JAIDS Journal of Acquired Immune Deficiency Syndromes:
Epidemiology and Social Science
Changing Behaviors and Patterns Among Cambodian Sex Workers: 1997-2003
Gorbach, Pamina M. MHS, DrPH; Sopheab, Heng MD, MPH; Chhorvann, Chhea MD, MPH; Weiss, Robert E. PhD; Vun, Mean Chhi MD, MPH
From the *Department of Epidemiology, School of Public Health, University of California, Los Angeles, CA; †National Center for HIV, AIDS, Dermatology, and STDs, Ministry of Health, Phnom Penh, Cambodia; and ‡Department of Biostatistics, School of Public Health, University of California, Los Angeles, CA.
Reprints: Pamina M. Gorbach, Department of Epidemiology, School of Public Health, University of California, Box 951772, Los Angeles, Los Angeles, CA 900095-1772. E-mail: firstname.lastname@example.org.
This work was supported by Cambodia's Ministry of Health, the United States Agency for International Development (USAID), and Family Health International. The content of this article does not necessarily reflect the views or policies of USAID.
Objective: Identify patterns and behaviors among direct and indirect female sex workers (DFSWs and IDFSWs, respectively) across Cambodia's 5 major cities from 1997 to 2003.
Methods: Interviews with DFSWs and IDFSWs followed random selection from clusters in 5 cities. Individual characteristics and condom use with clients and other partners were assessed in univariate and multivariate analyses.
Results: From 1997 to 2003, consistent condom use with clients increased from 53% to 96% among DFSWs and from 30% to 84% among IDFSWs. DFSWs reported staying in their profession longer, had fewer clients per day, stayed longer in each brothel, were in increasingly larger brothels, and were tested more for HIV. For IDFSWs, there were significant changes: more reported practicing commercial sex and testing for HIV. In adjusted models, reported condom use with clients was significantly higher among DFSWs in later survey years (odds ratio [OR], 2.17) and who were never married (OR, 1.69), were in larger brothels (OR, 1.02), and charged more for sex (OR, 1.27), but lower for DFSWs with sweethearts (OR, 0.68) and who reported abnormal vaginal discharge (OR, 0.52). For IDFSWs, in the adjusted models, reported condom use with clients was higher in later years (OR, 1.77) and for those reporting abnormal vaginal discharge (OR, 1.34) and HIV testing (OR, 1.46), and lower for those with sweethearts (OR, 0.49).
Conclusions: From 1997 to 2003, Cambodian direct and indirect sex workers increased their use of condoms each year with commercial as well as noncommercial partners, contributing to the evidence that HIV prevention programs can produce significant changes in risk behaviors.
Although Cambodia has experienced one of Asia's most severe epidemics of HIV/AIDS,1 it now seems to be on the decline in risk groups.2 The Cambodian National Center for HIV, AIDS, Dermatology, and STDs (NCHADS) HIV Sentinel Surveillance (HSS) reported HIV prevalence dropping in brothel-based direct female sex workers (DFSWs) from around 42.6% in 1995-1998 to 20.8% in 20033. This recent decline did not occur before the Cambodian epidemic became "generalized" as defined by the World Health Organization4 because HIV prevalence in pregnant women reached 2.2% in 2003.3 Although some of the recent decline in HIV prevalence reported in risk groups may be attributed to increasing AIDS mortality, the magnitude of the drop in prevalence could only have been possible if mirrored by a rapid change in risk behavior that precipitated decreasing incidence. Such behavior change is probable in Cambodia because a large-scale multisectoral HIV/AIDS prevention program that targeted core groups such as sex workers was initiated in 1995.1
Analyses of Cambodia's behavioral surveillance survey (BSS), a cross-sectional survey of many of the same risk groups as the HSS repeated annually since 1997, provide evidence of behavior change. BSS provides data on social conditions and sexual behaviors that drive HIV epidemics such as "bridging" behavior between high- and low-risk sexual networks5. Moreover, BSS is a source of comprehensive data on sex work, tracking changing conditions of sex work that influence whether DFSWs adopt safer sex behaviors. It also provides estimates of levels of risk behavior in commercial sex. The BSS collects information on brothel conditions as well as individual behaviors of sex workers, expanding the data available on sex work. Other studies on commercial sex in Southeast Asia have focused primarily on individual risk factors for HIV and anthropological descriptions of sex work6-8 and have shown that brothels are particularly high-risk settings. For example, brothel-based DFSWs are at higher risk of becoming infected with HIV compared with non-brothel-based sex workers in Thailand9 and in Vietnam.10 Working conditions11 and migration patterns12 have been described for Thai brothel-based DFSWs; however, the influence of such factors on sex workers' behaviors has not been well studied. In addition, longitudinal studies of sex workers' behaviors have assessed changes in condom use during commercial sex13 but not changes in conditions of sex workers or partner type. Cambodia's BSS includes such additional data.
The BSS was designed after formative research on Cambodian sexual behavior14 from which specific occupational groups were selected to be annually surveyed. Two high-risk groups of women were selected for inclusion. One group was brothel-based DFSWs selected because of their large number of sexual partners, high daily sexual activity, the Cambodian sex industry's organization around brothels, DFWs' high HIV prevalence reported in the HSS in the late 1990s, and their high prevalence of sexually transmitted infections (STIs) including gonorrhea or chlamydial infection identified in 1997.15 Women employed by beer companies to work in restaurants and bars as beer promoters were selected as "indirect female sex workers" (IDFSWs). These women have many sexual partners, including some commercial partners, suggesting that they are equivalent to the IDFSWs in Thailand who have a main occupation such as working in massage parlors but who also sell sex.13 The prevalence of HIV among Cambodian IDFSWs has remained relatively high and stable over time (20% in 1998 and 14.8% in 200216). Three groups of men were also selected but were not included in the succeeding analyses.
The following analyses of Cambodia's BSS address the need for greater contextual analysis to enhance the interpretation of surveillance data.17 The relationship between conditions of sex work, including brothel size and types of sexual partners, and condom use during commercial sex during the last 5 years was analyzed to enhance interpretation of changing patterns in HIV prevalence.
The BSS has been conducted annually in Cambodia's 5 largest cities (Phnom Penh, Battambang, Sihanoukville, Siem Reap, and Kampong Cham) since 1997. The cluster design used in other BSS surveys18 was used to sample each occupational group. Clusters (defined as naturally occurring units) were identified for each occupational group: brothels for DFSWs and beer companies for IDFSWs. Lists of all clusters and the number of individuals in each cluster were then generated in each city. Clusters were randomly selected from each list, and all members in each cluster were interviewed face to face by female interviewers with privacy ensured by NCHADS supervisors. The questionnaire included behaviors associated with STIs and HIV in 199615 and used in the Thai BSS for the following different types of partners: client, regular client (ie, a paying partner that has returned 5 or more times), and a noncommercial partner (known colloquially as a "sweetheart"). All questionnaires were checked in the field to improve data quality; STATA version 8.2 software was used for analysis. Oral informed consent was obtained from all participants who remained anonymous. The sample sizes in the Cambodia BSS I, II, III, V, and VI were 245, 804, 792, 569, and 822 for DFSWs1 and 581, 406, 379, 402, and 513 for IDFSWs, respectively. The surveys were conducted at 12-month intervals among rounds I, II, and III from 1997 to 1999 and then 16 months for BSS V and another 21 months until the BSS VI (BSS IV was a household survey, so it is not included in this analysis). The first round of BSS had a smaller sample of DFSWs because data were not collected in Battambang, Sihanoukville, or Kampong Cham because other studies were being conducted on female sex workers in those cities at that time.
Both DFSWs and IDFSWs were asked about condom use with paying partners (clients) and noncommercial partners (sweethearts). Associations between consistent condom use in commercial sex within each occupational group (in the past week for DFSWs and in the past 3 months for IDFSWs) and individual characteristics, partner type, and sex industry dynamics across survey years were tested for statistical significance using univariate and multivariate logistic regression and, for the latter, were stratified by the 5 cities using survey-based estimation models. The data sets from all 5 BSS years were concatenated into 1 data set with survey year included as a variable.
Table 1 tests used a score test for linear time trend for discrete variables and a generalized Kruskal nonparametric test of trend for continuous variables. Odds ratios (ORs) and tests in Table 2 are for the main effect of the covariate from a logistic regression predicting 100% condom use adjusted for year of survey. Multiple regression models of factors affecting the probability of always using condoms during commercial sex were stratified by city and include years since first survey as a continuous variable reflecting the time between surveys. BSS I serves as the reference year 0, BSS II is year 1, BSS III is year 2, BSS V is 3.3, and BSS VI is 5.1.
From 1998 to 2003, DFSWs reported a large increase in consistent condom use with clients, from 53.4% in 1998, to 78.1% in 1999, 91.7% in 2001, and 96% in 2003 (P < 0.0001). IDFSWs reported an even greater increase in consistent condom use with clients, from 29.7% in 1998 to 38.2% in 1999, 55.2% in 2001, and 84.4% in 2003 (P < 0.0001) (an increase of 55 percentage points). Figure 1 illustrates how condom use increased for both DFSWs and IDFSWs more with commercial partners than with sweethearts.
Across the 6 years (1997-2003), sex workers became older and conditions of sex work changed: DFSWs reported staying in their profession longer, stayed longer in each brothel, size of brothels increased, and more were tested for HIV (Table 1). For IDFSWs, significantly more reported practicing commercial sex in the past year (from 21% to 37%) and many more reported getting tested for HIV (from 21% to 70%) (Table 1). Finally, since 1998, respondents were asked if they know someone sick with HIV/AIDS, and these numbers increased considerably each year, from 7% to 64% among DFSWs and from 9% to 69% among IDFSWs (Table 1).
Condom use increased with both partner types over the 6 survey years (Fig. 1). DFSWs who worked longer in sex work, lived in larger brothels, had been in a particular brothel longer, did not have a sweetheart in the past year, never had been married, did not have abnormal vaginal discharge in the past year, and had been tested for HIV were significantly more likely to have reported always using condoms with clients in the past week than those who did not always use condoms (Table 2). DFSWs who reported a sweetheart in the past year compared with DFSWs who did not have such a partner were younger (22.3 vs 21.9 years, P = 0.008), had more clients in the last day (3.4 vs 2.7 men, P < 0.0001), were in smaller brothels (7.4 vs 8 other women, P = 0.004), spent a longer duration in that brothel (7.5 vs 6.7 months, P = 0.004), and had more years in sex work (1.8 vs 1.4 years, P < 0.0001).
As with DFSWs, with each successive survey, significantly more IDFSWs reported consistent condom use (P < 0.000) with both commercial and noncommercial partners (Fig. 1). Among IDFSWs reporting sex for money, those who were slightly older, had started sex later, had been sexually active longer, did not have a sweetheart in the past year, or had sex with a sweetheart in the past year, had vaginal discharge, were not from Battambang, and had an HIV test were more likely to report always using condoms with clients in the past 3 months than IDFSWs who did not always use condoms with clients (Table 2). Like DFSWs, those IDFSWs with sweethearts were more likely to be younger (22.4 vs 23.1 years) than those without sweethearts.
In the multiple regression models, DFSWs reported condom use with clients was significantly higher during later survey years (OR, 2.29 per year or per 4 years, as appropriate) and among those who were in larger brothels (OR, 1.02), and charged more for sex (OR, 1.29), but was lower for DFSWs with sweethearts (OR, 0.68) and those who reported abnormal vaginal discharge (OR, 0.50). In the adjusted models for IDFSWs, reported condom use with clients was higher in later years (OR, 1.80 per year or per 4 years as appropriate), among those who were tested for HIV (OR, 1.44) and lower for those with sweethearts (OR, 0.48) (Table 3).
Cambodia started an intensive monitoring and prevention program early in its HIV epidemic when the first evidence of high HIV prevalence among core groups became available. These findings demonstrate that, with each year, consistent condom use during commercial sex has increased among women working in the sex industry, that is, increases in condom use with clients from 53% to 96% for DFSWs and 30% to 84% for IDFSWs, suggesting intervention efforts by NCHADS, the Cambodian Ministry of Health, and nongovernmental organizations seem to have been effective especially in the groups practicing the most direct commercial sex. Continuing declines in HIV prevalence reported in the HSS further support this reported behavior change. Levels of behavior change can be compared with the first 3 years of Thailand's BSS (1993-1996), where increases in condom use during commercial sex among DFSWs were significant although small, 87% to 97%, and for IDFSWs, 56% to 89%.13 In addition, more sex workers reported consistent condom use with all types of partners: clients, regular clients, and even noncommercial partners (sweethearts).
These findings suggest that the Cambodian sex industry is in transition; sex workers seem to be staying longer in sex work and longer in each brothel and are older than 5 years earlier. Outside of the brothels, more IDFSWs are reporting that they sell sex. Such changes may translate into increasing competition among DFSWs for clients and increasing pressure among them to maintain clients. Many DFSWs report regular clients and noncommercial partners, and the DFSWs with the latter were less likely to use condoms consistently with their clients who are concurrent, perhaps as an attempt to convert them into regular or relational partners. These DFSWs may be concerned about aging and being outcompeted by younger women who can charge more for sex, which makes condom use more difficult to negotiate for older ones. DFSWs seeking a way out of sex work through a single partner may be cultivate regular clients and noncommercial partners (sweethearts) while they can still compete for such partners. Therefore, sex workers may be changing partners to change their behavior; a change in condom use behavior may represent the prelude to an exit out of sex work because decreasing condom use may be an attempt to seek and keep the partners who will help women leave the sex industry because sex without condoms is often interpreted as an expression of intimacy and trust. The findings also suggest that among IDFSWs, noncommercial, casual sex is on the rise.
DFSWs may be moving out of brothels that increasingly carry a stigma of high HIV risk into venues such as karaoke bars or massage parlors where women work as IDFSWs under the guise of another profession. The finding that more IDFSWs reported selling sex each year may also be a sign of an increasing market for IDFSWs or a greater willingness among this group to disclose their risky behavior. IDFSWs represent a vulnerable group of women who need more aggressive interventions that are designed specifically for them regarding behavior change.
Other measures of the conditions of sex work suggest the potential for structural/environmental level change to influence individual sex worker behaviors. Women from larger brothels and those who charge more for sex are more likely to use condoms. These larger and more expensive brothels may be better managed and therefore may have personnel who enforce condom use, serving as backup for individual sex workers' requests for clients to use condom. Earlier research on brothel size in Cambodia suggested larger brothels were safer environments for women because lower HIV prevalence was noted in the larger ones compared with that in smaller ones early in the Cambodian HIV epidemic.19
DFSWs and IDFSWs who reported abnormal vaginal discharge in the past year were less likely to report condom use with clients. Although such discharge may have been a symptom of an episode of an STI and DFSWs infected with STIs would be more likely to transmit or acquire HIV, self-reports of discharge are poor markers of STI. Such discharge could also be a symptom of being HIV positive because women with HIV are more likely to experience vaginal discharge20 caused by both STI and non-STI pathogens. Either case suggests such women may represent high transmission risks, and that they are less likely to consequently use condoms suggests they are putting their partners at possible exacerbated risk of infection. Moreover, these symptoms may suggest that their lack of condom use may already have had consequences for them such as acquiring an STI and may represent their risk for acquiring HIV if they have not done so already, suggesting a need for STI services for DFSWs.
That IDFSWs who report being tested for HIV are more likely to use condoms during commercial sex than those not tested suggests a strong potential for HIV voluntary counseling and testing programs. Women who test for HIV individually are less likely to reduce unprotected sex with a primary or nonregular partner than women who test as part of a couple21; nevertheless, women who are at particular risk may be more likely to change their behavior than those in the general population. Although women who test for HIV may be those already predisposed to higher condom use because they tend to practice protective behaviors in general, the potential for voluntary counseling and testing as a strategy for behavior change among female sex workers in Cambodia needs exploration.
Cambodia's BSS provides evidence that HIV/AIDS prevention efforts have been successful and are rapidly transforming the risk profile of a population, potentially stalling the spread of HIV in the country. In response to behavior change programs for sex workers, increases in condom use have also been noted in Abidjan, Cote d'Ivoire,22 and Cotonou, Benin23 concurrent with reductions in STD and HIV prevalence, and much of the success of Thailand's decrease in HIV prevalence has been attributed to sex worker-focused interventions such as the 100% condom program.24 It should be mentioned that Cambodia has gone beyond behavior change in its HIV prevention efforts by implementing programs in sexually transmitted infection screening and treatment, enforcement of 100% condom policies in brothels, and expanded HIV testing, and such programs have most likely contributed as well to stemming the course of their HIV epidemic. In the face of increasing pleas for biologic prevention, Cambodia contributes to the growing body of evidence that the course of HIV epidemics can be altered by effective behavior change programs, and resources must be maintained to sustain such efforts.
We thank Hor Bun Leng for his support and Mun Phalkun, Seng Sopheata, Kao Chantha, and Theng Thithara of the Cambodian Ministry of Health and the NCHADS for their contribution to data collection and management.
1. Phalla TLH, Mills S, Bennett A, et al. HIV and STD epidemiology, risk behaviors, and prevention and care response in Cambodia. AIDS
2. Saphonn V, Sopheab H, Sun LP, et al. Current HIV/AIDS/STI epidemic: intervention programs in car 1993-2003. AIDS Educ Prev
3. Surveillance Unit National Center for HIV/AIDS DaSN. HIV Sentinel Surveillance (HSS) 2003: Results, Trends, and Estimates
. December 04, 2004.
4. World Health Organization, and Joint United Nations Programme on HIV/AIDS. Second Generation Surveillance for HIV: The Next Decade
. Vol. 2000.5. Geneva: WHO/CDS/EDC; 2000.
5. Gorbach PM, Sopheab H, Phalla T, et al. Sexual bridging by Cambodian men: potential importance for general population spread of STD/HIV epidemics. Sex Transm Dis
6. Muecke MA. Mother sold food, daughter sells her body: the cultural continuity of prostitution. Soc Sci Med
7. Bishop R, Robinson LS. Night Market: Sexual Cultures and the Thai Economic Miracle
. New York: Routledge, 1998.
8. Cleo Odzer. Pattpong Sisters
. New York: Arcade Publishing, 1994.
9. Kilmarx PH, Limpakarnjanarat K, Mastro TD, et al. HIV-1 seroconversion in a prospective study of female sex workers in northern Thailand: continued high incidence among brothel-based women. AIDS
10. Thuy NTT, Nhung VT, Thuc NV, et al. HIV infection and risk factors among female sex workers in southern Vietnam. AIDS
11. Wawer MJ, Podhisita C, Kanungsukkasem U, et al. Origins and working conditions of female sex workers in urban Thailand: consequences of social context for HIV transmission. Soc Sci Med
12. Bond KC, Celentano DD, Phonsophakul S, et al. Mobility and Migration: Female Commercial Sex Work and the HIV Epidemic in Thailand. In: Herdt G ed. Sexual Cultures and Migration in the Era of AIDS
. Oxford: Claredon Press, 1997.
13. Mills SBP, Bennett A, Pattalung RN, et al. HIV risk behavioral surveillance in Bangkok, Thailand: sexual behavior trends among eight population groups. AIDS
14. Tarr CM. The Cultural Dimensions of Young, Unmarried Female's Sexual Practices in Cambodian Society
. The Third International Conference on AIDS in Asia and the Pacific, Thailand, 1995.
15. Ryan CA, Vathiny OV, Gorbach PM, et al. Explosive spread of HIV-1 and sexually transmitted diseases in Cambodia. Lancet
16. Leng HB, Sopheab H, Saidel T, et al. Report on HIV Sentinel Surveillance in Cambodia: 2002
. Phnom Penh: Ministry of Health-National Center for HIV/AIDS, Dermatology and STD; 2004.
17. McQueen DV. Surveillance of health behavior. Curr Issues Public Health
18. Mills S, Saidel T, Bennett A, et al. HIV risk behavioral surveillance: a methodology for monitoring behavioral trends. AIDS
19. Gorbach PM, Leng HB, Maloney-Kitts M, et al. Brothel-Based Sex In Cambodia
. International Conference for Sexually Transmitted Disease Research, Seville, Spain, Oct 19-21, 1997.
20. Sobel JD. Gynecologic infections in human immunodeficiency virus-infected women. Clin Infect Dis
21. The Voluntary HIV-1 Counseling and Testing Efficacy Study Group. Efficacy of voluntary HIV-1 counseling and testing in individuals and couples in Kenya, Tanzania, and Trinidad. Lancet
22. Ghys PD, Diallo MO, Ettiegne-Traore V, et al. Effect of interventions to control sexually transmitted disease on the incidence of HIV infection in female sex workers. AIDS
23. Alary M, Mukenge-Tshibaka L, Bernier F, et al. Decline in the prevalence of HIV and sexually transmitted diseases among female sex workers in Cotonou, Benin, 1993-1999. AIDS
24. Hanenberg RS, Rojanaithayakorn W, Kunasol P, et al. Impact of Thailand's HIV-control programme as indicated by the decline of sexually transmitted diseases. Lancet
behavioral surveillance; Cambodia; HIV/AIDS; sexual behavior
Copyright © 2006 Wolters Kluwer Health, Inc. All rights reserved.
What does "Remember me" mean?
By checking this box, you'll stay logged in until you logout. You'll get easier access to your articles, collections,
media, and all your other content, even if you close your browser or shut down your
To protect your most sensitive data and activities (like changing your password),
we'll ask you to re-enter your password when you access these services.
What if I'm on a computer that I share with others?
If you're using a public computer or you share this computer with others, we recommend
that you uncheck the "Remember me" box.
Highlight selected keywords in the article text.
Data is temporarily unavailable. Please try again soon.