Introduction
While much of the current attention to control the AIDS epidemic in sub-Saharan Africa is directed towards treatment, there is an at least equally urgent need to improve prevention efforts: 9% of the adult population aged 15-49 in sub-Saharan Africa may be infected and few countries show a decline in HIV incidence [1]. In severely affected countries such as South Africa, national HIV prevalence in women attending antenatal clinics has increased from less than 1% to more than 20% in a decade [2].
The lack of success in preventing HIV transmission has been attributed to the reluctance of governments to confront AIDS and a failure to prioritize and scale-up prevention programs [3]. Program effectiveness may be increased by prioritizing interventions and by focusing implementation on a core set of the most cost-effective interventions. Focusing on high-risk or core groups, especially commercial sex workers, has been promoted as a strategic and cost-effective intervention to limit HIV transmission in Africa and elsewhere [4]. The best evidence of the effectiveness of this approach is from Thailand, where a 100% condom policy was implemented early in the epidemic in regulated brothels and subsequently HIV prevalence among new military conscripts declined [5].
There are several reasons to question, however, whether core group approaches focusing on female sex workers will have a significant impact on the HIV epidemic in sub-Saharan Africa. First, there is no strong evidence in sub-Saharan Africa of the effectiveness of sex worker interventions on limiting the spread of infection to the general population. The only core group intervention in sub-Saharan Africa demonstrating an effect on sexually transmitted infections (but not HIV) among the non-core was a presumptive treatment intervention among sex workers in a mining community in South Africa, which is not appropriate for widespread replication [6]. Second, in most of sub-Saharan Africa, the epidemic is in a later phase [7] than in Thailand and consequently there is less opportunity to contain the epidemic within core groups. Consequently, the relative contribution of commercial sex to the epidemic in sub-Saharan Africa has decreased and the potential impact of future sex worker interventions in sub-Saharan Africa has probably been overstated. Third, targeted interventions with commercial sex workers in sub-Saharan Africa have proven challenging to implement and sustain. Even though commercial sex may be common in large urban areas, identifying commercial sex workers is difficult because commercial sex is rarely regulated. Many women occasionally exchange sex for money without turning it into a profession [8]. The stigma associated with both sex work and AIDS further complicates implementation of interventions with sex workers. Finally, national sexual behavior surveys and local quantitative and qualitative studies suggest that sexual mixing in much of sub-Saharan Africa is diffuse and dynamic and only to a limited extent concentrated on commercial sex work [9-11].
This paper presents an innovative method that changes the focus from risky people to risky places. This method, dubbed PLACE (Priorities for Local AIDS Control Efforts), can be applied in a district, town or city. It systematically identifies places where new sexual partnerships are formed and assesses HIV prevention program coverage in those places. The focus is on new partnerships because rates of new partner acquisition are a major force driving HIV and other sexually transmitted epidemics [12]. The PLACE method recognizes that it may be more feasible to reach individuals with high rates of new partnership acquisition with a place-based rather than a risk-group approach. The PLACE approach extends the core group approach to core areas and identifies core areas where people with high rates of new partnership acquisition can be reached without requiring intervention programs to grapple with either defining or assigning membership in groups that often carry stigma. This paper presents findings from a PLACE assessment in four urban applications in South Africa.
Methods
The first application was in a township in Cape Town (population 80 000) in the Western Cape Province [13]. Townships were previously all-black residential areas. Albeit improving, infrastructure, employment, and housing remain worse in townships than elsewhere. Official HIV prevalence estimates for antenatal patients are available at the provincial level. In the Western Cape, HIV prevalence among antenatal patients increased from 1.6% to 8.7% between 1995 and 2000.
The protocol was subsequently implemented in the Eastern Cape Province as part of an AIDS initiative to focus on areas with elevated sexually transmitted infection/HIV risk. It was implemented in an East London township (population 100 000), a Port Elizabeth township (population 60 000), and the Port Elizabeth business district (population 12 000). In the Eastern Cape, HIV prevalence among antenatal women increased from 6.0% in 1995 to 20.2% in 2000 [14].
There are three phases of fieldwork. In the first phase, interviewers asked over 250 informants selected by convenience in public places such as taxi stands, bars, and streets to identify public sites where people meet new sexual partners. In the second phase, interviewers visited these sites, interviewed someone knowledgeable onsite in a face-to-face interview about the characteristics of the site and its patrons, and mapped the location of the site on an aerial photograph (Cape Town) or using global positioning systems.
In the last phase, interviewers asked individuals socializing at sites in a brief face-to-face interview about their rate of new partner acquisition, condom use, and frequency of site attendance. In the Port Elizabeth business district, six individuals per site were interviewed. In Cape Town and East London, individuals were interviewed at sites most frequently reported by key informants and a stratified (by geographic zone) random sample of the remainder. In the Port Elizabeth township, individuals were interviewed at a random sample of sites. In the townships, all individuals at a site were interviewed unless there were more than 16 men and eight women. In these sites, interviewers selected a sample of 16 men and eight women based on whether they were standing along an imaginary 'X' drawn between opposite site corners.
The Institutional Review Board of the University of North Carolina and the University of Cape Town approved the protocol. Implementation in the Eastern Cape was done under the direction of the Eastern Cape Provincial Department of Health, which also reviewed and approved the protocol. Data were keyed locally and compared at the University of North Carolina using SAS version 8 (SAS Corp., Cary, North Carolina, USA) and STATA version 7.0 (Stata Corporation, College Station, Texas, USA). Data obtained from people socializing at sites were analyzed controlling for site. No sampling weights were applied.
Results
Each of the four studies was completed within 4 to 6 weeks between September 1999 and July 2000. More than 250 key informants, including students, out-of-school youth, petty traders, community health workers, security guards, alcohol sellers, taxi drivers, and teachers were interviewed in each area (Table 1). Over 75% of the 969 reported sites were located and an interview conducted with someone knowledgeable about the site. Refusal by site respondents was more common in the business district (14.1%) than the townships (range, 0.3-4.3). Township sites shared some similarities. More than 75% were small local drinking places, called shebeens. Approximately 25% had 30 or fewer patrons at busy times. In the business district, sites were more likely to have more than 100 patrons and bars, taverns, bottle stores, nightclubs, streets, and hotels replaced the informal shebeens. At virtually all sites weekends were reported as busy times.
Over half of the site respondents in each area (with the exception of the Port Elizabeth township) reported that people meet new sexual partners onsite. Female sex work and gay partnerships were reported at fewer than 6% of township sites but at over 20% of sites in the business district. Condoms and other AIDS prevention activities were rare everywhere. Figure 1 shows condom availability in Cape Town. A majority of township site respondents reported willingness to implement onsite prevention programs.
Approximately 4500 people socializing at over 180 sites were interviewed (Table 2) during weekends, when sites were reportedly the busiest, but before dark for safety reasons. Overall, for every woman visiting a site, there were two men. Refusal rates were below 2%, although there may have been incomplete recording of refusals. The mean age of people socializing at these sites was about 30 for both men and women; patrons under 20 years were more common in the East London township. In the three townships, about 90% of respondents resided in the study area and over 40% were unemployed. In contrast, one third of the business district patrons did not reside in the area and male unemployment was 12.6%.
Many patrons were regular customers and many visited more than one site per day. The mean number of sites visited per day was 2.5 or greater in all areas. The majority of individual respondents reported that other men and women come to the site to meet new sexual partners, confirming the key informant and site representative reports. Between 22 and 40% of men and women in the townships said they had personally met a new sexual partner at the site of the interview. In the central business district, 57% of women reported having met a new partner onsite.
The mean number of sexual partners in the past 4 weeks among men was just above two in Cape Town and Port Elizabeth, and 1.3 in the East London township. Women reported fewer partners than men in the townships, but more in the business district. In the townships, the proportion reporting two or more partners in the past 4 weeks ranged from 30 to 57% among men and from 19 to 46% among women. In Port Elizabeth's central business district, almost three times more women (39%) than men (14%) reported having four or more sexual partners in the last 4 weeks. Many sexual partnerships were new. In Cape Town, almost half of the reported sexual partners in the last 4 weeks among men and women were new. In all studies and for both men and women, having a new sexual partner in the past 4 weeks was associated with being younger than 30, living outside the study area and being employed (data not shown). Ever use of a condom ranged from one third of respondents in the East London study to two thirds of respondents in the central business district. Use of a condom at the last sexual intercourse with a new partner among those who acquired a new partner in the last 4 weeks was about one in three in the townships but almost twice as high in the central business district. In the latter study, almost a quarter of respondents reported having a condom with them at the time of the interview.
Discussion
The four studies showed that sexual networks in these four urban areas in South Africa are extensive, diffuse, and characterized by widespread high rates of new partnership formation and concurrency with little acknowledged commercial sex. Key informants identified hundreds of sites where people meet new sexual partners, many more than anyone involved in the study or local AIDS programs had expected. Subsequent interviews with site respondents and patrons confirmed a majority of these to be places where people met new sexual partners. Numbers of patrons reported by the site representatives and counting of patrons during site visits indicated that a large proportion of the population was involved in social and sexual networks evolving around the sites, with a ratio of one female to two males. Furthermore, individual interviews with men and women visiting the sites revealed high rates of new partner acquisition among both sexes. The four studies also showed that intervention efforts were at best modest and mostly absent in places where people go to meet new sexual partners. The lack of AIDS prevention messages and low availability of condoms stands in sharp contrast with the reported willingness of most site representatives to be part of AIDS prevention programmes and sell condoms.
The results suggest that key informants missed few public sites where people meet new sexual partners. Interviewers questioned hundreds of people until no new sites were named. There was no indication that informants found the question 'Where do people go to meet new sexual partners?' unacceptable. Few refused to participate. It is unlikely informants felt threatened about participating because no personal identifiers were requested and they were not asked about their own behavior or that of possibly stigmatized groups. Missed sites are likely to be small, transitory, or private. Such a systematic search involving a wide variety of respondents at the community level for sites where any type of new sexual partnership is formed is unique to the PLACE method, although other studies have identified places where sex work occurs [15,16].
Interviewers identified site representatives upon arrival at the site and usually selected an employee or regular patron. The criteria for being a site representative were intentionally broad to prevent deductive disclosure of identity and to limit the number of revisits to a site. Although some respondents were likely to be more knowledgeable about the site than others, the information requested was information that an employee or regular patron could answer. The response rate was high in the townships and not much lower in the business district, indicating that respondents were not reluctant to answer questions about onsite activities and patron characteristics. Validation of the site representative reports with interviewer observations and individual interviews during the third phase of the PLACE method showed that respondents were willing to report alcohol consumption and youth attendance. The age distribution of those socializing at sites was consistent with level of youth attendance reported by site representatives, suggesting that site representatives did not significantly under-report youth attendance.
Commercial sex may have been underreported in both the townships and the business district; however, the most plausible explanation for the much higher proportion of site representatives reporting commercial sex in the business district than in the townships is that commercial sex was actually more common in the business district. Moreover, women socializing in the business district were much more likely to report engaging in commercial sex. Further assessment of the extent to which commercial sex was under-reported was difficult partly because respondents had different definitions of commercial sex. Women socializing at sites were asked if they had received payment for sex in the past 4 weeks. A third of the women in the townships who reported receiving payment for sex in the past 4 weeks reported only one new sexual partner during that time, suggesting that they may have interpreted the question to include gifts from regular sexual partners.
The direction and extent of bias in data obtained from people socializing at sites are difficult to assess. The sites selected for interviews were not a random sample of sites (except in one township) and interviewers were unlikely to have succeeded in identifying a representative sample at selected sites. The dynamic nature of sites provided ample opportunity for individuals to elude interviewers. Efforts to minimize self- presentation bias included requesting verbal anonymous informed consent, assuring confidentiality, and designing brief and simple close-ended questionnaires without any skip patterns. Questions about commercial sex were asked at the end of the questionnaire and there were no questions on marital status to disclose infidelity. A 4-week reference period was used for the number of sexual partners. In spite of these efforts, self-presentation bias was likely to decrease the proportions willing to report multiple sexual partnerships. However, men may have over-reported partnerships. Although under-reporting sexual partnerships is more common in face-to-face surveys, these interviews took place at sites where norms, especially for men, may be perceived differently than in household settings. There is little reason to expect that women over-reported sexual partnerships. Under-reporting is much more likely. In a 1998 national survey in South Africa, only 2.9% of women aged 15-49 years reported more than one sexual partner in the past year [17]. This low rate appears incompatible with the high prevalence of HIV infection in South Africa and suggests that household surveys are likely to underestimate multiple partnerships. The internal consistency between the key informant, the site representative, and individual interview data is perhaps the strongest evidence that both men and women socializing at sites have a high rate of new partner acquisition.
There is some internal evidence that condom use at last coitus was over-reported. Ever use of condoms reported ranged from 33% among women in the East London township to 66% among men in the Port Elizabeth business district, yet a relatively high proportion reported using a condom at last coitus (ranging from 25% to 50%). The small differences between the proportions who had ever used a condom and using one at last sex implies a population divided into never users and consistent users. Other studies have shown, however, that the pattern of condom use is often inconsistent [18]. In addition, of those who reported using a condom at last coitus, only a fourth reported carrying a condom and fewer than 20% of these would show the interviewer the condom.
Although internal consistency checks offer some insight into the extent of bias, self-presentation bias in self-reported sexual behavior data is likely to vary by age, gender, and behavior pattern and is probably impossible to untangle. The PLACE method was designed to minimize reliance on such data. Key program indicators (e.g., condom availability, extent to which youth meet new partners at sites, extent to which people meet new sexual partners at the site) rely primarily on data from key informants and site representatives and only secondarily on data from people socializing at sites.
In all study settings it was considered too dangerous for the interviewers to visit sites and interview patrons at night. The earlier hour of the interview may have led to several biases in terms of male to female ratio, respondent behaviour, unemployment rates, etc. It is worthwhile to consider additional data collection on attendance patterns, alcohol abuse, and sexual partnership formation by on-site staff during the late hours.
Given the limited extent of commercial sex and the high rate of new partner acquisition, these data suggest that large sexual networks drive the HIV epidemic in the townships and that commercial sex currently has a limited role. In the townships, over a fourth of men and women had a new sexual partner within the past 4 weeks. Fewer than 10% of women acknowledged being paid for sex and those who did had a median of two new partners in the past 4 weeks, a much lower rate than the one to seven clients in the past week reported by self-identified sex workers in the four-city study [15].
Epidemiological models of the epidemic suggest that HIV prevalence in a population is the consequence of the pattern of contacts of the entire population rather than of certain individuals [19]. Based on the PLACE results, we estimated that over 4800 new sexual partnerships occurred in the East London township during the 4 weeks prior to the study. This estimate is based on applying the estimated mean rate of new partner acquisition reported by men at small, medium, and large sites to sites of similar size where interviews were not conducted. This does not include partnerships that were not new nor partnerships reported by the women not included in the number reported by the men. Using a similar method, fewer than 70 of these new partnerships involved commercial sex.
A site-based perspective can offer insights into sexual partnership formation and sexual mixing. Figure 2 is a map of sites in one of the townships. Men at smaller sites are more likely to live within a block of the site whereas men at larger sites are more likely to also come from outside the township. Men and women who visit multiple sites are more likely to have a greater number of new sexual partners, and estimating the ratio of men to women at a site provides insight into possible patterns of multiple partnerships not available using methods that are not site-based. Traditional sexual network methods often require individuals to name sexual partners and thus become vulnerable to self-presentation bias. Individuals with many sexual partners may be unable or unwilling to name their sexual contacts and more willing to identify low-risk than high-risk contacts [20-23]. The approach used in this study provides new insights into the context of new sexual partnership formation. Further work to investigate site-based patterns of sexual networks is ongoing.
Clearly, the partner acquisition rates reported at township sites cannot be generalized to the whole township population. The rate of new partner acquisition among women was an order of magnitude higher than that reported by women in household surveys. However, it is important to estimate what proportion of the population participates in these place-based networks and what proportion of the new partner acquisition rate is formed in the places identified by the study. Using the number of people counted at sites and the self-reported rates of attendance, we estimated about 30% of men and 15% of women in the Cape Town township visit at least one site per week. This is likely to vary during a month and over the course of a year, but there is little doubt that socializing at shebeens is a common leisure activity in the townships that involves a significant part of the population. A striking finding is the relatively high proportion of patrons who are female: about one-third of patrons in all four study locations. Better estimates of the proportions and characteristics of the population mixing at these sites will require longitudinal research and complementary data from improved household surveys.
The PLACE studies revealed that AIDS prevention messages and condoms were not available at places where potentially impact could be largest. Site-based interventions are appealing and have the advantage of reaching individuals at a critical time in condom use negotiation. A randomized controlled trial in Nicaragua showed that providing condoms in hotel rooms increased condom use [24]. In the townships, hotels are rare and not an option for condom provision. The last setting prior to sex accessible for prevention efforts in the townships is likely to be the sites described in this study, where people meet their new partners. In addition to offering a site for condom distribution, many sites offer walls for posters, rooms for conducting educational sessions, and video players for AIDS films. In addition, because the clientele of many sites visit at least weekly and often daily, site managers may know patrons well enough to provide a mechanism for education and social support that could be used by prevention as well as care and support programs. More research should be done to evaluate the effectiveness of site-based prevention strategies and how these strategies can best complement other prevention efforts such as school education, workplace education, and mass media campaigns.
We identified three types of priority site from the data: 'youth sites' where youth regularly visit and new partnerships are formed, 'high activity sites' where more than half the men reported a recent new sexual partner; and 'popular sites' reported by many key informants and where attendance exceeds 100. Using these criteria 30-40 priority sites were identified in each study area. Priority sites were more likely to be bars or taverns, allow hard alcohol consumption, and have patrons from outside the study area. Although prioritizing sites for intervention is useful, it must be noted that at least one person reported a recent new sexual partnership in almost all of the sites where interviews were conducted. The findings and recommendations from these assessments were used to develop and inform community-based interventions in the Eastern Cape. This included the development of peer education programs including women who visit sites frequently, improved condom distribution to sites, and improved access to treatment of sexually transmitted diseases.
Because resources for HIV prevention programs and monitoring and evaluation of AIDS prevention interventions are very limited, there is an urgent need for pragmatic methods for guiding public health program in selecting where to focus interventions. Any scale-up of prevention programs requires methods that are feasibly implemented at low cost in a reasonable period of time by non-professional staff. The PLACE approach appears to be such a method. Gaps in AIDS prevention programming were quickly identified as well as priority sites for condom distribution and AIDS education. Although these findings from urban South Africa are not readily generalizable, results from Burkina Faso, India, Mexico, Tanzania, and Uganda are confirming the usefulness of the method in other settings [25-28].
Acknowledgements
Our thanks are extended to C. Gilbert and A. M. M. Rossouw (University of Port Elizabeth, Port Elizabeth, South Africa) for implementation of the protocol in Port Elizabeth; A. S. Davids (Geography Department of the University of Port Elizabeth) and J. Spencer (University of North Carolina) for the maps of the township in Port Elizabeth; M. Pakade (KULA Development Corporation, East London, South Africa) for coordinating the implementation of the protocol in East London; D. Wilson (University of Zimbabwe), S. Mafani (Port Elizabeth), and R. Hegner (East London) for coordination with the intervention programs in East London and Port Elizabeth; C. Morroni (formerly of the University of Cape Town) for coordinating the study in Cape Town; J. Baumgartner (Carolina Population Center, University of North Carolina) for data cleaning and management.
Sponsorship: Funding was provided by USAID under the terms of the MEASURE/Evaluation Cooperative Agreement HRN-A-00-97-00018-00. Funding for data collection in Cape Town was provided by the UNC CFAR 9P30 AI50410.
References
1.Schwartlander B, Garnett G, Walker N, Anderson R. AIDS in a new millennium. Science 2000, 289:64-66.
2.Department of Health, Republic of South Africa. Summary Report National HIV Sero-prevalence Survey of Women Attending Public Antenatal Clinics in South Africa. Pretoria: Department of Health; April 2000:1414.
3.Ainsworth M, Teokul W. Breaking the silence: setting realistic priorities for AIDS control in less-developed countries. Lancet 2000, 356:55-60.
4.Jha P, Nagelkerke JD, Ngugi EN, Prasada R, Willbond B, Moses S, et al. Reducing HIV transmission in developing countries. Science 2001, 292:224-225.
5.Hanenberg RS, Rojanapithayakorn W, Kunasol P, Sokal DC. Impact of Thailand's HIV-control programme as indicated by the decline of sexually transmitted diseases. Lancet 1994, 344: 243-245.
6.Steen R, Vuylsteke B, DeCoito T, Ralepeli R, Fehlet G, Conley J, et al. Evidence of declining STD prevalence in a South African mining community following a core-group intervention. Sex Trans Dis 2000, 27:1-8.
7.Wasserheit JN, Aral SO. The dynamic topology of sexually transmitted disease epidemics: implications for prevention strategies. J Infect Dis 1996, 174(Suppl 2):S201-S213.
8.Mgalla Z, Pool R. Sexual relationships, condom use and risk perception among female bar workers in north-west Tanzania. AIDS Care 1997, 9:407-416.
9.Lamptey P, Kamenga MC, Weir SS. Prevention of sexual transmission of HIV in sub-Sarahan Africa: lessons learned. AIDS 1997, 11(Suppl):S63-S77.
10.Boerma JT, Urassa M, Nnko S, Isingo R, Zaba B, Mwaluko G. Sociodemographic context of the AIDS epidemic in a rural area in Tanzania with a focus on people's mobility and marriage. Sex Trans Inf 2002, 78(Suppl I):97-105.
11.Cleland J, Ferry B. Sexual Behavior and AIDS in the Developing World. London: Taylor and Francis; 1994.
12.Anderson RM. Transmission dynamics of sexually transmitted infections. In Sexually Transmitted Diseases, 3rd edn. Edited by Holmes KK, Sparling PF, Mardh PA, Lemon SM, Stamm WE, Piot P, et al. New York: McGraw-Hill; 1999: ch. 3.
13.Weir SS, Morroni C, Coetzee N, Spencer J, Boerma JT. A pilot study of a rapid assessment method to identify places for AIDS prevention in Cape Town, South Africa. Sex Trans Infect 2002, 78(Suppl I):106-113.
14.Meidany F. HIV and syphilis survey in the Eastern Cape. Eastern Cape Epidemiol Notes, April, 2001.
15.Morison L, Weiss HA, Buve A, Carael M, Abega S-C, Kaona F, et al. Commercial sex and the spread of HIV in four cities in sub-Saharan Africa. AIDS 2001, 15(suppl 4):S61-S69.
16.Behavioral Surveillance Surveys. Guidelines for Repeated Behavioral Surveys in Populations at Risk of HIV. San Francisco, CA: Family Health International; 2000.
17.Medical Research Council, MACRO International Inc, South Africa Department of Health. South Africa Demographic and Health Survey: Preliminary Report. Pretoria: Medical Research Council, MACRO International Inc, South Africa Department of Health; 1998.
18.Weir SS, Roddy RE, Zekeng L, Ryan KA, Wong EL. Measuring condom use: asking 'do you or don't you' isn't enough. AIDS Educ Prev 1998, 10:293-302.
19.Garnett GP, Anderson RM. Sexually transmitted diseases and sexual behavior: insights from mathematical models. J Infect Dis 1996, 174(Suppl 2):S150-S161.
20.Woodhouse DE, Rothenberg RB, Potterat JJ, Darrow WW, Muth SQ, Klovdahl AS, et al. Mapping a social network of heterosexuals at high risk for HIV infection. AIDS 1994, 8:1331-1336.
21.Ghani AC, Donnelly CA, Garnett GP. Sampling biases and missing data in explorations of sexual partner networks for the spread of sexually transmitted diseases. Stat Med. 1998, 17:2079-2097.
22.Coetzee N, Matthews C, McCoy D. Partner notification in the management of sexually transmitted diseases: options for South Africa. S Afr Med J 1996, 86:1478-1479.
23.Coetzee N, Visser H, Mofokeng M, Hennink M. Misses opportunities for partner notification in sexually transmitted disease clinics in Cape Town. S Afr J Epidemiol Infect 1996, 11: 44-47.
24.Egger M, Pauw J, Lopatatzidis A, Medrano D, Paccaud F, Smith GD. Promotion of condom use in a high-risk setting in Nicaragua: a randomized controlled trial. Lancet 2000, 355:2101-2105.
25.Salouka S, Nagot N, Khan M, Brown L, Ganou G, Bidiga J, et al. Rapid assessment methods for prioritizing HIV prevention interventions at the district level: the Burkina Faso experience. XVth International Conference on AIDS. Barcelona, July 2002 [abstract TuPpD2061].
26.Sengooba F, Ssekamatte J, Tate JE, Weir SS. Where are the gaps in Kampala's AIDS prevention program? XVth International Conference on AIDS. Barcelona, July 2002 [abstract TuPeE5211].
27.Patnaik P, Weir SS, Das B, Miller WC, Boerma JT. Assessment of the priorities for local AIDS control efforts (PLACE) method for facilitation of AIDS prevention in India. XVth International Conference on AIDS. Barcelona, July 2002 [abstract TuPeD5037].
28.Negroni M, Bassett Hileman S, Vargas G, Martinez C, Weir S, Bronfman M. Reaching mobile populations for AIDS prevention in southern Mexican border towns. XVth International Conference on AIDS. Barcelona, July 2002 [abstract TuPeE5209].
© 2003 Lippincott Williams & Wilkins, Inc.