In the past two decades, the commercial sex trade has dramatically increased in China, including in southern Yunnan where the HIV epidemic started. It is now well established among young, male, injecting drug users (IDU) [1–7]. Commercial sex plays a significant role in China in spreading HIV from drug injectors to the general population [8–11].
Studies conducted in neighbouring Thailand have shown that the promotion of consistent condom use among sex workers in natural settings can effectively reduce sexually transmitted diseases (STD) and HIV infection [12,13]. However in 1990s' China, this strategy was perceived as promoting prostitution and deteriorating China's ‘spiritual civilization’. At that time, education was only delivered in prisons and women's re-education centres, with assistance from public security officers, but was largely ineffective. According to data collected by the Ruili Centre for Disease Control (CDC), 80% of sex workers continued prostitution after being released, and the majority did not learn to use condoms consistently (Ruili CDC, unpublished survey). Moreover, arrestees only account for < 10% of all sex workers (Public Security Bureau, unpublished data). Given these limitations, this was the first study to explore a community-based education program for sex workers without public security involvement.
Ruili, Longchuan and Chengjiang counties in Yunnan Province were selected for the study. Ruili and Longchuan are located in southwest Yunnan along the Myanmar border, and were among the counties first hit by HIV in China . Chengjiang is a county located 40 km from Kunming, the capital of Yunnan Province. Chengjiang was selected because it is a tourist destination and HIV cases had been reported in the years prior to the intervention.
A previous study in Dehong  as well as interviews with key informants prior to the study identified that the majority of sex workers work in bars, barber shops, beauty parlours, massage centres, karaoke clubs and dance halls. Most establishments are owned by local residents, but the girls working in them typically come from rural areas of neighbouring counties and provinces. Thus, although ethnic minorities such as the Dai and Jingpo, comprise more than 50% of the population in Ruili and in Longchuan, more than 95% of sex workers are Han (the ethnic majority in China), because they are not local residents. The services offered in these establishments include massage, rubbing, and sex, however sex workers often encourage their clients to buy sex since they can charge more money. Consistent condom use was low .
The objectives of the study were to test the feasibility of conducting intervention activities among sex workers in community settings, and to evaluate the efficacy of the intervention to increase knowledge of HIV transmission routes and prevention measures, perceived risk of HIV infection, and to promote condom use.
The study was conducted between March and October 1997. Approval for the study was granted by the institutional review board of the Chinese Academy of Preventive Medicine (now the Chinese Centre for Disease Control and Prevention).
Thirty establishments in Chengjiang and 34 in Ruili participated. These establishments were chosen because they were concentrated in one or two streets in each site and had roughly 120 to 140 sex workers working in them. In Longchuan, the number of establishments was relatively small, so all 23 establishments in the study township were included.
A qualitative study involving face-to-face interviews with sex workers, and focus group discussions with sex workers and establishment owners (separately), was conducted to understand the commercial sex business and the life styles of sex workers, and to develop a specific work plan for the targeted intervention. Sex workers and establishment owners were approached at their workplace by study staff who explained the objectives and procedures of the study and then invited them to participate. Participants gave verbal consent and were given ¥50 RMB (≈US$6.30) for participation.
Interviews were conducted with three sex workers in Longchuan, two in Ruili and three in Chengjiang. Only one of the nine sex workers approached refused to be interviewed. Interviews followed an interview guide, which included questions about: working history, number of establishments worked at and how often they moved on, services offered and amounts charged; condom use with clients and other partners, preferred brands and where they usually purchased them; STD history and testing/treatment preferences; abortion history; and advice on how, where and when to conduct the intervention. Interviews lasted for 1–2 h and were conducted in private rooms.
Two focus groups were conducted in each site; one with establishment owners, the other with sex workers. In Longchuan seven establishment owners and five sex workers participated in separate focus groups. In Ruili, these numbers were 13 and 14, and in Chengjiang they were 12 and 15, respectively. Two of the sex workers and none of the establishment owners refused to participate. Focus group discussions lasted for 1.5 to 2 h. There were two local facilitators, one male health worker and one female health worker, to chair the discussions. Focus groups began with two videos about HIV from Thailand and then followed a similar interview guideline as that used in the interviews.
A survey instrument was developed based on information gathered from the literature and the qualitative study. The questionnaire was pretested on 15 sex workers and modified accordingly. Information collected included: demographic characteristics; basic knowledge of STD/HIV transmission routes and prevention measures; perceived risk of infection with STD/HIV; history of STD and testing; and condom use.
Interviews were conducted in private rooms by trained, married, female health workers using an interview-administrated questionnaire. The purpose and nature of the study was explained, participants were invited to ask any questions and were then asked for their signed informed consent form. Participants were provided with free packets of condoms, lipsticks and perfumes for participation in the study. Once the questionnaire was completed, the interviewer went through the knowledge questions one by one together with the sex workers to reinforce what they already knew, and to give them correct answers for questions that were answered incorrectly. They also encouraged those who used condoms consistently to keep doing so, and those who did not or only sporadically used condoms to use them consistently.
Roughly 1 month after the baseline survey, a 3-month intensive, comprehensive intervention was implemented in the selected establishments targeting both establishment owners and sex workers. Twelve trained, married, female health workers (four in each study site) visited the establishments on a fortnightly basis. Six face-to-face intervention sessions (one every 2 weeks) were designed and implemented, with each session focusing on one topic with the following titles: (1) ‘Everyone needs to know about AIDS’, a cartoon educational folder; (2) ‘Everyone needs to know about condoms’, a cartoon educational folder; (3) ‘Ant nibbles house and HIV kills your body’, a cartoon educational folder that explains why it is not possible to know if a person is infected with HIV by looking at him/her; (4) ‘What are Thai sex workers doing to protect themselves?’ a video cassette dubbed into Chinese that describes prevention of STD/HIV among Thai sex workers; (5) ‘Xie-Er, Ping-Mei and Me’, a cartoon educational folder that uses three personal stories to emphasize the importance of consistent condom use; and (6) ‘A handsome fellow gives me HIV’, a cartoon educational folder that uses a story to illustrate that we cannot judge whether or not a person is free from HIV infection because he is good looking or physically strong. In addition, health workers asked the group about condom use with their last clients. They encouraged sex workers who were using condoms to continue using them and asked sex workers who had not used condoms the reasons why and encouraged them to begin using them. Educational flyers were distributed and condoms were sold to sex workers at one-third the market price.
Establishment owners were also trained on how to encourage sex workers to use condoms with clients and how to find STD services for testing and treatment. Each establishment owner was given one copy of ‘House of Charm’, a video cassette dubbed into Chinese that describes prevention of STD/HIV among Thai sex workers, and a copy of ‘Everyone needs to know about AIDS and everyone needs to know about condoms’, a recorded tape cassette that could be played for the sex workers. In Chengjiang County, an Establishment Owners Association was set up to facilitate intervention activities. Health workers met with the association on a regular basis to discuss the progress of the intervention and to deliver lectures.
During the two-week interval between scheduled discussions, establishment owners, selected sex worker peer-leaders and the sex workers themselves carried out intervention activities. These included three prevention strategies:
1. playing the distributed video and tape;
2. giving printed educational materials to sex workers newly arrived at the establishment or who had not received educational materials;
3. discussing the previous night's business and whether or not they used or tried to use a condom, how they successfully persuaded their clients to use a condom, and possible explanations for not being able to persuade their clients, if applicable.
The key message of consistent condom use with every client was emphasised using multiple channels, including face-to-face lectures, questions and answers, educational folders, audio and vidio tapes, during the 3-month intensive intervention.
Process evaluation was used to monitor implementation of intervention activities. In each twice-weekly contact with establishments, information was collected about the number of sex workers participating in the discussion and participation to date, the number of new sex workers at the the establishments, the number of sex workers who had left the establishments since the last meeting, the number of educational materials distributed, and the number of condoms distributed.
Outcome evaluation was used to determine the effectiveness of the programme. The changes in knowledge of HIV transmission routes and prevention measures and condom use with the most recent and last three clients were compared pre- and post-intervention. Categorical variables were compared using chi-squared tests, and continuous variables by t-test. To determine the impact of the intervention on condom use, two logistic regression models were developed for condom use with the last client and condom use with the last three clients. Residency, age group, ethnic group, education level, marital status and intervention (participated in health worker-chaired activities at least once) were treated as independent variables. Univariate analyses were first used to identify variables (α = 0.2) for the multivariate model. Statistical analyses were performed using Stata™ V6 (StataCorp, LP, College Station, Texas, USA).
The number of sex workers per establishment varied widely, from as few as three to more than 30, and was lowest in the rainy season. The relationship between sex workers and establishment owners was unstable. Sex workers typically paid owners ¥50 RMB (≈US$6.30) for using the establishment to find clients, and charged clients ¥200 to ¥300 RMB (≈US$25–37) for sexual services. Sex workers frequently moved from one establishment to another, usually every 2 to 3 months. If the owner treated them well, some would stay for more than 12 months. Consistent condom use varied: less experienced sex workers, with a low level of education, were less likely to use a condom and would accept extra payment to not use a condom. Most sex workers reported a history of abortion and symptoms of sexual or reproductive health problems in the past 12 months.
In terms of delivering the intervention, both establishment owners and sex workers prefered female health workers aged around 30 to 45 years rather than receiving the intervention from their peers, establishment owners, or police officers. Reasons given included that they preferred women and respected the knowledge of health workers. Both establishment owners and sex workers reported that between 3:00pm and 7:00pm would be the most convenient time for delivering intervention activities. They also reported that they preferred a format that involved mini-training workshops, with question and answer time, plus cartoon, audio and video materials.
One hundred and nineteen sex workers in Chengjiang, 128 in Ruili and 66 in Longchuan were interviewed at the baseline survey. One hundred and thirty four sex workers in Chengjiang, 112 sex workers in Ruili and 51 sex workers in Longchuan were interviewed at the evaluation survey 2 weeks after delivery of the last session of intervention. Seven sex workers refused to participate in the baseline survey but none refused in final evaluation survey.
All sex workers participating in the evaluation survey reported had participated in intervention activities at least once. Fifty-one percent attended one session, 28% attended two sessions, 13% attended three sessions, and 8% attended at least four sessions. Participation in Chengjiang was better than in the other two sites, with 74% attending at least two sessions compared with 26% in Ruili and 35% in Longchuan (P < 0.001).
The demographic characterstics of the participants are outlined in Table 1. In general, demographic characteristics were similar at baseline and final survey, except in Ruili for residency, marital status and education. At baseline, in Chengjiang and Longchuan, the majority of sex workers were Yunnan residents, while in Ruili, they were mostly migrants from other provinces. Most sex workers were aged 20 years or older in Chengjiang and Ruili but younger than 20 years old in Longchuan. In all three study sites the majority were Han at both baseline and final survey, but there was a slightly higher proportion of other ethnic groups in Longchuan compared with Ruili and Chengjiang. This may reflect that Longchuan had a higher proportion of local residents. No sex workers had received more than 12 years of school education. Sex workers in Longchuan had a lower education level than in Chengjiang and in Ruili. The proportion of married sex workers were relatively small. However, there was a relatively higher proportion of married sex workers in Ruili than in Chengjiang or Longchuan.
Participants showed significant improvements in HIV knowledge, perceived risk and condom use (all P < 0.001). Knowledge of HIV transmission routes increased from 25% at baseline to 88% at the end of project. Knowledge that eating with a person living with HIV/AIDS (PLHA) and that mosquitoes cannot transmit HIV increased from 45 to 87% and from 25 to 70%, respectively. Knowledge that using a condom reduces the risk of STD/HIV increased from 56 to 94%. Perceived risk of HIV infection increased from 20 to 52%. Self-reported condom use with the last client and with the last three clients increased from 61 to 85% and 41 to 70%, respectively. Changes in knowledge and risk behaviours for the individual sites are listed in Table 2.
The logistic model showed that age, education level and marital status were related strongly to condom use (Table 3). After controlling for age, educational level, and marital status, women participating in the intervention were 3.7 times more likely to have used a condom with the last client (95% confidence interval = 2.1–6.4) and were 4.0 times more likely to have used condoms with the last three clients (95% confidence interval = 2.6–6.3).
Among the 297 sex workers who participated in the evaluation survey, we further compared those who participated in at least two intervention sessions with those who participated in only one session. No differences were observed between the two groups in terms of knowledge that condoms reduce risk of transmission (92 vs. 96%, P > 0.05), but condom use with the last client (91 vs 79%, P < 0.05) and with the last three clients (78 vs 72%, P < 0.05) was significantly higher in the group who attended more than one session.
This was the first intervention study in China designed to increase knowledge of STD/HIV transmission routes and prevention measures and to promote safer sexual behaviours among sex workers in a community-based setting. The study demonstrated that it was both feasible and effective to conduct such interventions.
The changes in knowledge and in sexual behaviours varied in the three study sites, but were greatest in Chengjiang. There are several possible explanations for this. First, sex workers in Chengjiang were relatively stable in comparison with the other sites and had therefore participated in more of the intervention activities. Ruili and Longchuan have a rainy season from late June to September, whereas Chengjiang does not. Since the heavy rains reduce business, there is more sex worker migration in Ruili and Longchuan than in Chengjiang. Second, although it was not a component of the intervention, in both Chengjiang and Ruili, local health workers referred all participants to STD check and treatment services. Third, an Establishment Owners Association was set up in Chengjiang while no such association was set up in Ruili or in Longchuan. This association played an additional role in the intervention, and gave the establishment owners a greater sense of partnership and responsibility for the intervention. Given that the sex worker population is highly mobile, but that establishment owners are relatively stable, the participation of owners was crucial for continued delivery of intervention activities, even after the trial had ended.
The study had several limitations. First, condom use behaviours were self-reported. Given that the intervention emphasized condom use, responses may have been influenced by social desirability bias which may lead to overestimation of the intervention's impact. Second, we used two condom use behaviours, condom use with the last client and condom use with the last three clients, as outcome measurements. In both baseline and final surveys, the rate of condom use with the last three clients was lower than the rate of condom use with last client, implying that consistent condom use might be even lower. Third, the study was designed as a three-month intensive intervention and had neither involved the local police departments nor other government sectors; therefore, expanding this model in large scale might have unexpected challenges. Fourth, there are different types of sex workers in China [10,15]. Some work at big hotels, some at smaller establishments such as hair dressing salons, some along the streets, some from home, and some from a car. We only studied sex workers who were stationed at establishments and so our results may not be generalizable to sex workers who operate independently of an establishment. Different intervention formats are likely to be needed for these other populations.
This study evaluated the feasibility and impact of an intervention targeting sex workers. The results indicated that the program was both feasible and had the desired effect of increasing HIV/STD knowledge and condom use with clients. Training from local health workers and engagement of establishment owners in intervention activities is recommended for future programs.
Sponsorship: This study was supported by the World Bank. Preparation of the manuscript was supported by the China Multidisciplinary AIDS Prevention Training Program with NIH Research Grant # U2R TW06918 funded by the Fogarty International Center, National Institute on Drug Abuse, and the National Institute of Mental Health.
1. Zheng XW. A preliminary study on the behavior of 225 drug abusers and the risk factors of HIV infection in Ruili county Yunnan Province [in Chinese]. Chin J Epidemiol 1991; 12:12–14.
2. Ma Y, Li ZZ, Zhang KX, et al
. Identification of HIV infection among drug users in China [in Chinese]. Chin J Epidemiol 1990; 11:184–185.
3. Zhang JP, Cheng HH, Zhao SD, et al
. An epidemiological study on HIV infection in Ruili county Yunnan Province [in Chinese]. Chin J Epidemiol 1991; 12:9–11.
4. Zhao SD, Cheng HH, Zhang JP, et al
. AIDS surveillance in Yunnan Province in China (1986-1990) [in Chinese]. Chin J Epidemiol 1991; 12:72–74.
5. Sun X, Nan J, Guo Q. AIDS and HIV infection in China. AIDS 1994; 8(Suppl 2):S55–59.
6. Cheng C, Zheng X, Liu K, Kang L, Hao R. Epidemiology of HIV/AIDS in China [in Chinese] In: Office SCR (editor). Beware of AIDS: for the survival of the Chinese nation
. Beijing: Xinhua Publishing House; 1993, pp. 35–58.
7. Zheng X, Tian C, Choi KH, Zhang J, Cheng H, Yang X, et al
. Injecting drug use and HIV infection in southwest China. AIDS 1994; 8:1141–1147.
8. Wu Z, Zhang J, Detels R, Li VC, Cheng H, Duan S, et al
. Characteristics of risk-taking behaviors, HIV and AIDS knowledge, and risk perception among young males in southwest China. AIDS Educ Prev 1997; 9:147–160.
9. Kaufman J, Jing J. China and AIDS-the time to act is now. Science 2002; 296:2339–2340.
10. Yang H, Li X, Stanton B, Liu H, Liu H, Wang N, et al
. Heterosexual transmission of HIV in China: a systematic review of behavioral studies in the past two decades. Sex Transm Dis 2005; 32:270–280.
11. Xiao Y, Kristensen S, Sun J, Lu L, Vermund SH. Expansion of HIV/AIDS in China: lessons from Yunnan Province. Soc Sci Med 2007; 64:665–675.
12. Wongkhomthong S, Kaime-Atterhog W, Ono K. AIDS in the developing world: a case study of Thailand. Bangkok: Hoslistic Publishing; 1995.
13. Rojanapithayakorn W, Hanenberg R. The 100% condom program in Thailand. AIDS 1996; 10:1–7.
14. Wu Z, Zhang J, Dong L, Li Z, Wu B, Duan S, et al
. Commercial sex establishments and sex workers in Dehong Prefecture, Yunnan Province [in Chinese]. Chin J Prev Control STD and AIDS 1997; 3:14–16.
15. Tucker JD, Henderson GE, Wang TF, Huang YY, Parish W, Pan SM, et al
. Surplus men, sex work, and the spread of HIV in China. AIDS 2005; 19:539–547.