system (an infrastructural constraint)
(Bernard 2002). Although they do not use the term "brand theory", Flora et al. (1989) concur
that behavior change which depends primarily on an increase in knowledge (a superstructural
choice) can be affected by mass media alone (a superstructural intervention), whereas change
that aims to affect a deeper level (structural, infrastructural) must combine mass media with
considerable interventions at the structural level.
One example of the brand theory in action is a report by Soumerai et al. (2002) on the
effect of publicity in reducing aspirin use for childhood fevers. When a link was found between
aspirin use in children and the development of Reye' s syndrome an often fatal illness both
academic journal and mass media reports in the United States disseminated that information.
This led to federally-mandated warning labels on aspirin bottles and a subsequent drop in both
childhood aspirin use and the incidence of Reye's syndrome. Because other brands of equally-
effective fever-reducers were already on the market, a superstructural brand switch was
comparatively easy to make in response to superstructural information. According to brand
theory, abandonment of aspirin for childhood fevers would have been much more difficult to
achieve in the absence of any good alternatives.
Education can potentially influence behavior if prior motivation for change already exists
(Backer et al. 1992). "Effective messages", write Piotrow et al. (1997:36), "build on people's
current thoughts, feelings, and needs and do not disregard or contradict them." Provided an
individual was intending to make a large behavior change in the first place, education can
persuade someone one way or another in the implementation of that change. If someone was
planning to spend an expensive vacation in Florida to begin with, an infomercial might sway that
person to stay in one resort over another. But a person without the intention (or finances) to
vacation in Florida in the first place will probably not be moved to book a reservation based on
One example of the EMSC successfully influencing behavior is a hygiene promotion
program in Burkina Faso (Curtis et al. 2001). The program was promoted for three years in
Bobo-Dioulasso in an effort to change behaviors that spread diarrhea. The program was tailored
to local customs using local communication channels. It also targeted specific behaviors. But
most importantly, it built on people's existing motivation for hygiene; that is, it didn't have to
create motivation for change in the first place, but simply had to influence that motivation in the
right direction through education (the "brand theory"). Rogers (2003) calls this motivation
"recognizing the relative advantage" of a new behavior, which speeds up its adoption.
Success was measured by proportion of targeted mothers who had been reached (three-
quarters), proportion that could properly cite two main messages of the campaign (one-half), and
- most importantly actual behavior change: hand-washing with soap after cleaning a child's
bottom (an increase from 13% to 31%) and hand-washing after mothers themselves used the
latrine (an increase from 1% to 17%). One targeted behavior change that did not occur to a large
degree was the safe disposal of children' s stools (an increase of only four percentage points from
80% to 84%) (Curtis et al. 2001). According to brand theory, this might be explained by a lack
of initial motivation for safer disposal of children' s stools.
In a step-by-step guide to developing health behavior change campaigns, Witte et al.
(2001) contrasted early reports on the effectiveness of fear appeals (nearly all showed strong
evidence of overwhelming efficacy) with later rej section of the method by health educators
(nearly all balked at using it; see Geller 1989, for example). Upon discovery that earlier
reporting had selectively favored successful outcomes rather than reporting all results the
EPPM (Extended Parallel Process Model) was developed to explain in which cases fear appeals
work, and in which ones they do not.
The first stage of the EPPM is a cognitive threat appraisal; when faced with information in
a fear campaign, people first try to determine if they are at risk. If they feel that they are not,
then they disregard the fear campaign. By contrast, if they feel that they are at risk, then they
become fearful and the second stage of the EPPM kicks in: appraisal of the efficacy of the
recommended response. If people determine that the recommended response is possible to
implement and truly does reduce the risk they feel threatened by, then they take action to protect
themselves (danger control response). If, however, people determine that the recommended
response is impossible to implement (due to structural constraints) or does not impact the risk
they feel threatened by, then they take action to control their fear (denial, avoidance,
development of conspiracy theories), as danger control appears useless. Thus, fear appeals can
be successful if people correctly perceive their risk and perceive efficacy in controlling the
danger and attendant anxiety (i.e. do not feel hampered by structural or infrastructural barriers to
mitigating their risk) (Backer et al. 1992).
Altering one's environment, providing access to resources, and removing or circumventing
obstacles is a necessary prerequisite for successful behavior change (Maibach and Cotton 1995;
Valente 2002; Backer et al. 1992; Adhikarya 1989; Reardon 1989). Valente calls this the
ecological perspective, and argues that "these [environmental] contexts are often beyond
personal control, yet they influence behavior. The ecological perspective shifts the emphasis
from individuals to a broader understanding of the environmental and social context of behavior
and in so doing sometimes identifies barriers to behavior change" (2002:45). Only the removal
of a government hostile to capitalism allowed private enterprise to flourish in the former Soviet
Union, though Soviet citizens were knowledgeable about capitalism long before the fall of
communism. Similarly, if disincentives to change are replaced or overridden by large enough
incentives, then that specific kind of barrier to change might be breached (Reardon 1989). Many
people see disincentives in the use of condoms (reduced pleasure, implication of mistrust in a
relationship, loss of clientele for sex workers), but some of those same people would use
condoms if they were paid enough to do so each time they had sex. This would not mean that
barriers to condom use didn't exist anymore, but the incentives would supersede the
disincentives. This is part of the method used by the Chinese government (tax incentives, child
allowances, education, and j ob placement) to encourage families to limit themselves to one child
only (Rice 1989). Realizing this principle, "some donors and other institutions are emphasizing
long-term structural changes intended to increase motivation for smaller families, especially
more education, employment, and other opportunities for women," write Piotrow et al.
Valente (2002) identified eight different health intervention strategies, some of which aim
to increase knowledge (such as training local health providers, who in tumn provide health care to
the community) and others that aim to change behavior directly. One of the latter types of
strategies is community-based distribution or outreach, which combines social learning theory
(Bandura 1977) which argues that people leamn by copying role models with what Rogers
(2003) calls "trialability" the opportunity for potential adopters to try new products or ideas
before committing to behavior change. Community-based distribution or outreach is where
agents of change personally approach members of the target population, identify with their
clientele, and then encourage adoption of a new behavior, often passing out free samples of a
product associated with that behavior (e.g. condoms, makeup products, or cigarettes). This
strategy removes access barriers, not only for the products themselves, but also for contact with
an actual practitioner of the new behavior who can answer adoptees' questions and allay their
fears of change. This strategy has been highly successful in the United States for companies
such as Tupperware and Avon, and has also worked well for drug rehabilitation, when former
users personally recruit their friends to the program that helped them kick the habit (Valente
Such an approach is based upon the diffusion of innovations theory (Rogers 2003; Valente
2002) which describes how new ideas and behaviors spread throughout a community,
categorizes people by how early they adopt a new practice (innovators/pioneers, early adopters,
the early maj ority, the late maj ority, and the laggards), and elucidates the stages individuals
undergo in the process of adoption (knowledge, persuasion, decision, implementation/trial, and
confirmation/adoption). Diffusion of innovations theory also describes what is known as the
KAP-gap (KAP stands for Knowledge, Attitudes, and Practice). "The percentage of the
population aware of the innovation increases most rapidly," writes Valente (2002:38). "It can
take somewhat longer for that same percentage to have a positive attitude, and still longer for the
same percentage to adopt it." The KAP-gap partially explains why education by itself may
increase knowledge, but not necessarily change behavior. To narrow the KAP-gap requires
removal of structural inhibitions to adoption, as community-based distribution or outreach
attempts to do. In addition, if the results of adopting a new behavior are readily visible to
onlookers (Rogers 2003) and the early adopters are similar to their peers in beliefs, education,
socioeconomic status, et cetera (known as "homophily") (Rogers 2003; Backer et al. 1992), then
the innovation is more likely to spread quickly.
Finally, a meta-analysis (Snyder and Hamilton 2002) was conducted on 48 health behavior
campaigns, in order to generalize the results of multiple individual case studies. In addition to
reinforcing the effectiveness of enforcement (combined with dissemination of information, as
reported above) on behavior change compliance, the meta-analysis concludes that inclusion of a
message that contains new information increases campaign effect size. In addition, the authors
found that the wider the reach of a campaign, the larger the effect size.
Two of the most famous health behavior change campaigns that are repeatedly cited in the
literature for their effectiveness are the North Karelia (Finland) and the Stanford (California)
heart disease risk factor reduction programs. Though they predate most of the studies cited in
this section (both started in the early 1970s), they are efficient summaries of the information
presented above because of their comprehensiveness. In the highly successful North Karelia
Project (as reported by Backer et al. 1992), residents petitioned the government public health
agency to assist them in developing strategies to reduce their heart disease risk; prior motivation
was already present. Apart from a mass media campaign promoting the adoption of healthier
behaviors (the educational component), other components of this seminal proj ect included
community participation and management, inclusion of a prestigious agency that could serve as a
role model, integration of the proj ect into the existing health service structure of the community,
widespread promotion of the intervention, community-based outreach (using homophilous peers
and opinion leaders), long-term investment in the proj ect (with continuous oversight and follow-
up for over 20 years), targeted information, and, perhaps most importantly, the removal of
structural barriers to implementation (reliably stocking supermarkets with low-fat foods, for
example). Similarly, the Stanford Heart Disease Prevention Program utilized not only mass
media promotion, but also the networking of opinion leaders, bilingual approaches (tailoring the
campaign to the audience), targeted information, incentives, and ongoing oversight and
evaluation of the project (Flora et al. 1989; Backer et al. 1992).
Education for AIDS prevention must be multi-faceted and combined with strategies that
eliminate structural and infrastructural barriers to successful implementation. Many developing
nations and the NGO's toiling within their borders are dedicated to ramping up the infrastructure
and improving structures, as well. But must successful AIDS prevention wait for the structure
and infrastructure to support it? What can one do in the meantime, while waiting for structural
and infrastructural changes to occur? The answer is to identify indigenous practices and
movements that are protective against AIDS and reinforce them. If an approach is homegrown,
it is safe to assume that motivation for change is already present. And it is also likely that the
approach must not be hampered by structural or infrastructural constraints, or else it wouldn't
have arisen in the first place. If that energy can be channeled and supported, education may
make a difference where it hasn't before.
Many indigenous practices in West Africa do have protective effects for a variety of
diseases. These practices have been documented and published for a wider audience, and many
important principles can be drawn from the studies reporting them. The following chapter
investigates ways that West Africans have developed homegrown approaches to public health in
the absence of outside intervention.
THE ANTHROPOLOGY OF DISEASE INT WEST AFRICA
Cultural evolution is an idea that never really died out. Popular in the latter part of the
19th century in anthropology, various iterations of the theory posited that all cultures pass
through the same stages of development, so that one can ascertain how advanced a society is by
looking for the presence or absence of certain traits or cultural markers. Early theorists in
anthropology like Edward B. Tylor and Lewis Henry Morgan used the theory of unilineal
cultural evolution to explain differences in customs, family structure, and material artifacts
across societies. The most advanced societies were identified as those possessing the most
cultural traits that were used as markers. Perhaps not surprisingly, the traits chosen as markers
inevitably pointed to Western civilization as the most advanced society.
A great deal of thought and research went into the development of unilineal evolutionary
theories. However, as time went on and more ethnographic data were collected, it became
increasingly apparent that unilineal cultural evolution was a weak explanation for cultural
At the turn of the 20th century, Franz Boas challenged the notion of cultural evolution with
his dual emphases on historical particularism and cultural relativity. Insisting that
anthropological theorists base their ideas on voluminous data collection instead of speculation
and opinion, Boas argued that societies and their cultural traits are products of their particular
history and circumstances. Each society has its own unique background, and this determines
how and which traits each will develop. Because each society can only be fairly assessed on its
own terms and in reference to its own particular background, one cannot judge the advancement
of a society in reference to one's own (cultural relativism), and one should not expect that
customs and adaptations developed in one society will necessarily prove adaptive and useful in
To a large degree, cultural relativism has prevailed over cultural evolution in contemporary
anthropology. And yet, cultural evolution is an idea that has been repeatedly revived since the
time of Boas in various new forms. Some of the practical problems of a cultural evolutionary
orientation include appearing politically incorrect (it's hard to be diplomatic when one is
basically saying that another society is not as progressive as one's own) and denying the
ingenuity of indigenous adaptations to particular circumstances and environmental stressors. For
example, a cultural evolutionist might consider a revolver to be a superior example of a gun over
a blow-dart. After all, the proj ectile flies faster, is more destructive upon impact, and can be
accurate over great distances. However, if one's primary use for a gun is to hunt small birds and
frogs in a forest, a blow-dart may prove more adaptive, as it is relatively silent, accurate at ranges
close enough to see small animals, and does not obliterate one's prey upon impact. The
ethnographic record is replete with similar examples where a trait initially appearing maladaptive
or primitive is actually more advanced than originally thought when placed in historical and
environmental context (for more examples, visit the Human Relations Area Files online at
http://ets.umdl .umich.edu/e/ehrafe/ and search using OCM subj ect code 178" for cultural
The same is true when assessing public health cross-culturally. Are Western models and
approaches to disease prevention always better than indigenous ones? Given different
environmental and historical contexts, can one make such a judgment obj ectively? Should
anthropologists, public health practitioners, and outside experts trust indigenous wisdom?
Cultural relativity would seem to suggest so. "Let us go beyond the ethnocentric assumption,"
writes Green (1999b,:224), "that any health beliefs and practices that differ in substance and in
idiom from ours is [sic] ipso facto deficient and something that stands [sic] in the way of
In studying the health of cultural others, it is instructive for anthropologists to consider not
only the effects of Western interventions on wellbeing, but also the effects of indigenous
approaches to public health. Which cultural taboos have beneficial effects on wellbeing? What
is the role of religious prohibitions on the health of a population? Which behavior modifications
positively impact the health of a society? Do some practices produce better health outcomes
though better health was not their original intent?
This chapter focuses on how West Africans have developed strategies (outside the
biomedical paradigm) to avoid numerous kinds of diseases, and have coped and adapted
culturally to mitigate these diseases' impacts on society, productivity, and public health. The
focus is exclusively on disease prevention (and not treatment), and cases analyzed are largely
(but not exclusively) limited to West Africa. The studies analyzed below demonstrate that the
practice of indigenous disease prevention is widespread, multi-faceted, and has the potential to
be quite effective in the absence of Western medical interventions.
Inhorn and Brown (1990) have written an extensive review of the anthropology of
infectious disease, highlighting many cases around the world where simple, indigenous strategies
of disease avoidance resulted in significantly positive public health outcomes. As an example,
they cite a report from J. M. May's 1958 book, The Ecology ofHuman Disea~se, where it was
observed that native hill peoples in North Vietnam had adjusted to the presence of malaria
vectors in the area by building homes on stilts over ten feet high the flight ceiling of the
But, as Inhorn and Brown point out, not all behaviors that prevent disease are deliberate.
They cite examples of malaria prevention that happened quite circumstantially, as in the use of
alkaline laundry soaps that destroy mosquito breeding sites and the timing of shepherds in
Sardinia moving their flocks to higher elevations for summer grazing placing the shepherds out
of harm's way during peak malaria months. Similarly, Inhorn and Brown cite an unpublished
manuscript by Linda Collier Jackson which hypothesizes that regular cassava consumption in
Liberia might limit parasite development in humans because it contains minute amounts of
cyanide (Inhorn and Brown 1990). And conventional wisdom in northwestern Benin holds that
the consumption of pima (a small, very spicy red pepper used in abundance for cooking) is
protective against memingitis.
Even if behaviors are deliberate, however, the rationales behind them are not necessarily
biomedical. "These and other traditional behaviors", Inhorn and Brown note in reference to
malaria prevention, "based on the folk theory of miasma probably had preventive effects"
(1990: 101). This counteracts "the entrenched belief in the biomedical community that
indigenous beliefs and practices are irrelevant to the problem at hand" (1990: 104).
The effect of cultural taboos upon public health can be great, both positively and
negatively. Tab oos that have an impact on health functi on primarily as prevent on-by-avoi dance
- that is, distancing oneself from sources of disease transmission, either knowingly or
unwittingly. Three studies from Nigeria shed light on the power of cultural taboos to promote
health in the West African context.
The first study (Olusanya 1969) investigates attitudes toward birth control among the
Yoruba. Children are not a disease to be prevented, per se, but family planning (namely, spacing
children and limiting family size) has long been recognized in the West as a way to increase the
overall health of each family member and to make sure that resources are allocated sufficiently.
The Yoruba have long practiced birth control, too, but it is "traditionally not regarded as a means
of achieving a pre-meditated size of family or the spacing of births for economic reasons, but is
closely linked with their belief system" (Olusanya 1969: 14). Regardless of the reasons for
practicing birth control in the first place, the positive effects of limited and controlled family size
upon public health can still be enjoyed.
Olusanya mentions three indigenous methods of preventing conception: a charm (either a
ring on the finger or a belt stuffed with charms) worn by a woman during sexual intercourse,
drinking very salty water immediately after intercourse (on the woman's part), or coitus
interruptus. More common than prevention of conception are the traditional methods of
abortion. These consist either of drinking mixtures of different substances (potash, lime juice,
cactus juice, fresh eggs, gin) in various combinations, depending on the progression of the
pregnancy, or the insertion of a vaginal suppository (made from locally-grown seeds and leaves)
that destroys the fetus (Olusanya 1969).
But according to Olusanya, "traditionally, abstinence is the main method of birth control; it
stems from the lactation taboo and has the effect of spacing out successive children born by a
woman and consequently controlling the ultimate size of family" (1969: 14-15). "Lactation
taboo" refers to the belief that a nursing child will fall ill and die if the mother' s milk is made
harmful by sexual intercourse. It is considered better to wait until the child is old enough not to
depend on its mother' s milk anymore before resuming sexual intercourse, thereby lowering the
lifetime number of pregnancies. Thus, this taboo is specifically geared toward promoting the
health of a specific individual the nursing child and this ultimately affects the community as
a whole, as children from any given family are spaced out more manageably.
Hughes (1976) also explores the role of cultural taboos for promoting public health in
Nigeria. He discusses indigenous health practices among the Egba Yoruba, including disease
prevention methods. Contagion of some sort is recognized as being the causative agent in
smallpox, as well as in certain mental disorders. Consequently, the practice of quarantining an
infected individual is well known among the Yoruba. The cultural taboo which isolates a sick
person from human contact extends to contaminated obj ects, as well. Members of healing cults
who have had smallpox themselves in the past (and are therefore immune) are appointed to
remove the sick person's clothing and personal belongings from general contact with the rest of
society (Hughes 1976).
This emphasis on isolation of ill individuals is also highlighted in Oladepo and Sridhar's
1987 discourse on public health practices in Nigeria. The authors examined traditional health
practices of the Yorubas, Hausas, and Ibos in Nigeria through interviewing about four hundred
persons in each ethnic group. They found that isolation is advocated for tuberculosis and
cholera, and since those diseases are considered to be hereditary, a taboo against marriage
between infected and uninfected families exists, as well. Thus, the isolation can extend beyond
the individual and their immediate circumstances (Oladepo and Sridhar 1987). Isolation is
practiced in Mozambique in the context of sexually-transmitted diseases, as well. Contaminated
individuals are kept apart from other people until they have been ritually purified, which may
also involve herbal medicines (Green 1999a). "Perhaps the isolation, avoidance, or social
marginalization of people in polluted states serves to quarantine those who, in fact, could be a
health threat to others because of their contagiousness" (Green 1999b:14).
Oladepo and Sridhar (1987) also mentioned taboos regarding water sanitation; respondents
advocated not drinking from water sources that are used for sacrifice, children not being allowed
to collect rain water with their bare hands, menstruating women not being permitted to fetch
water from the village stream, avoiding submerging wounds in a river, and avoiding collecting
water simultaneously in large groups. From these examples, it appears that contamination of the
water source is a maj or concern for many Nigerians, and these cultural taboos are geared toward
protecting those sources to promote the general health of the public.
Regarding the sanitary disposal of human waste, Oladepo and Sridhar mention a taboo
against defecating upon other people's feces, which probably limits the collection of large
amounts of sewage in one place (a potential public health hazard). It is believed that unsanitary
conditions attract snakes and scorpions, and can cause leprosy. They also mention the
prohibition amongst Muslims against using the right hand during anal cleaning with water
(1987). This limits fecal-oral cycles since Muslims eat with the right hand. Oladepo and Sridhar
also mention taboos whose health significance is not immediately apparent to a Western
layperson: a pregnant woman cannot use a pit latrine or she'll become barren, and if a child
urinates in the mother' s food while she is eating it, she must not abandon it or the child will die
They next consider taboos relating to food handling, preparation and storage. For
example, nobody should eat food that is stored under the bed of a female, and a woman should
not cook or eat while braiding her hair. Pregnant women should avoid certain foods, such as
snails, bush meat, and certain vegetables, but after delivery, they should eat okro and melon to
improve digestion (Oladepo and Sridhar 1987).
These taboos are communicated and reinforced through parents educating their children on
moonlit nights, during crises or community festivals, and during visits to traditional healers
(Oladepo and Sridhar 1987). Thus, cultural taboos are numerous in the West African context,
and their knowledge is widespread. Their potential impact upon public health and disease
prevention is great, and should not be overlooked by anthropologists attempting to understand
how West Africans promote good health outside the biomedical context that so many of us are
familiar and comfortable with.
Behavioral Prevention Factors
Behavioral responses to disease threats are a common method of disease prevention. Once
risk factors are identified in disease transmission, people tend to engage in behaviors that
minimize their association with those risk factors. It' s the law of common sense, if you will, and
it is practiced to some degree by virtually all people worldwide in some form.
MacCormack (1984) relates how human behavior has facilitated the spread and
intensification of malaria in sub-Saharan Africa, as inhabitants of sparsely-populated areas
settled into ever larger groups adj acent to bodies of water. Cohen and Armelagos (1984),
summarizing cross-regional comparisons, concur that infection, malnutrition, and episodic stress
all increased and mean age at death decreased among populations that transitioned from hunting
and gathering to agriculture. Caldwell and Caldwell (2003), however, argue that the transition
was too slow to produce drastic changes in health status and claim that the evidence for Cohen
and Armelagos' conclusions is selective and sketchy. Nevertheless, MacCormack concludes that
"first, we sometimes modify our environment to gain nutritional or convenience benefits to the
detriment of our disease status; and second, we then adapt to the new disease risk" (1984:81).
Adaptations can come in the form of official government programs administered by medical
experts, or in the form of lay practices, which usually long precede bureaucratic efforts at disease
control. These lay efforts are commonly based on technologies that are accessible to almost
every household a key factor in their sustainability and reproducibility (MacCormack 1984).
First of all, physical barriers to prevent mosquitoes from feeding on humans have been
developed by various peoples across the continent. In many parts of rural Africa, tight-fitting
doors and window shutters are closed as the evening descends. Prior to the colonial period, the
Fulani people were known to sleep under fine-mesh mats that permitted air circulation while
blocking mosquito access to their skin. In swampy coastal Sierra Leone, children who are
weaned sleep on the floor outside their mother' s bed net, and are thus susceptible to mosquito
bites. Consequently, many children completely envelop themselves in thick cotton cloth that is
locally woven. Tests have demonstrated that this cloth is too thick for mosquitoes to penetrate.
Similar practices have been observed in northern Nigeria, where cool nights cause the thick cloth
coverings to render double duty warmth and mosquito blockage (MacCormack 1984).
Repellents are also used, though they are not the kind found emanating from an aerosol
can. Rural south Sudanese mix cow dung ash and cow' s urine and apply it to the skin in the
evening to ward off mosquito bites. Smoky fires can also keep mosquitoes at bay especially
cow dung fires, which are purported to keep even snakes away in Kenya. Cow dung is also used
to plaster the inside walls of houses in an effort to repel mosquitoes (MacCormack 1984).
Hughes (1976) highlighted not only cultural taboos, but also addressed behavioral disease
prevention factors. He discussed healing cults among the Yoruba, specifically in regard to the
smallpox god, Shopona. These cults developed indigenous practices of vaccination, which have
been practiced for a long time both by these groups and others throughout West Africa.
Unfortunately, Hughes did not go into further detail to describe what type of vaccinations were
implemented (subcutaneous, oral, or otherwise), but it is apparent nonetheless that concepts of
behavioral interventions to increase immunity and prevent disease are not unknown in West
Hughes also discussed hygienic behaviors practiced among the Yoruba that promote good
health and prevent disease. He wrote that "sanitation in villages is generally adequate, and the
Yoruba themselves live by standards of personal cleanliness which, considering the supply of
water and cleansing materials are rather remarkable" (1976:17). Hughes also mentioned the
common practice throughout West Africa of chewing on a fibrous stick to promote good dental
hygiene, a behavioral preventive measure that sustains the relative health of many people' s teeth
absent the services of Western dental interventions.
Indigenous responses to AIDS and other STD's have recently stolen the limelight of
studies focused on behavioral disease prevention activities. Because AIDS in Africa is most
highly concentrated primarily in the southeast portion of the continent, there are more studies
dealing with AIDS in those contexts than in West Africa. But the lessons learned about
indigenous responses to AIDS can apply elsewhere continent-wide.
In an extensive study of STD's in Mozambique, Green (1999a) discussed the role of
traditional healers in preventing widespread outbreaks of illness. When treating those who came
to them with STD's, some healers reported seeking out and treating recent sexual partners of the
diseased patient, which is itself a preventive measure (considering that those former partners may
also be infected though still asymptomatic and may pass it on to others if the sickness is not
nipped in the bud). Thus, proactive aggressive treatment is a behavior that is used to prevent
disease transmission. In addition, traditional medicines are available that are said to always be
effective for STD prevention if taken just prior to sexual intercourse with an infected individual
(Green 1999a). Thus, traditional medicines are said to have a preventive effect not just a
Green also notes that several other behavioral methods of STD prevention exist in the
traditional medicine paradigm. These include avoiding adultery, avoiding sexual intercourse
with strangers, and avoiding sexual intercourse during menstruation (1999a). These preventive
behaviors focus on disease avoidance, rather than risk reduction, such as the use of preventive
traditional medicines while still exposing oneself to an infected individual.
Stoneburner and Low-Beer (2004) represent only two of literally hundreds of researchers
who have investigated which behavioral disease prevention practices played the biggest role in
explaining Uganda' s steep HIV prevalence decline in the late 1980s and 1990s. The jury seems
to still be out on a definitive answer, as consensus appears to be elusive amongst the various
academic and practitioner players with AIDS research expertise. Nonetheless, Ugandans have
demonstrated several indigenous response strategies regardless of the magnitude each strategy
played (or Westerners' reluctance to acknowledge them) which have unquestionably impacted
the AIDS epidemic.
As cited in Stonebumner and Low-Beer 2004, Ugandans increased the average age of their
sexual debut before the arrival of Western intervention, a strategy which ultimately leads to
fewer lifetime partners and a consequent reduction in risk of exposure to a potentially HIV-
infected individual. Those who had already made their sexual debut appear to have reduced the
number of their casual partnerships, and there seems to have been an increase in condom use. Of
these three, the first two can rightly be viewed as indigenous behavioral responses for the
purpose of disease prevention as they appeared before Westemn interventions really got
underway in Uganda whereas the last is non-indigenous, having been introduced from the
outside. This demonstrates that, even in the face of a new and modernn" disease, Africans have
the ability to generate indigenous responses that can be effective in promoting and maintaining
good public health.
A rise in sexual debut is a form of prolonging one' s abstinence, which fits well into the
African concept of specific abstinence, such as abstinence while pregnant and abstinence while
lactating. There is much disagreement as to how long any of the various types of abstinence can
be maintained or whether they can be practiced at all but abstinence as a mode of disease
avoidance is hardly a new idea; it is simply a long-standing cultural tradition (or taboo, if you
will) that has been recycled in some cases to combat the scourge of AIDS.
In Guinea, abstinence until an accepted age for sexual debut has been seen as a way to
avoid weakness or illness that can spring from starting intercourse too early (Goirgen et al. 1998).
The view was also expressed that if young men started having sex too early, they would use up
their available sperm before siring children. However, young women have been encouraged to
marry early (especially if they are not in school) in order to reduce the Einancial burden of care
upon the parents. This has the effect of causing a significant number of early sexual debuts to be
marriage-related, which reduces the number of casual pre-marital partners a girl is likely to have.
Early marriage, then, is an indirect strategy of reducing a girl's exposure to many partners
(which could be a health risk).
Goirgen et al. (1998) also discuss Guineans' avoidance of great age differences between
partners. Whereas older, wealthier men are considered attractive for purposes of acquiring
Financial support, girls typically do not want an older man for a regular partner, unless he is not
more than a few years older than she:
Young women accept a few years' age difference, but they do not want their partner to be
much older, for fear that such a relationship will endanger their health, destroy their youth
and contribute to early aging. Young men try to avoid contact with older women. They
believe that a relationship with an older woman makes a young man grow old or causes
diseases or even an early death, while it rejuvenates the woman and makes her more
Thus, health is an important consideration for young people in Guinea when choosing a
sexual partner. Spouse selection behavior based upon age differences between partners is
directly tied to perceptions of health risk and disease avoidance. The indigenous strategy of
spouse selection is specifically geared toward promotion of well-being.
Pre-marital pregnancy is considered to be a significant threat to a young woman's well-
being in Guinea. This is discouraged through a system that ostracizes unmarried pregnant young
women, but also demands (through the agency of the girl's father) that the girl's partner
acknowledges his paternity and takes at least some Einancial responsibility for it. This
indigenous response to a threat to well-being encourages behavior change that minimizes risk-
taking for both young men and young women (Goirgen et al. 1998).
The Role of Tangential Practices in Disease Prevention
Finally, it is important to consider the role of practices which may have a public health
outcome, though that may not have been their original intent. A prime example of this is Etkin
and Ross' 1982 report on foods normally consumed by the Hausa of northern Nigeria which
happen to have disease prevention qualities, as well. "While therapeutic value is ascribed by the
Hausa to a wide range of species," Etkin and Ross write, "their consumption of these plants is
not necessarily limited to a medical context" (1982:1560). Instead, plants that have
gastrointestinally therapeutic qualities are often eaten as snacks throughout the day, or as relishes
accompanying grain-based dishes in regular meals that are shared by all in the household. It is
only when these plants are specifically consumed for their medical purposes that they are
restricted to consumption by the ill person alone (Etkin and Ross 1982).
Though these plants are acknowledged medically for their curative qualities, they also have
preventive effects in warding off gastrointestinal distress that is common among the Hausa when
consumed in the course of everyday meals or snacks. For example, tannins and astringents have
the protective effect of coating the alimentary tract, protecting inflamed mucous membranes and
minimizing the absorption of toxic substances. Gum resins and mucilages also create a physical
barrier that prevents damage to the gastrointestinal tract by certain bacteria. Fixed and volatile
oils can stimulate pancreatic secretions and gastric fluids, as well as salivation, generally
promoting smoother digestion. Sulphur-containing volatile oils are antimicrobial and inhibit
fermentation in the stomach, warding off gastric discomfort and bacterial disease. Various other
components of common Hausa food items have mild laxative effects, are effective against
amoebae and protozoa, exhibit antifungal characteristics, inhibit intestinal worms, alleviate
acidity, and restore electrolyte balance (Etkin and Ross 1982).
That these foods and their components do have preventive effects on gastrointestinal
diseases is demonstrated by the fact that, as certain foods become available or unavailable due to
seasonal rainfall fluctuations, specific diseases wax and wane in tandem with the availability of
the specific plant food items that protect against them. "On the other hand," write Etkin and
Ross, "environmental factors important in the transmission and development of the other
intestinal disorders remain relatively constant over the course of the year" (1982:1571). This
demonstrates, then, that plants long praised for their curative qualities can have preventive
effects, as well. In addition, this study demonstrates that tangential practices of West Africans
can have as much of an effect on disease prevention as intentional practices.
West Africans (like many other anthropological "others" around the world) have
established well-developed indigenous systems of disease prevention that are effective,
culturally-acceptable, and easy to implement (since they use readily available resources).
Cultural taboos such as the lactation taboo, isolation taboo, and various water, human waste, and
food taboos can have preventive effects that protect adherents from disease. Behavioral
measures, such as the erection of physical barriers to malaria-carrying mosquitoes, development
of mosquito repellents, indigenous vaccinations, hygienic practices, aggressive STD treatment,
and sexual behavior change also have preventive effects against disease transmission. And
tangential practices whose original intent was not necessarily disease prevention can also
protect individuals from the onset of various diseases, as in the case of Hausa utilization of
medicinal plants for everyday food.
The implications of these examples for AIDS research and practice are great. Western
anthropologists (and other AIDS practitioners) need to seriously consider the possibility that our
research subjects may have come up with indigenous solutions to AIDS prevention that we have
overlooked. They may have developed modalities that are well-suited to their cultural context,
are easy to implement, and are highly effective. We must not limit our inquiry to the effects of
biomedical interventions that were developed in the Western cultural context, far removed from
the reality on the ground in AIDS-ravaged sub-Saharan Africa. If we are to make any inroads
into understanding and defeating this disease, we need to take African stakeholders seriously,
and treat them as partners in prevention efforts.
This literature review of indigenous methods of disease prevention is by no means
exhaustive. But it is sufficient to demonstrate the depth and breadth of indigenous wisdom and
of the potential for non-Western, non-biomedical models of disease prevention to significantly
and positively impact public health. As historical particularism has predicted, West Africans are
quite capable of developing customs, practices, and technologies that fit their historical context
and are adaptive to environmental constraints and possibilities.
RESULTS AND ANALYSIS OF DATA COLLECTED INT THE BOURE
Condoms are at the forefront of many Westerners' minds when they think of AIDS
prevention, even though there is no documented case where condoms have brought down a
population-level AIDS rate. Where condoms have made a noticeable impact on AIDS rates is
among specific, targeted high-risk groups, such as prostitutes in Thailand (Green 2003b) and
male homosexuals in the United States and Europe. Thus, condoms deserve consideration for
the role that they may have played in keeping AIDS infection rates low thus far among the high-
risk migrant mining communities of Guinea. In order for condoms to have played a maj or
contributing role in this stasis, at least three conditions must have been met: (1) condoms must be
available in mining communities in sufficient quantity to be used in most high-risk sexual
encounters; (2) understanding of the proper use of condoms must be high among both men and
women; and (3) women must be able to enforce proper condom use by their male sexual
Undoubtedly there are additional factors for successful AIDS risk reduction using
condoms, but these three are particularly significant because a default in any one of them would
severely reduce the effectiveness of the other two. For example, if there was insufficient
availability to meet demand, that would undercut the condom explanation regardless of whether
or not understanding of proper use was strong or enforceability by women was high.
Availability considers issues such as distribution (are large stashes of condoms locked in a
warehouse?), convenience (can people find condoms when they need them?), and price
accessibility. Knowledge of proper condom use is also critically important because the best
estimates place condom efficacy for AIDS prevention at 90% when used correctly (figures range
from 80% to 90%) (Weller and Davis 2002; Hearst and Chen 2004; Davis and Weller 1999).
That means that one out of every ten sexual encounters using a condom will result in a
compromised condom which exposes one of the partners to HIV if the other is infected, even if
the condom is used properly. If condoms are used improperly, effectiveness plummets.
Depending on the circumstances, condom effectiveness could drop to zero percent in a given
encounter, but even a ten percent failure rate is cause for concern when dealing with a disease
that is, at present, incurable.
To test the three principle factors (availability, knowledge of use, and women's agency),
460 randomly selected respondents were interviewed in the villages of Fatoya, Boukariah,
Balato, and Kintinian (collectively called the Boure) near Siguiri, Upper Guinea. In addition, 16
health centers and pharmacies were visited to gain an understanding of the role they played in
assisting people in their pursuit of protection from HIV/AIDS. Some of the best methods of
random sampling (using a random number generator with a census list, drawing names from a
hat/bowl) were impossible to use in the Boure due to its remote location and poor infrastructure
and the transient nature of respondents interviewed. A variation of the space sampling method
(Lang et al. 2004; Bernard 2000; Duranleau 1999; Handwerker 1993, 1999; and Mukhopadhyay
1999) proved most appropriate; research assistants independently walked in a randomly chosen
straight line toward the edge of the village, interviewing every nth person encountered, whether
walking in the street, cooking in the courtyard, or sitting at a vendor' s stand (where n is a
randomly chosen integer between 5 and 20 drawn from a hat each day). When they reached the
edge of the village, assistants chose a new traj ectory and continued in the same manner,
crisscrossing the village at all angles and directions. This produced a sample ranging in age from
18-62 years (mean = 32.8); 49.8% female, 50.2% male; 81.3% married, 16.5% single, 0.652%
widowed (3 respondents out of 460), 0.217% divorced (one respondent), with six respondents
not reporting marital status at all; education ranging from 0 years of schooling to 16 years (mean
= 2.30); and respondents hailing from as close as the Boure itself to as far away as Abidjan, C8te
d'Ivoire with professions as diverse as mining, housekeeping, teaching, truck driving,
governmental administrating, blacksmithing, tailoring, farming, traditional healing,
photographing, hairdressing, and artistic painting.
For the prefecture of Siguiri as a whole (which includes the Boure), the average age is
20.5, and the median age (for rural populations) is 14.5 (Direction Nationale de la Statistique
2005b), indicating a population significantly skewed toward youth. Our sample mean is
different than this reported figure because we excluded respondents under age 18. As with age,
proportions of our sample's marital status and years of education are likely to be different than
figures calculated for all residents of the prefecture of Siguiri because respondents under age 18
were specifically excluded. However, proportions of gender are not as susceptible to the
influence of age as other demographic characteristics may be. According to the Direction
National de la Statistique (2005b), males account for 50.7% of the population of the prefecture
of Siguiri, and females for 49.3%. In this regard, our sample's proportions resemble the
population as a whole. Since the sample was randomly chosen and represents a cross-section of
those present in the Boure at any given time, then there's a well understood chance that the
results can be validly extrapolated to the population of those older than 18 and present in the
Boure during the mining season.
In a 2001 study, it was estimated that there were an average of 4.6 condoms per man (aged
15-59) per year available in all of sub-Saharan Africa (Shelton and Johnston). While this
average is low, the same 2001 study indicated that condom availability is nearly as high as 17 per
man per year in some countries, indicating that there is great variability in condom availability
from one country to another. Likewise, there may be even greater variability within regions of a
given country such that large supplies of condoms cluster around urban or industrial centers.
To address condom availability in the Boure, the sample of 16 pharmacies and health
centers was visited regularly to count how many condoms were in stock. Condom tracking
enabled not only an estimate of the sufficiency of supply, but also determination if any condoms
that were available were actually being bought (and, by extrapolation, used). The results are
shown in Figure 6-1.
Given that the population estimates for each village varied from around five thousand to
eight thousand (with no exact figures available) and given that it is impossible to know precisely
how many sources for condoms there are in any given village, it is difficult to exactly determine
sufficiency of condom supply. However, supposing that the figures for Boukariah are
representative of average condom outlets in that village, one could say that condoms would be
hard to come by if there was any demand for them. By the same token, if the figures for
Kintinian are representative of average condom outlets in that village, one could say that
condoms are probably readily available if there is a demand for them.
The issue then centers on whether these samples of condom outlets are representative of all
condom sources in each village or not. Given the small average number of outlets visited in each
village on a regular basis (16 outlets divided by 4 villages = an average of 4 outlets per village),
this sample may not be statistically representative. However, it is representative by virtue of the
method of selection. Before beginning any research in each village, the chief was visited and he
appointed one of his aides to guide the research team to the main condom outlets. In addition to
this personal endorsement, condom vendors identified by the chief were asked to point out their
colleagues, too. In this way, major condom vendors were identified through social networking.
Secondary sources not known by the chief and the vendors he identified are probably not
primary outlets. Pharmacies and health centers appeared to be the main sources of condoms, yet
many identified pharmacies did not carry them at all, either. Pharmacies which did not carry
condoms at all and didn't intend to begin doing so were discarded from the sample in order to
avoid tracking outlets that would record a "O" each week. Thus, if anything, the sample is
skewed toward representing the villages as more thoroughly stocked with condoms than they
When considering the information in Figure 6-1, it is important to note the variance around
the mean. For example, condom outlets visited in Fatoya during the first observation period
reported an average of 275.33 condoms in stock. Yet the standard deviation was 573.16, and the
range in actual values was 0 to 1,440. In fact, three out of the six condom outlets monitored at
that time in Fatoya reported zero condoms available, even though they identified themselves as
usually having condoms in stock and intending to sell them in the future (which allowed them to
be included in the sample). Thus, availability aside, distribution throughout the village is grossly
disproportionate. If one lives in a neighborhood where none of the nearest pharmacies carry
condoms on a given week, will they go out of their way to the pharmacy in another
neighborhood which is carrying 1,440? Will they even know that that pharmacy is carrying
1,440 condoms that week? Given that this enormous variance about the mean is common for
nearly all of the villages monitored at each observation, it appears that condom distribution is
uneven across the board throughout the villages in the Boure.
For most of the villages, overall condom supply seemed to stay relatively stable, either
indicating that no one is buying condoms, or outlets that sell a lot of condoms are reliably
restocked. The one major exception to the stability rule is Kintinian, which plummets from an
average of 402.67 condoms at observation 3 (standard deviation: 335.05) to a mere 75 condoms
at observation 4 (s.d. 35.36). This is another thing that the tremendous variance about the mean
hides. One out of the three outlets observed on a regular basis in Kintinian is responsible for
some 800 condoms at every observation. At the fourth observation, however, this one outlet
reported only 50 condoms. This either implies that this outlet sold some 750 condoms in one
week (an extreme outlier when compared to the change in stock from every other observation in
every other village) or there was a measurement error for the fourth observation. But since the
averages are at the mercy of few observation points in each village, one outlet can make an
enormous difference in the calculated means.
Boukariah' s low condom counts appear to be representative of the village as a whole; the
research team had by far the most difficulty identifying any condom outlets whatsoever in this
village. Following the same protocol as the other villages (visiting outlets identified by the chief,
visiting networked outlets, discarding those that never carry condoms), nearly every pharmacy
visited was discarded from the sample due to widespread lack of intention to ever carry
condoms. The three outlets finally selected for the sample were the only three outlets found
carrying condoms at all in Boukariah after a far more extensive search effort than was required in
any of the other villages. Perhaps this is why Boukariah has a reputation for promiscuity
amongst the other villages of the Boure; neighboring villagers avoid taking brides from
Boukariah if at all possible.
Another important observation about Figure 6-1 is that means hide fluctuations of
individual outlets. In Balato's case, one sole outlet was responsible for the only condoms
reported at all for observation 1 (116 condoms; the others all had 0 in stock). By observation 2,
that same outlet dropped to only 5 in stock, and at observations 3 and 4, this one outlet had 0
condoms left. By observations 3 and 4, however, a different outlet had picked up the entire slack
while the rest (including the original well-stocked outlet) had 0 in stock. Thus, the average
jumped back up at observations 3 and 4 to levels similar to observation 1, but it was a different
outlet entirely that was responsible for the rebound in average. This inconsistency in individual
outlets' stock may discourage people who have come to rely on a single tr-usted source for their
condom needs. 10
These individual complexities for each observation in every village make it difficult to
generalize about condom availability across the board. Nonetheless, certain key conclusions can
be drawn to some degree. Across the board, condom availability depends to a large extent on
which mining village you live in. Even in villages that are well-stocked, supply may vary greatly
from pharmacy to pharmacy (poor distribution). And certain condom outlets' supplies are not
dependably stable. Nevertheless, if a villager is persistent enough and condom usage is
important enough to them, condoms are technically available in the Boure.
An important question to address is why condom stock fluctuates so much. Why isn't
there a consistent supply? Though this research project did not directly address this question,
three potential answers emerged in the course of informal conversations with condom vendors.
First of all, many vendors get their stock directly from Conakry (a 12 to 14 hour drive
away) instead of Siguiri (a one hour drive, at most) because bulk condoms are cheaper in
Conakry and the profit margin for resale is larger. Several outlets went for weeks with no
"' There seems to be a certain level of shame associated with condom use. Euphemisms such as "boubou" (a
traditional full-body-length robe) are used by buyers to refer to condoms, as in, "I'd like to buy two boubous,
please" (when addressing a pharmacist whom they know full well does not sell clothing). Other pharmacists, aware
of their customers' discomfort with talking about condoms at all, place large boxes of condoms in prominent
displays, so that customers can surreptitiously point to the boxes and mutter, "I want one of those." This
embarrassment suggests that some people don't want their condom use to become common knowledge, and
therefore perhaps visit the same vendor repeatedly, to minimize the pool of people "in the know" (many vendors
reported repeat customers). If a given vendor runs out of condoms, their repeat customers might have to risk
exposing their secret to another vendor in order to meet their needs.