“… But Then He Became My Sipa”: The Implications of Relationship Fluidity for Condom Use Among Women Sex Workers in Antananarivo, Madagascar

Abstract

Increasing evidence indicates that sex workers use condoms less consistently with regular (i.e., nonpaying) partners than with clients. Few studies have examined the extent to which these 2 categories are mutually exclusive. In an ethnographic study of women's sex work in Antananarivo, Madagascar, we examined how the meaning of women sex workers’ sexual relationships could shift among 3 different forms of sex work. Condom use was less likely in forms in which the distinction between client and lover (sipa in Malagasy) was fluid. For many sex workers, therefore, relationships they understood to be intimate imparted the greatest health vulnerability. It is important to examine the influence of the meaning of sexual relationships on condom use for HIV prevention. Policy implications for HIV prevention work with sex workers are considered.

In countries where the HIV epidemic is driven primarily by heterosexual intercourse, sex workers are typically considered a primary core group responsible for transmitting HIV to male clients who then pass HIV on to their wives and the general population.¹ Many conventional HIV-prevention efforts involve campaigns for individual behavior change targeting high-risk groups and incorporating the so-called ABCs (abstinence, be faithful, and condom use). Combining the core group perspective with the ABC approach has resulted in programs that focus narrowly on individual behavior change to encourage consistent condom use among sex workers and their clients. Research conducted under this paradigm focuses on sexual behavior and examines the determinants and barriers to condom use.

These approaches, however, often rest on and reproduce oversimplified understandings of sex work and sex workers’ relationships. Behavioral studies have been criticized for failing to recognize the diversity of sex work present in a given setting, leaving an impression that sex workers are a homogenous population.² In addition, these studies often assume that clients are sex workers’ only sexual partners—or their only risky partners.

Increasingly, studies are incorporating a distinction between sexual interactions with clients and sexual interactions with at least 1 other partner type, commonly described as a nonpaying partner (also described as boyfriend, regular partner, personal partner, or intimate partner). Condom use has been shown to be less consistent with nonpaying partners in a number of contexts in almost every region around the world, including Asia,³ sub-Saharan Africa,⁴ the West,⁵ and Central America.⁶

In most of these studies, sex workers’ multiple relationships are treated as belonging to separate, mutually exclusive categories. In the Western context, these categories do, in fact, appear to be largely mutually exclusive. Studies from Australia and Europe, for example, indicate that most sex workers distinguish clearly between clients and nonpaying partners.⁷ Women sex workers in Glasgow report almost 100% condom use with clients and nearly 0% condom use with regular partners.⁸

In other contexts, there is indication that these partnership categories are less distinct. Evidence suggests that women practicing more informal forms of sex work in sub-Saharan Africa do not easily differentiate between clients and nonpaying partners⁹; this pattern has been observed in Asia as well. A study from Cambodia describes how women practicing informal sex work within bars had difficulty distinguishing between “guests” (clients) and “sweethearts.”¹⁰ In a study from Kenya subtitled “Regular partners may contribute more to HIV transmission than clients,” Voeten et al. state that many women did not differentiate consistently between these groups, and perhaps this fact contributed to their increased vulnerability with regular partners.¹¹

What some of these studies imply, without adequately addressing, is the important assertion that clients and nonpaying partners may not always represent distinct types of relationships. Rather, these are fluid relationships, distinguished from one another through the development of affective ties over time. We addressed this gap in the literature by using ethnographic data to explore how sex workers and the clients and partners of sex workers interpreted their relationships and the circumstances that led to relationship change. We also examined the implications of relationship fluidity for women's health vulnerabilities. This understanding is critical to the development of effective strategies to prevent HIV and other sexually transmitted infections.

THE STUDY SITE

A growing city of nearly 2 million people located in the central highlands of Madagascar, Antananarivo is still surprisingly ethnically homogenous.¹² Women's employment opportunities in Antananarivo are concentrated in lower-paying jobs in the formal economy (e.g., textiles) or the informal economy (e.g., produce sales, embroidery).¹³ Sex work provides some women monthly earnings that far surpass what they could expect to earn through these other alternatives. Antananarivo is a globalizing city with Internet access on nearly every corner. Foreign media and clothing are ubiquitous. Western notions of love and romance are everywhere, reflected in Malagasy (as well as American and French) songs blasting tales of love and loss on the radio and in video.

Prevalence of HIV in Madagascar is still low, but rising—estimated at less than 1% in recent national surveillance studies.¹⁴ As elsewhere, HIV prevention efforts in Antananarivo are largely focused on individual behavior change targeted at high-risk groups—sex workers being first on the list. Sex workers and their clients are encouraged to use condoms for their sexual interactions.¹⁵ In addition, mass media campaigns encourage the general population to delay sexual debut, to be faithful to 1 partner, and to use condoms with nonregular partners (the ABC prevention strategy).¹⁶ Although a few studies indicate some diversity in sex work in Antananarivo¹⁷ and that condom use is lower among nonpaying partners than among clients,¹⁸ the meanings of sex work and sexual relationships within different forms of sex work remain relatively unexplored.

RESEARCH DESIGN

Women sex workers’ sexual relationships and condom use were explored within a broader project examining the social organization of women's sex work in Antananarivo. Data were collected by a small research team led by K. S. over the course of 14 months of fieldwork (January–December 2003; May–June 2004). As is the case in most of the world, sex work is highly stigmatized and sex workers are marginalized in Antananarivo. Therefore, the community perception of sex work has bearing on sex workers’ lives and opportunities.

The research team first conducted in-depth and informal interviews with experienced community informants (n = 24), and semistructured interviews with general community respondents (n = 60) to gain the community's perspective. Three forms of sex work were described by the community as ambany (low), antonony (middle, in-between), and ambony (high; the English is a direct translation of the community-generated terms). Women sex workers were purposively selected to capture the breadth of the social organization of sex work depicted by the community.

The 3 sex work forms identified by community respondents were examined through “key participant” women sex workers (n = 10) and matched non–sex worker comparison women (n = 8). Sex workers were either recruited by the research team or with the assistance of a nongovernmental organization and a sex work association. Because the “high” category of sex work was practiced in establishments patronized by tourists, K. S., who speaks Malagasy, was able to approach these women within these establishments. It was more difficult to establish relationships with women sex workers who practiced on the streets at night or in poor neighborhoods. The research team relied on a nongovernmental organization conducting outreach work and members of a sex work association for initial contacts.¹⁹

Key participant women's lives were examined via participant observation, life histories, in-depth interviews, and sexual relations diaries. Life histories and in-depth interviews were conducted by K. S. together with a research assistant in a quiet place of the participant's choosing (in 2 to 3 sessions over a total of 1.5 to 2.5 hours for life histories, and a 45-minute to 1-hour session for in-depth interviews).

Additional sex workers were included through participant observation (when possible) in sex work locales (n ≈ 50) and semistructured interviews (n = 13). Finally, semistructured interviews were conducted with a convenience sample of sex workers’ clients (n = 20) by P. G. R. in 4 different sex work locales. A detailed description of the study populations and methods used is presented in Table 1.

TABLE 1.

Study Populations, Survey Methods, and Data Gathered: Antananarivo, Madagascar, 2003

Population	Methods Used	Data Gathered
Community informants	Informal and in-depth interviews and repeated in-depth interviews with 2 physicians working with sex workers	Insight into the social organization of sex work in Antananarivo and how HIV prevention efforts have been formulated and conducted
Past and present Ministry of Health officials (n = 2); other government officials (n = 2); sex work association members, peer educators (n = 4); additional NGO/ intergovernmental agency employees (n = 11)
Physicians working with sex workers (n = 5)
Community respondents	Semistructured interviews with 30 men and 30 women in 5 neighborhoods (6 men and 6 women in each neighborhood) selected to roughly represent the age, gender, and socioeconomic status distribution of Antananarivo	Understanding of the local social organization of sex work, including the meaning of sex work and the different forms or social categories of sex work
Women sex workers	Interviews (some in-depth), observation, and diaries	Understanding of women sex workers' lives and concerns, including their perceptions of their lives and identities, the social organization of sex work, their health needs and concerns, and their relationships with men
Key participants (n = 10)	Life history interviews; participant observation of key participants in their homes and sex work locales; sexual relations diaries (n = 5); and in-depth interviews: social organization of sex work (n = 8); clients (n = 8); gender and power (n = 7); and life and health concerns (n = 7)
Additional women sex workers (n = 50)	Participant observation of women in key participants’ sex work locales: low form in a very poor neighborhood; middle form on the street with key participants via outreach efforts; and high form in bars/nightclubs; semistructured interviews (n = 13; 2 low, 10 middle, and 1 high sex work forms)	Corroboration of key participants’ information and further understanding of the social organization of sex work
Key participant non–sex workers (n = 8)	Life history interviews; sexual relations diaries (n = 3); and in-depth interviews (n = 7): social organization of women's work; gender and power; and life and health concerns (n = 7)	Further understanding of the meaning of sex work through comparing the lives and identities of non–sex working women matched on age, background socioeconomic status, and ethnicity to key participant sex workers
Clients	Semistructured interviews with convenience sample of 20 men (5 per locale) located in or near 4 sex work locales: 1 locale where low form is practiced, 2 middle form, and 1 high form	Clients’ perspectives on and interpretations of sexual interactions with sex workers and clients’ sexual behavior
Additional community informants: anthropologists, sociologists, historians, demographers, TV producers, hotel managers, police officials	In-depth and informal interviews	Insight into women's rights and racial and ethnic identities in Antananarivo, regulations and laws regarding sex work, and sexuality and sex work in the media in Antananarivo

Note. NGO = nongovernmental organization.

Interviews were transcribed, translated and then brought into Atlas/ti version 5.2 (Scientific Software Development, Berlin, Germany) along with field notes for data coding and analysis. Data were read for themes relevant to condom use, discussion and interpretation of sexual relationships, and description of health concerns. Condom-use reporting bias, a concern with many HIV prevention studies, was minimized through triangulation of methods and study populations, as well as through an extended period of data collection, which both built trust and provided opportunity for internal data checks.

DIVERSITY, FLUIDITY, AND CONDOM USE

The analysis of the social organization of sex work has been described in greater detail elsewhere.²⁰ In general, women's sex work in Antananarivo during the study took place without the involvement of a third party of any kind; there were no brothels or pimps in Antananarivo. Women controlled their earnings and, for the most part, were their own managers.

Community respondents identified 3 forms of sex work (low, middle, and high) based on sex work location, price per client interaction, and women's ethnic and racial identity. Racial and ethnic identity formation in Antananarivo may be traced back to precolonial and colonial history and organizes the population of Antananarivo into 2 racial categories—those who self-identify as descendants of nobles or commoners (who are “white”) and those who are identified as descendants of slaves (who are “black”).²¹

The ethnic distinctions made in Antananarivo by the local ethnic group, the Merina, differentiate the Merina from those identified as non-Merina or côtier (coastal people).²² Women identified with the low form of sex work tend to be considered “black”; women identified with the middle form are more likely to self-identify as “white”; and those identified with the high form are often described as côtier.

The women practicing the low and the middle forms of sex work operated in poor neighborhoods and on main city streets, respectively, and described the sex work they practiced as mitady vola (looking for money). The women practicing the high form of sex work did so in bars and nightclubs and described what they did as mitady vady vazaha (looking for foreign husbands).²³ A detailed description of the forms of sex work is presented in Table 2.

TABLE 2.

Characteristics of the Social Organization of Women's Sex Work: Antananarivo, Madagascar, 2003

	Low Sex Work	Middle Sex Work	High Sex Work
Marital status	Generally not legally married	Mixed, but more often legally married	Generally not legally married
Background socioeconomic status	Usually very poor	Varied	Varied
Mean age (range), y	27.0 (18–35) (n = 7)	31.4 (19–48) (n = 16)	25.2 (19–40) (n = 9)
Mean education (range), y	4.4 (0–9) (n = 7)	6.2 (3–11) (n = 15)	8.4 (3–13) (n = 8)
Mean no. children (range)	1.3 (0–4) (n = 7)	2 (0–4) (n = 16)	0.70 (0–2) (n = 9)
Major life concerns	Disease, particularly STIs and HIV	Being outed as a sex worker to the family and the community	Getting married
Racial/ethnic identity	Mainty (“black,” slave descent)	Fotsy (“white,” nonslave descent) and Merina (local ethnic group)	Côtier (“coastal,” outside Antananarivo province)
Stated goal of sex work	To make money (mitady vola)	To make money (mitady vola)	To find a foreign husband (mitady vady vazaha)
Locale of sex work	Poor neighborhoods during the day, sometimes at night	Streets in nicer parts of town, mostly at night, occasionally during the day	Nightclubs, bars, restaurants, hotels
Price per sexual interaction	Less than US $1.00	Between US $2.50 and US $17.00	Between US $17.00 and US $50.00
Typical male client type	Poor Malagasy	Middle-class to wealthy Malagasy	European tourists or business travelers
Dress of sex worker	Poorly dressed, not often in provocative clothing	Some dress provocatively, not in the latest fashions, some dress modestly	Provocatively dressed in the latest expensive fashions
Language spoken	Malagasy	Malagasy, may know some French	Must know some French
Typical discussion of price	Discuss price, but client may not pay for every sexual interaction when it takes place	Discuss price and specifics of sexual interaction before sexual exchange	No discussion of price or particulars before sexual interaction

Notes. STI = sexually transmitted infection. Statistics provided above for participants (n = 33) include 10 key participants, 13 additional sex workers formally interviewed, and 10 sex workers met through participant observation whose background characteristics were recorded in field notes. Other characteristics (stated intentions, client type, dress, language, discussion of price) represented “ideal types,” and there were exceptions to every case (e.g., not every woman in a nightclub was always able to wear “expensive” fashions).

Condom Use in Antananarivo

Condom use between sex workers and their sexual partners was structured by popular understandings of condoms within Antananarivo. Condom use has been influenced by a number of factors including, but not limited to, family planning communication efforts and HIV prevention efforts, which both began in the early 1990s.²⁴ Evidence from other sub-Saharan African countries indicates that targeting strategies and the ABC approach to HIV prevention may be inadvertently symbolically associating condoms with immoral or commercial sexual encounters.²⁵

Following this logic, couples wishing to understand their relationships as meaningful and moral may be less inclined to use condoms. Research participants described 2 relationships within which condoms were not necessary or not appropriate: sipa (lover, such as a boyfriend) and vady (spouse) relationships. As one key participant sex worker stated: “With the client you use a condom; with a sipa you don't use a condom.”

Sipa and vady relationships are ideally characterized by love and trust. Love, trust, and the fear of disease were the major themes raised by research participants in descriptions of condom nonuse within sipa and vady relationships compared with condom use in other sexual relationships—findings similar to those from studies in other regions of the world.²⁶ As a 25-year-old man with nearly a completed high-school education explained:

With women who are “looking for money” [sex workers] I use a condom. With my sipa, when I trust her, and I'm certain, because she isn't a sex worker, then I can trust her and I don't use a condom.

Participants from all the sampled groups—sex workers, clients, non–sex workers—employed this same logic to describe any exceptions to expected condom-use practices. For example, some participants asserted lack of trust tied to concerns about disease to explain why they used condoms with their intimate partners.²⁷ In other cases, condom nonuse between clients and sex workers was justified through emotional intimacy, or the anticipation of future emotional intimacy. The same man quoted previously, initially describing the clear distinctions between with whom it is and is not important to use a condom, went on to say, “Sometimes if I really like the woman [sex worker] I'm with, then I don't use a condom.”

Sipa relationships may be characterized by love and trust, but they are also unions that can be formed and severed quickly. For some women sex workers, the distinction between a client and sipa relationship was also highly fluid. We show how relationship fluidity contributed to women's health vulnerability given that condom use was deemed less appropriate with sipa or vady than with clients.

Fluidity in Women Sex Workers’ Sexual Relationships

Women practicing the low form of sex work.

The low form of sex work was practiced within slumlike neighborhoods throughout the city during the day.²⁸ The women stood in areas of pedestrian traffic and accompanied the men they met to nearby small shacks rented by the half hour for the purpose of sexual intercourse. Women sex workers hoped that the men with whom they had intercourse would come to consider them as a sipa or vady (i.e., someone for whom a man should provide), so they could rely on the men in times of need. Men, in turn, have learned that women sex workers often make more money than they themselves can expect to earn. As a result, men and women each look to the other for some financial stability in their lives, fostered through romantic relationships that often begin as a sexual interaction with a client.

Kara, a 22-year-old key participant with 3 years of education called home half a bed in one of the shacks used for sexual interactions during the day. Her relationships with men illustrate how the meaning of sexual relationships can shift. She described how her relationship began with a regular client:

He courted me. He took me for a passage [brief sexual interaction] and after that he bought me some toaka [local moonshine] and beer… . And he told me “I won't look for other women anymore. I will always go with you.” … Sometimes, he buys some clothes for me… . Sometimes, he comes even without going for any passage but just takes me to the bar, or asks me to go with him to buy some things.

Kara's description is more suggestive of a romantic relationship than a client interaction: He “courted” Kara, and he took her for a passage; he sometimes invited her out to the bar or to go shopping without any exchange of sexual activity. Kara also stated that he only paid her for a little less than half of their sexual interactions. She soon began to consider this man her sipa rather than her client:

Yes, he was a client, but then he became my sipa… . We went and had a picture taken together … it was then that he told me … “I love you, but I have a wife.”

He no longer paid her for sexual relations, but she did not mind because she loved him.

The following month Kara had a fat lip and a sore back—her sipa had beaten her with a piece of firewood. Kara explained: “When he sees that I have been with someone else … then he beats me, but he doesn't give me any money … and so I go with other men.”

This relationship, which began and ended within a little more than 2 months, illustrates the struggles faced by women in an extremely economically depressed setting. This man was not in a position to economically provide for Kara and Kara could not forgo her means of income generation to be only with this man. The cycle then repeated itself. Two months later, Kara was caught in another abusive relationship with a man whom she called a client, then a sipa, and then “a client who was together with me.” Client relationships could very quickly become understood as sipa or vady relationships, and often these categories of relationships were not clearly delineated.²⁹

Women practicing this form of sex work frequently expressed (unsolicited) fears about acquiring HIV or other sexually transmitted infections. Most, however, also indicated that they did not insist on condom use with the men with whom they had sexual relations. It is generally found that a frequent explanation for condom nonuse is the pressure that immediate financial need creates, making women more likely to acquiesce to men's interests in having unprotected sexual intercourse. Condom nonuse could also be influenced by the shifting definition of relationships between client and lover or spouse. By failing to insist on condom use, a woman was inviting the sexual relationship to be understood as a relationship between trusting partners—in other words, as sipa or vady. As relationship meaning shifted, condom use became more and more unlikely.

Women practicing the middle form of sex work.

Women who practiced the middle form of sex work were more likely to self-identify as “white,” Merina (members of the dominant local ethnic group), and to have families who originate from the area and live nearby. As such, these women hold social capital that they would certainly lose if they were to be discovered as sex workers. Therefore, most hid this identity from their communities and families, and sometimes their own husbands.

To this end, most of these women practiced sex work on dark streets at night. This environment encouraged very formal transactions. Women awaited men on the street corner, they engaged in a discussion of the particulars of the sexual interaction, and then accompanied their client to a cheap hotel or back to his car to exchange sexual intercourse for immediate financial remuneration.

Most of these women made an effort to keep their romantic relationships distinct from their client relationships. Unmarried women pursued romantic relationships with men they met in other contexts. Unlike women practicing other forms of sex work, these women had a more pejorative view of those who allowed client relationships to transition into other kinds of relationships. They thought sex work should be about making money, not building relationships.

Compared with women in the low form, participants practicing the middle sex work form were much more likely to report consistent condom use with their clients. These women were also less likely to express concerns about sexually transmitted infections but more likely to express concerns about violence on the streets, as working on isolated streets at night made them easy targets for physical and sexual assault.³⁰

Despite the formality, women who practiced the middle sex work form were sometimes pulled into more affective relationships with established “fixed” clients (regular clients). Some women described having a difficult time maintaining more formal client relationships, but that it was often the clients who encouraged such shifts.³¹

Women practicing the high form of sex work.

The high sex work form was practiced in bars and nightclubs, where women went to “look for foreign husbands.”³² These women made a conscious effort to maintain very informal interactions with the men they met, so that these interactions could be interpreted as a (Western construction of a) boy-meets-girl encounter in a bar or nightclub.

The women usually avoided discussion about money or the particulars of sexual intercourse before it took place. Participants described the importance of the woman's waiting for a man to approach her rather than trying to actively attract a man so as to ensure that he would be more receptive to any (financial) concerns she might communicate to him later (communicated as money needed for her family, e.g., a sick parent). In addition, they believed that there was an increased likelihood that if a man approached the woman first, his interest in her might extend beyond a 1-time sexual encounter.

Although women in clubs and bars insisted that they used condoms with their clients, they also explained that there were many women who did not always use condoms. Women were more likely to insist on using a condom if they knew they were not interested in the man as a potential husband. Alternatively, they were more likely to risk sexual intercourse without a condom if they were romantically interested in the man, or if they believed he might be romantically interested in them.

The result of these strategies was not only exposure to HIV and other sexually transmitted infections but also pregnancy. Women openly discussed abortions they carried out after learning that a potential husband was not in fact committed to their shared future.³³ For example, Simone, a 24-year-old key participant with 7 years of education, explained that she had had to abort recently. The father was an older European man who had been sending her money between visits to Madagascar. She was led to believe he was interested in marriage. When he learned she was pregnant, instead of sending more money as promised, on his next visit he found another woman at the bar.

Soon after that, Simone met Pierre: a young, fit European tourist, he represented the ideal potential husband. Once he left Madagascar, he began financially supporting her and promised to return. Around her friends and family members she referred to him as her vady. However, in a formal interview, she referred to him as a fixed client, explaining that this was because they had not yet formalized their commitment (i.e., through formal marriage). Although he seemed committed, she hedged her bets by continuing to look for a foreign husband on weekends at the bar. She also explained that she was lonely when not going out and so she sought out Malagasy sipa for company.

Pierre did return for a long visit, during which time they did not use condoms.³⁴ He and Simone began compiling the paperwork toward a travel visa for her to visit him. Once he left the country Simone returned to going to the bar. Although she suggested that she was now more interested in money than in a potential husband, sex work practice in this setting was informally structured; therefore, shifts in relationship understanding and potential for condom nonuse would remain likely.

Key participant Denise had a Malagasy “spouse” off and on for 7 years who served as an emotional safety net while she looked for a foreign spouse. Denise said that they never used condoms, as they were spouses. However, they had separated on numerous occasions, always based on the same complaint. Denise explained: “He comes home with love-bites, but I didn't give them to him.”

Women who practiced the high form of sex work were therefore exposed to sexual health risk both through their relationships with potential foreign husbands and through romantic relationships with Malagasy men.

DISCUSSION

Our findings show that in Antananarivo the meaning of sexual relationships was highly fluid for women practicing certain forms of sex work. Once affective ties were introduced into a relationship between a client and a sex worker, condom use became less appealing, particularly if that relationship became understood as one with a sipa or vady. The result was that the health of many women sex workers was put at greater risk with men whom they understood to be sipa or vady rather than with men whom they understood to be clients.

Women practicing what the community respondents described as the middle form of sex work were more likely to maintain formal relationships with their clients, structured by their interest in discrete sex work practice. Women practicing the low and high forms, alternatively, either allowed or sought fluidity in the understanding of their relationships.

Such fluidity was integral to women's efforts to improve their situations through building intimate relationships with men who could provide stable economic support, and condom nonuse served toward building that intimacy. For these women, distinctions prevalent in conventional public health discourse between regular partners and clients did not exist.

This finding has implications for research and programs that assume that sex workers conceptualize 2 discrete groups (clients vs regular partners). Efforts to prevent HIV could benefit from more carefully accounting for the interpretation and fluidity of sexual relationships and condom use within those relationships. In addition, it is important to account for different contexts in which sexual relationships take place and to determine which relationships are most likely to increase health vulnerabilities for the individuals involved.

We found that significant health vulnerabilities were also generated through violence—whether structured as intimate partner violence, street violence, or sexual assault from relatively unknown perpetrators (such as formal clients). Prevention programs working with women sex workers should make efforts to address these women's pressing health concerns, such as violence, in addition to promoting condom use.

Not all of sex workers’ sexual relationships began as client relationships. Women practicing the high form of sex work, for example, took emotional refuge in relationships with Malagasy men while awaiting the ideal European husband. Women practicing the middle form were sometimes legally married to men they met outside of (or before) their sex work practice.

These relationships also contributed to women sex workers’ health vulnerability. Mounting evidence shows that marriage is a risk factor for HIV for women in sub-Saharan Africa (as well as other regions).³⁵ This important finding affects any woman in Antananarivo who is married or in an intimate partnership (whether or not it began as a client relationship) through the cultural expectation that men have extramarital relationships that women should quietly tolerate.³⁶

Women sex workers experience persistent gendered social, political, and economic inequalities that are fundamental to their disproportionate health vulnerabilities. Given the reality of these gendered power structures, interventions encouraging use of the male condom need to place more of the onus for condom use on men. In addition, continued encouragement of the development and distribution of women-controlled methods for HIV prevention is needed.³⁷

In Antananarivo, the health vulnerability resulting from condom nonuse was structured, in part, by the general understanding of when it was and was not appropriate to use condoms. This understanding may result from HIV prevention messages grounded in idealized Western assumptions about sex work and sexual behavior in relationships.³⁸

The ABC approach to HIV prevention has been criticized as overly reliant on a Western conception of marriage, assuming sexual monogamy with only rare episodes of infidelity (assumptions being tested by evidence that marriage is an HIV risk factor for women).³⁹ In addition, within the current policy environment, the shift to strategies emphasizing abstinence and fidelity and away from condom use potentially exacerbates the association between condom use and immoral sexual interactions, further narrowing the description of with whom it is necessary to use condoms. Our findings indicate that the assumptions on which the ABC strategy is based do not correspond to the realities confronted by our study population.

The prevention and research community must find ways to take the symbolic power away from the condom. If prevention efforts continue to rely on categories such as client versus regular partner or primary versus occasional partner for the promotion of condom use, evidence suggests that a sizeable number of sexual interactions that should be protected will go on being unprotected (particularly in non-Western settings). The challenge becomes one of restructuring messages about condoms, moving away from promotion of population categories with whom to use them (and therefore implicitly with whom not to), and toward nonjudgmental encouragement of their use in sexual relationships more generally.

Inequalities ranging from a country's position within the global economy to local gender-based discrimination explain many (but not all) women sex workers’ participation in the sexual economy around the world. Until such inequalities are greatly diminished, sex work will remain a popular constrained choice for income generation. It is therefore imperative that the public health community be in a position to work with this population toward understanding their diverse needs and concerns.

However, this notion is incompatible with current funding restrictions tied to the US Global AIDS Act of 2003, which founded the President's Emergency Plan for AIDS Relief. Under this legislation, prostitution is equated to trafficking and in need of eradication, and both foreign and US organizations that receive US Agency for International Development funding toward HIV prevention are required to have a policy explicitly opposing prostitution.⁴⁰ Our findings call into question the appropriateness of this interpretation of prostitution and suggest that these restrictions could seriously compromise HIV prevention efforts with sex workers, particularly efforts attempting to account for sex workers’ diverse health concerns. A bill reauthorizing the Global AIDS Act was passed in 2008. While the reauthorization bill relaxed earmarks on abstinence, it maintained the restrictions on funding applying to sex workers.

Our research underlines the recommendations of the Institute of Medicine's evaluation of the President's Emergency Plan for AIDS Relief in anticipation of the renewal and increased funding of this program.⁴¹ The Institute of Medicine's recommendations emphasized the need to allow regions to have flexibility in the substance and goals of their intervention efforts. Intervention efforts must be able to address the needs of each specific context, rather than apply the same strategy to every context on the basis of one assumption about sexual behavior and one assumption regarding the interpretation of the meaning of prostitution.