Commercial sex and the spread of HIV in four cities in... : AIDS (original) (raw)
Introduction
Female sex workers (FSWs) constitute a group often socially stigmatized and economically disadvantaged, with a high rate of sexual partner change. They are highly vulnerable to infection with HIV and other sexually transmitted infections (STIs), and consequently are at high risk of transmitting HIV to their clients and other sexual partners [1]. Mathematical models have suggested that, in populations where many men have sex with FSWs in addition to other less sexually active women (e.g., their spouses), explosive epidemics of HIV can be expected [2]. However, simulations also suggest that the importance of FSWs in HIV epidemic decreases over time as HIV levels become higher in the general population [3,4]. These hypotheses are based on mathematical models, and there are few empirical data on the relative contribution of commercial sex to HIV incidence in different populations at different stages of the epidemic.
This study was part of a larger study that used standardized cross-sectional surveys of the general population to determine which sexual behaviour patterns and cofactors appeared to be the most important in determining the differential spread of HIV in four cities: two with very high HIV prevalence (Kisumu, Kenya and Ndola, Zambia) and two with relatively low HIV prevalence (Cotonou, Benin and Yaoundé, Cameroon) [5]. In addition to the population surveys in each of the four cities, qualitative and quantitative data were collected on the characteristics and extent of sex work. The aim of this paper was to examine the hypothesis that, in the high HIV prevalence cities, commercial sex transactions were more common and/or that transmission of HIV between sex workers and their clients was more efficient than in the low prevalence cities.
Methods
Data collection
In each of the four cities, qualitative data were collected on the types and characteristics of sex workers in different parts of the city [6]. All the places where sex work was known to occur were mapped and the number of FSWs present (identified by local informants) in each place at the time of the visit was recorded. The number of FSWs counted in this way was used to estimate the total number of FSWs working in each city at any one time.
The surveys took place between September 1997 and May 1998, and we aimed to collect data on around 300 women in each city. The maps and lists already mentioned were used to select samples that were representative in terms of types of sex workers and place of work. In Cotonou, Kisumu and Ndola, the places where FSWs worked were randomly selected and all sex workers present at the time of the team visit were invited to participate. In Yaoundé, the number of sex workers sampled from each zone of the city was proportional to the number of FSWs identified during the mapping. In each zone, sex workers were identified and asked to participate in the study by other sex workers. In Cotonou, brothel and street-based FSWs were deliberately overrepresented in the sample (for the purposes of a separate study), necessitating adjustment during analysis.
Following informed consent, interviewer-administered questionnaires were used to collect demographic and socio-economic data, history of sex work, and information on 'steady partners' and 'clients'. Information gathered on clients included: number in the past 24 h, total number in the past week, number of city-resident clients, number of first-time clients, sexual practices and condom use. FSWs were also invited to attend a specially set-up field unit or a designated health centre for a full genital examination. In Cotonou, there was an initial resistance to attend for genital examination and therefore extra women were included to obtain the required sample size. Blood samples were collected and tested for HIV, syphilis and herpes simplex virus type 2 (HSV-2), and urine was collected and tested for gonorrhoea and chlamydial infection. During the clinical examination, a vaginal swab was taken that was immediately inoculated into a culture medium for Trichomonas vaginalis. Laboratory analysis of specimens was conducted as for the general population surveys [5].
Data from random samples of approximately 1000 men aged 15-49 from the general population in each city were used to examine the extent of contact with sex workers [5]. Men (and women) were asked whether they had had intercourse with any non-spousal partners in the past year. If so, they were asked to provide further details on up to eight non-spousal partners. These details included type of relationship, age of the partner, duration of relationship, exchange of money for sex, and number of partners of the partner. Men were not directly asked whether they had any contact with a sex worker in the past year. In addition, data was analysed from in-depth interviews carried out in Cotonou and Ndola with a random sample of 25 men aged 15-30 years selected from the original sample of men in the survey. These interviews were conducted within a few months of the initial survey in order to test the reliability of the data collected in the main survey.
Statistical methods
Data from the most recent population census for each city [7-10] was extrapolated to estimate the total adult male population aged 15-59 years in 1997. This was combined with data from the mapping of sex workers to estimate the number of sex workers per 1000 males in each city.
All survey data were double-entered and validated using EPI-INFO version 6 (CDC, Atlanta, Georgia, USA). Stata version 6 (Stata Corp, 1999, College Station, Texas, USA) was used for all statistical analysis. To describe characteristics of the sex workers, summary statistics (percentages, means and medians) by city were produced. Weighting was used to adjust for the deliberate over-sampling of brothel and street-based FSWs in Cotonou.
The extent of contact of sex workers with clients was estimated initially using data collected during the survey of sex workers. The annual number of contacts with city-resident clients for each FSW was estimated by multiplying the mean weekly number of city-resident clients by the average number of weeks a FSW was resident in the city. The number of FSW contacts per 1000 men per year was estimated by multiplying the annual number of contacts by the number of FSWs per 1000 men. The number of contacts with clients per 1000 men per year that were unprotected by a condom was also calculated, assuming that the proportion of FSWs using a condom with the last client approximated the proportion of all client-contacts protected by condoms. These should be regarded as minimum estimates because some sex workers did not include clients in the past 24 h in their weekly totals.
Independent estimates of the extent of contact with sex workers were also obtained using the survey of approximately 1000 men in each city. Three definitions of contact with a sex worker were derived based on the characteristics (such as exchange of money) of nonspousal partners of men. These definitions were first used to estimate the percentage of men having sexual contact with a FSW. Two of the definitions were then used to estimate the total number of FSW partners per 1000 males per year. These estimates were multiplied by the average number of times each man reported intercourse with each FSW partner to estimate total sexual contacts per 1000 men per year. For some nonspousal partners, the detailed characteristics were not available, either because there were more than eight non-spousal partners or because the man did not give details for all non-spousal partners. A maximum estimate of extent of contact was therefore calculated by assuming that all these non-spousal relationships were contacts with sex workers.
Results
Table 1 presents the estimated number of FSWs working at any one time within each city and the estimated size of the male population in each city. When these were combined to estimate the number of sex workers per 1000 men, Kisumu and Ndola had higher ratios of sex workers to men than Cotonou and Yaoundé.
Estimated number of female sex workers (FSWs) per 1000 men in each of the four cities
Socio-economic characteristics of sex workers
Sex workers in the low HIV prevalence cities were more likely to be older and to be single than those in the high HIV prevalence cities (Table 2). FSWs in Yaoundé and Kisumu tended to have migrated from within the country, while over one-half of those in Ndola were born in the city. In contrast, 86% of sex workers in Cotonou were migrants from other West African countries (Ghana, Nigeria and Togo). Women in Cotonou started as sex workers at an older age, on average, than in other cities. The median duration of sex work was longest in Yaoundé and Ndola (4 and 3 years, respectively), with a median of only 1 year for Cotonou and Kisumu. A higher proportion of women in Kisumu reported sex work as the main source of income than for the other cities. After converting local currencies to dollars, sex workers in Kisumu and Ndola reported a higher average price per sex act ($5) than in Cotonou and Yaoundé ($3), although this may not reflect differences in socioeconomic status as the cost of living is likely to vary between cities. Cotonou was unique in that almost one-fifth (18%) of FSWs worked in brothels and received a very low price per client ($0.87 compared with around $3 for other types of sex work). Brothel-based work was rare elsewhere, with women meeting their clients in bars, clubs or on the streets. Higher prices per sex act were associated with higher level of education in each city except Yaoundé.
Demographic and socio-economic characteristics of sex workers in the four cities
Sexual behaviour reported by sex workers
A higher proportion of women in Cotonou and Ndola than in Yaoundé and Kisumu reported sexual contact with a client in the previous 24 h (Table 3). In Cotonou, the median number of clients was three among those who had worked in the past 24 h, whereas in the other cities it was one. Sex workers in one of the low HIV prevalence cities, Cotonou, reported the greatest number of clients in the past week (59% had at least five clients), and the highest proportion of new clients. In contrast, sex workers in one of the high HIV prevalence cities, Kisumu, reported the lowest number of clients with 23% reporting no clients in the past week (Table 3). There was little change in reported numbers of clients after excluding the women who reported being off work sick during the past month. Reported condom use with clients was highest in Cotonou, moderate in Kisumu, and lowest in Yaoundé and Ndola. A higher proportion of sex workers in Ndola and Kisumu had refused a client in the past week than in Cotonou and Yaoundé. Refusal to use a condom was more commonly cited as a reason in Cotonou than in the other cities. In each city except Cotonou, around 90% of women reported at least one steady partner, with the highest median number of such partners reported in Yaoundé and Kisumu (Table 3). In Cotonou, only 57% had a steady partner, and one-half of these had not had sexual intercourse with their partner in the previous 4 weeks. There was no indication that high-risk sexual practices such as anal sex, the use of substances to dry or tighten the vagina, or sex during menstruation were more prevalent in the high HIV prevalence cities (Table 3).
Sexual behaviour reported by sex workers: frequency and types of partnerships, estimated number of unprotected contacts, and sexual practices
Prevalence of HIV and other STIs among sex workers
As expected, HIV among sex workers was highest in the two high HIV prevalence cities (69-75%), but the HIV prevalence among sex workers in Cotonou was also high (55%). HIV prevalence was lowest among sex workers in Yaoundé (34%) (Table 4). The prevalence of genital ulceration found on clinical examination was higher in the high HIV prevalence cities. However, after restricting the analysis to HIV-negative women, Kisumu still had the highest levels (20%) and Yaoundé the lowest (0.5%), but the results for Cotonou (10%) and Ndola (7%) were similar. Cotonou had the lowest prevalences of recent/untreated syphilis, gonorrhoea and chlamydia, while prevalences were highest in Ndola and Yaoundé. The prevalence of recent/untreated syphilis in Ndola (42%) and chlamydia in Yaoundé (18%) were particularly high. Trichomoniasis was markedly more common among sex workers in the two high HIV prevalence cities. Higher proportions of sex workers in Yaoundé and Ndola reported a STI episode in the past month than in the other cities, but only one-half of the affected women sought modern treatment. Around 80% of sex workers in Cotonou and Kisumu who reported a STI episode had sought modern treatment.
Prevalence (%) of HIV, genital ulcers and other sexually transmitted diseases among sex workers
Extent of contact with sex workers
Sexual behaviour data from the population surveys of adult men were used to try and identify which had been clients of a sex worker over the past year. Men rarely described non-spousal partners as prostitutes (between 0.4 and 1.8% in each city reported a prostitute as a sexual partner), limiting the use of this in identifying clients. In each city, money was reported to be exchanged in 16-44% of non-spousal relationships not described as sex work (and including fiancées) and was exchanged only in 25-72% of relationships described as sex work. It was therefore considered not to be a useful identifier of sex work contact if used alone, and was used instead in combination with other characteristics of the relationship. The following three definitions of contact with a FSW were used.
A. Male partner described a partner as a prostitute when asked what type of relationship it was.
B. Money was always or often exchanged and the partner was reported to have 10 or more other partners in the past year. However, between 21 and 58% of responses for the question on partners of the partner were missing or the man reported that he did not know.
C. Male partner described a partner as a prostitute, or it was a relationship where money was always or often exchanged and any one of the following: duration of relationship 1 day or less; sex on the same day they met; female partner was reported to have 10 or more partners; or female partner reported to exchange sex for money with others.
Table 5 presents the percentage of men identified as having contact with a sex worker in the past year for each of the three definitions used. By all three definitions, the proportion of men having contact with a FSW was greatest in Yaoundé, next highest in Ndola, and lowest in Cotonou and Kisumu.
The percentage of men defined as having contact with sex workers during the past year for each city
The number of sex work contacts per 1000 men per year was calculated based on the latter two definitions of sex work contact for men and on client data from the survey of sex workers (Table 6). Regardless of the method of computation, Ndola and Yaoundé have a much higher estimated number of contacts per 1000 men per year than Kisumu. Using the data from the sex workers Cotonou has a very high number of contacts, but high levels of condom use offset this bringing numbers of unprotected contacts below those in Ndola and Yaoundé.
Estimated contacts per 1000 men per year with sex workers using data provided by the male population and by sex workers
Comparing estimates of client contacts per 1000 men per year between the data from the sex workers and the data from the male population, rough agreement is seen for Yaoundé, substantial discrepancies for Ndola and Kisumu, and a very large discrepancy for Cotonou. Where there are discrepancies, the estimates based on the data from the male population are lower than those based on reports of the sex workers.
Qualitative data for Cotonou and Ndola from subsamples of men re-interviewed after the survey showed high levels of contact with FSWs (9/32 = 28% in the past year for Cotonou, and 5/25 = 20% over a lifetime in Ndola). These results suggest a higher level of contact with sex workers than that suggested by the survey of men.
Discussion
Depending on the context, it may be difficult to define and differentiate professional sex work from the range of relationships where the exchange of money is common. Methods of identifying and sampling sex workers will therefore have influenced the characteristics of our samples. For the mapping, some types of FSW and some establishments may have been missed, resulting in minimal totals and possible bias. For the survey, only women who self-identified as sex workers participated. Occasional sex workers (for example, market women in Kisumu who occasionally sell sex) and higher class sex workers were less likely to be included in our survey [6]. Other sources of error may have affected the calculation of the extent of sex work based on the reports of the FSW. The population censuses for Yaoundé and Kisumu that were used to estimate the size of the male population were very out of date. Also, it was probably an oversimplification to multiply the weekly number of clients by the number of weeks the sex worker was present in the city to calculate the annual number of contacts. However, unless the overall effect of these possible biases varied greatly between the four cities, the patterns of estimates over the four cities should be meaningful.
The difficulty in differentiating commercial sexual transactions from other types of relationship was also apparent when we attempted to estimate rates of contact with sex workers using questionnaire data from the male population. In local terms, very few men described any of their partners as prostitutes and we therefore attempted to identify clients of sex workers using characteristics of non-spousal partners. Exchange of money was found to be common in many types of non-spousal relationships, including those engaged to be married, and therefore more complex definitions of sex work contact were derived. The percentage of men identified as clients varied greatly depending on the definition used (although the broad patterns by city remained similar). In addition, estimates of the extent of sex work based on these definitions were substantially lower than those based on reports of sex workers for all cities except Yaoundé. There are several possible explanations for these discrepancies. The differing age ranges on which the estimates are based [population census data (15-59 years) and the survey data (15-49 years)] may have affected the discrepancies, but this is unlikely to be substantial. Another possible explanation is that men who have contact with FSWs might be less likely to be found at home and hence to participate in the study. However, in Cotonou where the discrepancy was greatest, participation rates were 95% for men, making this an unlikely explanation. Another possibility is that clients might be from high-risk groups (e.g., military or truck drivers) who would not be included in a household study. However, in Cotonou, a recent study of clients of sex workers found no military personnel among them and only 15% were truck drivers [11] so this also seems an unlikely explanation for the discrepancies. The lack of any other explanation and the high levels of contact with sex workers in Cotonou and Ndola found in the qualitative data suggest that under-reporting is the most likely explanation for the discrepancies. Under-reporting by men could occur because of embarrassment at having sex with a sex worker, because very casual contacts might not be remembered, or because a contact with a sex worker was not considered to be a 'partnership'. The best way of ascertaining contact with a sex worker in population surveys of males needs to be addressed for future studies.
The aim of this paper was to use standardized empirical data to examine whether present patterns of commercial sex transactions might give some indication of how sex work has influenced the spread of HIV in the four cities. Specifically, we examined whether commercial sex transactions were more common and/or whether transmission of HIV between sex workers and their clients was more efficient in the high compared with the low HIV prevalence cities. The estimated number of sex workers per 1000 men was higher in the cities with high HIV prevalence. However, when the estimated number of clients per sex worker was taken into account, Cotonou and Ndola had the highest number of contacts per 1000 men per year. When considering contacts unprotected by a condom, Ndola and Yaoundé had the highest number. Therefore, there were no clear differences in the extent of sex work between the high and low HIV prevalence cities. Considering differences in transmission probabilities, examination of data from clinical examination showed that genital ulceration was lower in the low HIV prevalence cities. However, as STI and HIV mutually enhance each other, we examined rates in HIV-negative women only. When this examination was carried out, the original pattern for ulceration disappeared with similar prevalences being observed in Ndola and Cotonou. The laboratory method we used results in underestimates for gonorrhoea [12] but the rate of underestimation would be similar in all four cities. Recent/untreated syphilis, gonorrhoea and chlamydia were not obviously lower among sex workers in the low prevalence cities. They tended to be higher in Yaoundé and Ndola where the proportion of women seeking modern treatment for a STI episode was lowest. The markedly high prevalence of syphilis seen in sex workers in Ndola and chlamydia seen in sex workers in Yaoundé mirror the pattern in the general population [12]. Trichomoniasis followed a different pattern being higher in the high HIV prevalence cities, again mirroring patterns among women in the general population [13]. The role of trichomoniasis in HIV transmission needs further research. In summary, apart from trichomoniasis and perhaps observed genital ulceration, no other STI were found more commonly in the high prevalence cities. Differences between the high and low prevalence cities in sexual practices that might affect transmission probabilities, such as anal or dry sex, were also not obvious. There were some differences in sociodemographic characteristics between high and low HIV prevalence cities. Sex workers in the low prevalence cities tended to be older and were more likely to be single, but it is not clear how this in itself would affect the role of sex work in driving the HIV epidemics. Condom use among sex workers was highest in Cotonou but joint lowest in Yaoundé and Ndola, again not showing a simple association with high and low HIV prevalence in the four cities. In conclusion, at the time of the study, neither the extent of sex work nor factors affecting transmission showed clear differences between the high and low HIV prevalence cities.
The lack of clear differences between the high and low HIV prevalence cities does not rule out the possibility that differences in patterns of sex work between the four cities were important in driving the different HIV epidemics. Changes over time may have occurred related to the HIV epidemic and/or other factors. For example, mortality and illness due to HIV/AIDS might reduce numbers of FSWs while increased dependence of large numbers of women on sex work for money (if a large proportion of husbands die of AIDS) might increase their numbers. Similarly, in the later stages of the epidemics when the effects of AIDS are more obvious, men might reduce their contacts with sex workers, and both sex workers and clients might be more likely to use condoms. It is impossible now to fully assess whether and how the patterns of sex work might have changed over time. However, given the limitation of possible changes over time, our data do not suggest any obvious differences in the characteristics of sex work in the four cities that could explain the differential spread of HIV within them.
The simple comparison of factors between the high and low prevalence cities described does not demonstrate a greater extent of contact with sex workers or higher transmission probabilities in the high prevalence cities. However, an examination of combinations of factors within each city does give some clues as to how sex work is influencing the HIV epidemic in each one.
The group of sex workers in Cotonou was the smallest relative to the population size and was characterized by the highest number of clients (with two-thirds being new clients in any week) of any of the four cities. Also, there was a substantial proportion of brothel-based workers (18%), whereas brothel-based work was rare in the other cities. This combination of low numbers of FSWs per 1000 men and the high number of clients may explain the relatively high HIV prevalence among sex workers in Cotonou (55%) given the relatively low HIV prevalence in the male population (3%). Another possible explanation is that sex workers are infected before they come to Cotonou (86% migrate in from other countries). Recent/untreated syphilis, gonorrhoea, chlamydia and trichomoniasis were lower in sex workers in Cotonou than in sex workers in the other cities, perhaps reflecting high levels of reported condom use (69% with last client) and high proportions seeking treatment for STI episodes (80%). Transmission of HIV from sex workers to the general population may be lower than in the other cities because of lower levels of STI in the FSWs and the high level of condom use. The fact that virtually all males in Cotonou are circumcised is also likely to reduce transmission from sex workers to men [14,15].
Ndola was characterized by high numbers of sex workers per 1000 men and high levels of contact with FSWs per 1000 men per year compared with the other cities. The FSWs also tended to have worked for a relatively long time, to have high numbers of clients (many new) and to have high numbers of steady partners. The FSWs in Ndola also reported low levels of condom use, had the highest levels of recent/untreated syphilis and had relatively high levels of ulceration. It was not therefore unexpected that they had very high levels of HIV infection. The lack of condom use, high levels of genital ulceration among the FSWs and low levels of circumcision among the male population are likely to be responsible for a high HIV transmission probability to clients.
Reported patterns of sex work in Yaoundé were similar to those in Ndola in some ways (long-term workers, high numbers of contacts and low condom use), but not others. The FSWs in Yaoundé tended to have more regular clients and had a very low prevalence of genital ulceration on clinical examination (2%). The low prevalence of genital ulceration might, however, be due to method of ascertainment as relatively high levels of recent/untreated syphilis were found in this study and a higher prevalence of genital ulcer (15%) was reported in another study [16]. The almost universal circumcision of males in Yaoundé is another interesting difference from Ndola and may help to explain differences in the HIV prevalence between the two cities.
Kisumu was characterized by a relatively high number of FSWs per 1000 men, many widowed or divorced FSWs, and FSWs highly dependent on income from sex work. At the time of the survey, they reported relatively few clients (many regular) and more contacts with men they classed as steady partners than with clients. Condom use was moderate, with 50% of women reporting condom use with their last client. A notable feature of the FSWs in Kisumu was the high level of ulceration seen, and this has also been reported by other researchers [17]. The low prevalence of recent/untreated syphilis suggests that other causes for the ulceration are more important including Haemophilus ducreyi, genital herpes and possibly HIV infection itself. High levels of ulceration and the low prevalence of male circumcision are likely to lead to high transmission probabilities from FSWs to men.
Mathematical modelling suggests that, although the importance of sex work decreases as HIV prevalence rises in the general population [3,4], interventions targeted at sex workers and their clients are still relatively effective because of the large number of sexual contacts of sex workers [18]. Interventions consisting mainly of improved access to STI treatment and/or encouraging the use of condoms among sex workers and clients have been shown to reduce HIV incidence among sex workers, and are thus presumed to have prevented cases in the general population [19,20]. STI prevalence among miners in South Africa was found to decrease following an intervention among local sex workers [21]. At a national level, the campaign in Thailand to promote condom use among sex workers seems likely to have reversed a major epidemic there [22]. The high reported levels of condom use by sex workers in Cotonou described in this paper, which are the result of interventions targeting sex workers since the early 1990s, are likely to have had an impact on HIV transmission to the general population in that city.
Conclusions
This paper provides standardized, cross-sectional data on the characteristics and extent of sex work in four African cities. No clear patterns between the high and low HIV prevalence cities were found in terms of STI prevalence (apart from trichomoniasis), sexual practices, condom use or extent of contact with sex workers. Thus, there are no obvious differences in present patterns of sex work that might explain the different spread of HIV in the high and low prevalence cities. However, combinations of factors relating to sex workers and their clients give some hint as to how commercial sex may have contributed to the epidemics in each city. The high rate of partner change of FSWs and their high levels of STIs and HIV underline the importance of interventions among sex workers and their clients. High levels of condom use among sex workers in Cotonou are likely to have slowed the epidemic there.
References
1. Plummer FA, Nagelkerke NJ, Moses S, Ndinya-Achola JO, Bwayo J, Ngugi E. The importance of core groups in the epidemiology and control of HIV-1 infection.AIDS 1991, **5 (suppl 1):**S169-S176.
2. Anderson RM, May RM, Boily MC, Garnett GP, Rowley JT. The spread of HIV-1 in Africa: sexual contact patterns and the predicted demographic impact of AIDS.Nature 1991, **352:**581-589.
3. Robinson NJ, Mulder D, Auvert B, Whitworth J, Hayes R. Type of partnership and heterosexual spread of HIV infection in rural Uganda: results from simulation modelling.Int J STD AIDS 1999, **10:**718-725.
4. Kamenga M, Denison J, White R, MacNeil J, Sweat M. Targeting interventions for the prevention of HIV infection in developing countries: when to begin targeting non-core groups. Paper presented at the 3rd USAID HIV/AIDS Prevention Conference. Washington, DC, 7-9 August 1995.
5. Buvé A, Caraël M, Hayes RJ, et al.Multicentre study on factors determining differences in rate of spread of HIV in sub-Saharan Africa: methods and prevalence of HIV infection.AIDS 2001, **15 (suppl 4):**S5-S14.
6. Huygens P, Carael M. Multicentre study on determinants of differences in HIV prevalence: anthropological assessment.UNAIDS/PSR Research Report 1997. Geneva, Switzerland; 1997.
7. National Institute of Statistics and Economic Analysis. Volume II: analysis of results, parts 1-4. In Second National Population Census of Benin 1992. Ministry of Planning, National Institute of Statistics and Economic Analysis, Benin; 1994.
8. Central Statistical Office. Part 3, volume III: population structure by age and sex (December 1992); part 3, volume II: Central Province (April 1993); part 12, volume II: towns of Douala and Yaoundé. In Second National Population Census of Cameroon 1987. Central Statistical Office, Cameroon; May 1993.
9. Central Bureau of Statistics. Kenya Population Census 1989, Vols 1 and 2. Office of the Vice-President and Ministry of Planning and National Development, Central Bureau of Statistics, Nairobi, Kenya; March 1994.
10. Central Statistical Office. Descriptive tables, volume 2: Copperbelt Province. In Census of Population, Housing and Agriculture 1990. Central Statistical Office, Demography Division, Zambia; 1994.
11. Lowndes CL, Alary M, Gnintoungbé CAB, et al.Management of sexually transmitted diseases and HIV prevention in men at high risk: targeting clients and non-paying sexual partners of female sex workers in Benin.AIDS 2000; **14:**2523-2534.
12. Buvé A, Weiss HA, Laga M, et al.The epidemiology of gonorrhoea, chlamydial infection and syphilis in four African cities. AIDS 2001, **15 (suppl 4):**S79-S88.
13. Buvé A, Weiss HA, Laga M, et al.The epidemiology of trichominiasis in women in four African cities.AIDS 2001, **15 (suppl 4):**S89-S96.
14. Auvert B, Buvé A, Lagarde E, et al.Male circumcision and HIV infection in four towns in sub-Saharan Africa.AIDS 2001, **15 (suppl 4):**S31-S40.
15. Weiss HA, Quigley MA, Hayes RJ. Male circumcision and risk of HIV infection in sub-Saharan Africa: a systematic overview and meta-analysis.AIDS 2000, **14:**2361-2370.
16. Roddy RE, Zekeng L, Ryan KA, Tamoufe U, Weir SS, Wong EL. A controlled trial of nonoxynol-9 film to reduce male-female transmission of sexually transmitted diseases.N Engl J Med 1998, **339:**504-510.
17. Kaul R, Kimani J, Nagelkerke NJD, et al.Risk factors for genital ulceration in Kenyan sex workers.Sex Transm Dis 1997, **24:**387-392.
18. Van Vliet C, Holmes K, Singer B, Habbema D. Effectiveness of HIV prevention strategies under alternate epidemiological scenarios: evaluation with the STDSIM model. In Confronting AIDS: Evidence from the Developing World. Selected Background Papers. Belgium: World Bank, Washington and European Commission; 1998:39-51.
19. Laga M, Alary M, Nzila N, et al.Condom promotion, sexually transmitted diseases treatment and declining incidence of HIV-1 infection in female Zairian sex workers.Lancet 1994, **344:**246-248.
20. Ngugi EN, Wilson D, Sebstrad J, Plummer FA, Moses S. Focussed peer-mediated educational programs among female sex workers to reduce sexually transmitted disease and human immumodeficiency virus transmission in Kenya and Zimbabwe.J Infect Dis 1996, **174 (suppl 2):**s240-s247.
21. Steen R, Vuylsteke B, DeCoito T, et al.Evidence of declining STD prevalence in a South African mining community following a core-group intervention.Sex Transm Dis 2000, **27:**1-8.
22. Hanenberg RS, Rojanapithayakorn W, Kunasoi P, Sokal DC. Impact of Thailand's HIV control programmed as indicated by the decline of sexually transmitted diseases.Lancet 1994, **344:**243-245.
**Section Description
This publication is sponsored by the Joint United Nations Programme on HIV/AIDS (UNAIDS) and the Agence Nationale de Recherches sur le SIDA (ANRS) Paris, France. The Editors of this supplement wish to acknowledge the referees who provided peer review of the manuscripts.
The study was supported by the following organizations: UNAIDS, Geneva, Switzerland; European Commission, Directorate General XII, Brussels, Belgium; Agence Nationale de Recherches sur le SIDA/Ministère français de la coopération, Paris, France; DFID, London UK; The Rockefeller Foundation, New York, USA; SIDACTION, Paris, France; Fonds voor Wetenschappelijk Onderzoek, Brussels, Belgium; Glaxo Wellcome, London, UK; and BADC, Belgium Development Cooperation, Nairobi, Kenya.
Keywords:
Sex workers; prostitutes; heterosexual transmission; Africa
© 2001 Lippincott Williams & Wilkins, Inc.