Home / Topics / Collaboration / Research Teams / Building and Maintaining Teams / Establishing an International Community-Based Research Study
Margaret R. Weeks

Establishing an International Community-Based Research Study

Posted on May 18, 2011

Margaret R. Weeks (bio) discusses how she and a team of U.S. and Chinese researchers initiated a participatory community intervention study to promote female condoms among rural sex workers in southern China.


Q: What is the focus of this project?
A: This project, titled the "China/U.S. Women's Health Project" and funded by the U.S. National Institute of Mental Health, is a study of HIV risk and prevention among women working in the sex industry in four towns in southern China. Its goal is to design, implement, and test an establishment-based and community-wide intervention to promote the use of the female condom to prevent HIV and other STDs.

We're currently in the last year of this five-year project. We recently completed intervention and process evaluation efforts in four study sites: two rural towns, a small city, and a larger city.
Q: What is the composition of the research team?
A: The research team includes investigators from the Department of Epidemiology at Peking Union Medical College (PUMC) in Beijing and the Institute for Community Research (ICR), a nonprofit community research organization in Hartford, Connecticut.

It also includes health educators and public health professionals from the Centers for Disease Prevention and Control (CDCs) in the provinces and counties where the research is being done. Local nurses, gynecologists, public health educators, and outreach workers are involved.

This U.S./China partnership has been in place for almost 10 years, and this collaboration has been central to the success of our HIV prevention efforts.
Q: What kind of staffing structure was established?
A: The staffing structure is based on meetings between the co-PI from the Beijing location, the HIV division leaders of the provincial and county CDCs, and me, the PI from the Hartford location. Research oversight is split between the co-PI and myself, and each of us has a community ethnographer and statistical data analysts, in addition to a few other staff.

The provincial and county CDCs handle data collection and management and manage intervention related activities. Township-based staff were assigned to implement the intervention, since they have the best access to local sex work establishments.
Q: What were some of the staffing challenges the team encountered?
A: Finding qualified staff for roles such as outreach, health education, and data management, was difficult for several reasons. First, the time-limited nature of the employment opportunities on the project was a barrier. In addition, we needed local staff experienced in reproductive health care to organize key project activities in the community, but most experienced health workers were already committed to other work in the local hospitals.

Furthermore, full-time, field-based local project staff who had less experience with the target population were not equipped to handle many of the sex workers' questions. The newer project staff also needed more time to establish a rapport with the women in the study and the establishment owners than the more experienced staff, who had already developed these relationships through prior interventions in the establishments.
Q: How were the staff trained?
A: We began each cycle of activities with an intensive 3-5-day training program to educate local staff on the study design, research and intervention components, and protections protocols for human subjects.

New staff were trained in women's anatomy, reproductive health, and the use of the female condom, and they accompanied experienced staff to the sites until they were comfortable with the implementation of the program. Ongoing project meetings also help to support the implementation staff.
Q: How were intervention sites chosen?
A: Ethnographers from both China and the U.S. conducted a formative ethnography to learn about the context and dynamics of the sex industry in the study towns, the women's HIV and STD risks and protections within the sex-work establishments, and other characteristics of the towns in the study. Through in-depth interviews and field observations in the establishments and in local health facilities, they evaluated the potential for the female condom intervention.

Local outreach staff were instrumental in helping the ethnographers conduct their research. They leveraged their familiarity with many of the sex-work establishments, their standing relationships with some of the establishment owners, and their knowledge of the local residents and the regional dialects to facilitate access to the establishments and, at times, ease the apprehension of owners and sex workers toward the study.
Q: What was your recruitment strategy?
A: Our recruitment goal was a minimum of 80% of all women in the establishments selected for the study. In some establishments, which also offer non-sex services such as massage and beauty services, it's challenging to differentiate sex workers from other workers. In addition, engagement in sex work is often under-reported because of the illegal nature of the work. Given these factors, all sexually active women working in these establishments were invited to participate in the study. One advantage of this approach is that it decreased the emphasis on sex workers, thus minimizing the potential stigma of participation in the project.
Q: What were some of your initial recruitment challenges and how did the staff resolve them?
A: We faced a number of recruitment challenges, including the following:
  • The study's focus on sex-work establishments, despite significant efforts to downplay this emphasis, made some of the sex workers and establishment owners reluctant to participate.
  • Some establishment owners hesitated to allow sex workers to go off-site for the hour-long surveys conducted as part of the study, for fear of losing business during the women's absence.
  • Some women didn't want to go to the interview site for fear of being recognized as sex workers.
  • Many women simply preferred to spend their free time doing things other than participating in the study activities, such as playing majiang, a Chinese board game used for gambling.
Luckily for us, the local outreach staff was highly successful in helping us identify and address these barriers. For example:
  • They addressed fears of stigma among prospective project participants by broadly promoting the use of the female condom, emphasizing that it can help protect all sexually active women from STDs or unwanted pregnancies.
  • They protected the privacy of participants by arranging for interviews to take place in rented rooms in a local hotel. And in some cases, they arranged for surveys to be conducted in safe, private rooms on site at the sex-work establishment, minimizing the possibility of drawing attention to the sex workers as well as decreasing time away from the worksite.
  • They came up with appealing incentives, such as beauty products, to encourage participation in the survey.
Q: Were you successful in achieving your initial recruitment goals?
A: We were very pleased with our recruitment rate. Within only one month, we got representation from every known sex work establishment in our first study town and over 85% of the establishments in the second town. The third study site was even better; we were able to recruit the sample we needed in only a few weeks. We attribute much of this success to the knowledge and dedication of the county-level staff and local outreach workers.

However, survey recruitment in the fourth study site – a larger city – was much more challenging and took several months to complete. Less outreach experience among staff and more challenges with establishing rapport slowed the process.
Q: What challenges did follow-up recruitment present and how did the staff address them?
A: We had two primary challenges when we conducted our six-month follow-up surveys.

First, we had to deal with the high rate of turnover among the study participants. To address this issue, we incorporated methods from both longitudinal and cross-sectional design. Unique, easily reconstructed IDs were given to participants on a card, which they were asked to bring on their follow-up visits to help us link baseline and follow-up data.

We primarily used a cross-sectional design, recruiting at least 80% of women in the sex industry in each town at each time point, so our outcome evaluation will be based on the study population, not just a sample. In the larger city, we sampled districts of the city and still attempted to recruit all of the women working in those districts.

Additionally, our intervention is intentionally focused on the establishments; therefore, sex workers who began working in the town after the baseline recruitment still may experience the intervention through peers or ongoing outreach. The cross-sectional outcome evaluation takes into account both the intervention (at the establishment level) and individual exposure (time spent in the establishment).

Secondly, based on negative feedback about the time required for the baseline survey, field staff were concerned that the time and effort involved in the study over time would be perceived as too burdensome for sex workers as well as establishment owners. Research and outreach staff addressed this concern by offering additional incentives that included a special incentive for women who participated in both baseline and follow-up surveys. As with the baseline surveys, staff conducted some of the follow-up surveys at the sex-work establishments – rather than an off-site location – to minimize the burden on participants.
Q: Were these follow-up recruitment strategies successful?
A: Yes, they were very successful. We were able to recruit almost 98% of eligible sex workers at the six-month survey interviews in the first three study sites. We have not yet started the follow-up surveys in the fourth site.

Although we had more participants willing to take the follow-up study than we expected, the sex worker turnover was, in fact, significant. Almost half of the women had started working in the establishments after the baseline survey was conducted. In addition, only 29 of 154 women in the first two sites presented their unique ID at the follow-up survey.

Unfortunately, this combination of factors resulted in a relatively low rate (about 20%) of matched surveys, so our outcome evaluation will be based on analysis of cross-sectional data at different time points.

The good news is that many women who were not present for the baseline were exposed to the intervention and other project activities. Additionally, we were able to gather quite a bit of systematic data through both qualitative and quantitative methods, thanks to the strong presence of ethnographers, interviewers, and local outreach staff in the study locations.

To learn more about how the intervention was implemented and how the team addressed related challenges, read Implementation Challenges in an International Community-Based Research Study.



Weeks, M., Liao, S., Li, F., Li, J., Dunn, J., He, B., ... Wang, Y. (2010). Challenges, strategies, and lessons learned from a participatory community intervention study to promote female condoms among rural sex workers in southern China. AIDS Education and Prevention, 22(3), 252-271.


Please note that the feedback is viewed only by 4researchers staff and is not intended for communication with individual contributors.


Use the form below to submit feedback about this article. If you would like a response, please be sure to include your e-mail address.

More About "Building and Maintaining Teams"

15 People; 15 Different Projects


A Collaboration of Systems Thinkers

Show All...


More From Margaret R. Weeks (bio)


Applying Anthropological Principles


Continuous Feedback

Show All...