PropertyValue
is nif:broaderContext of
nif:broaderContext
is schema:hasPart of
schema:isPartOf
nif:isString
  • The study protocol was approved by the Institutional Review Boards at the Wayne State University in the USA and Beijing Normal University in China. The study was conducted in Guangxi Zhuang Autonomous Region in southwest border of China. One of the five ethnic minority autonomous regions, Guangxi is ranked the second in terms of total HIV prevalence and first in terms of new HIV infection cases [28]. Guilin City located in northeast Guangxi and Beihai City in the southern coast were selected as research sites for the study. Both cities are medium size with population over one million and both are tourists spots attracting 4–10 million visitors to each city every year. Local CDC estimated that more than 2,000 FSW worked in at least 150 commercial sex establishments in each city. We used the same sampling and recruitment strategies as we did in previous studies with FSW in China [26], [27], [29]. In the formative phase of the study, our research team worked closely with the local CDC for ethnographic mapping, in which all known commercial sex venues were updated on the maps. Participants in the current study were recruited from 60 establishments (31 in Guilin and 29 in Beihai) that represented all known commercial sex venues in Beihai and Guilin, including saunas, night clubs, bars, karaoke (KTV), bars, massage parlors, hair salons, restaurants, mini hotels, and the streets. We first contacted the owners or managers of each representing establishment for permission to conduct research on their premises; if they declined (less than 30%), we identified other venues of the same type and similar size until the targeted sample size (i.e., approximately 500 FSWs in each city) was achieved. Once we obtained permission from the manager, we invited FSW in the site to participate; about 20% of FSW in participating venues declined to participate. We conducted the survey in separate rooms or private spaces in the establishments; only the interviewer and the participant were allowed to stay in the room. After the informed consent, the participant completed a structured questionnaire in a paper-pencil format; the interviewer only provided assistance when necessary. It took about 45 minutes to complete the questionnaire and each participant received a small gift equivalent to 30 yuan ($4.5). Among 1,022 women who completed the survey, 937 provided information on violence from clients and were included in the analysis; 279 women did not report having stable partners and therefore did not provide data on PV from stable partners. Demographic information collected in the survey include age, ethnicity (Han or ethnic minority), residency (rural or urban household registration), education (completed middle school or not), marital status (ever married or not), length of migration (months of living in the city), length of working in the city (months of working in the city), living with partners or not, working venue, and monthly income. Because of high correlation between length of migration and length of working, we only included in the latter in the data analysis. Because of the hierarchy of commercial sex establishments, the venues were categorized into four types by the mean income of FSWs at each venue: Level 1 included those venues with mean income higher than 3,000 yuan each month (in this study, only FSWs working sauna had a mean income higher than 3,000 yuan); Level 2 consisted of those venues with mean income between 2,000 to 3,000 yuan (night club, KTV, bar, dancing halls); Level 3 included those venues with mean income between 1,000 to 2,000 yuan (message parlor, hair salon); and Level 4 consisted of those venues with mean income less than 1,000 yuan (restaurant, mini hotel, and streets). Partner violence scales were adapted from the WHO’s Multi-Country Study on Women’s Health and Domestic Violence [1]. PV measured in the current study covered three dimensions: physical violence (e.g., slapped you or threw something at you that could hurt you; pushed you or shoved you or pulled your hair; kicked you; dragged you or beat you up), sexual violence (e.g., forced you into intercourse; inserted something into your genitals), and psychological abuse (e.g., belittled or humiliated you in front of others; threatened to hurt you or someone you care about). PV from stable partners included three additional items, for example, ignoring you for a long time, threat of separating you from your children or terminating your pregnancy, and restriction of your freedom. All the items were assessed using a 4-point scale: 0 = never, 1 = occasionally, 2 = sometimes, and 3 = frequently. The PV from clients included 17 items with a composite score ranging from 0 to 51; the PV from stable partners included 20 items with a composite score ranging 0 to 60. Because of big range and skewed distributions of both variables, we used binary score (dichotomized never or ever experience PV) in data analysis. Psychosocial distress was measured by five indicators: alcohol intoxication, drug use, suicidal behavior, depression, and loneliness; all indicators were validated in Chinese language including our own studies of FSW in China [26], [27], [29]. Alcohol intoxication was measured by frequency of alcohol intoxication (e.g., every day, once every 2–3 days, once a week, once every 2–3 weeks, never). Respondents were categorized into never or ever had abuse intoxication. Drug use was measured by a single item: “Have you ever used illicit drugs? (Yes/No)”; respondents were categorized into never or ever groups. Suicidal behavior was measured by two items of suicidal ideation (“had seriously considered killing yourself”) and suicidal attempt (“had tried to kill yourself”). Respondents who answered “yes” to either of these two questions were categorized into ever had suicidal behavior and the rest into never. Depression was measured by the Center for Epidemiologic Studies Depression scale (CES-D) [30], which was validated in Chinese culture.31 The Cronbach’s alpha for the current study sample was.89; and the mean score was 17.35 (ranged 0 to 53); a cut-off point of 16 was used for dichotomization [31]. Loneliness was measured by the UCLA Loneliness scale [32]. The Cronbach’s alpha of the scale in the current study was.74; the mean score was 43.30 (ranging from 20 to 75); a cut-off point of 44 was used for dichotomization [32]. First, we used Chi-square and t-test to compare demographic characteristics between those who had never and ever experienced PV from stable partners and clients. Second, multivariate logistic regression models were built to examine the independent relationship between PV and key demographic variables. To control for potential intra-class correlation (ICC) by venue due to cluster-sampling, we used random effect modeling. Adjusted odds ratio (aOR) and 95% confidence intervals (95% CI) were used to depict an independent relationship between dependent and independent variables. Third, the association between PV and each of the five indicators of psychosocial distress was analyzed using Chi-square or t-test. Finally, multivariable regression models were built to examine further the relationship between PV and psychosocial distress (each of the five indicators and the total psychosocial distress score) while controlling for potential confounders. Variables of significance in Step 2 were included in the multivariable models. Similar to Step 2, random-effect modeling was employed to control for ICC; adjusted odds ratio (aOR) for logistic regression models and 95% CI were used to depict the independent relationship between dependent and independent variables. Data management and analysis were performed using Stata 10.0.
rdf:type