Wang, Bo PhD*; Li, Xiaoming PhD†; McGuire, James PhD*; Kamali, Vafa PhD*; Fang, Xiaoyi PhD‡; Stanton, Bonita MD†
A substantial research literature hastedly identified STD/HIV knowledge, susceptibility to STD and HIV infection, and condom use self-efficacy as important predictors of consistent condom use among female sex workers (FSWs).1–4 Condom-use skills have been found to be positively related to condom use.4 However, a behavioral survey among brothel-based FSWs in Indonesia revealed that knowledge about STDs and AIDS, perception of susceptibility and severity of the disease, and self-efficacy in using condoms did not significantly predict consistent condom use, whereas perceived clients’ and managers’ attitudes towards condom use was a significant determinant.5 Another study among FSW in Indonesia also found that self-efficacy was not significantly associated with consistent condom use.6 Inconsistent results among existing studies suggest that work remains to be done to determine the complex relationship between multiple determinant factors of condom use.
The exploratory conceptual framework for this study has been developed based on our understanding of the relationships between determinant factors of condom use within commercial sex, which combines theoretical concepts from the health belief model7 and social cognitive theory.8 In the context of this study’s model, beliefs including perceived susceptibility and severity of STD and HIV determine an individual’s action (e.g., condom use) according to the health belief model. Self-efficacy, an essential component of social cognitive theory, mediates the effects of STD/HIV knowledge and condom use skills, and determines whether people adopt safer sex behavior.
Prior research has demonstrated an important link between perceived benefits of unprotected sex and inconsistent condom use among adolescents.9,10 Limited research has been conducted to examine this connection within the realm of commercial sex. A recent cross-sectional survey conducted in the Democratic Republic of Congo revealed that 26.5% of FSWs engaged in unprotected intercourse for extra money.11 Financial strain was identified as the major obstacle to condom use in a study among FSWs in South Africa.12 Perceived costs and benefits of using condoms were found to be consistently associated with consistent condom use in the past week in a study among a group of FSWs in Indonesia.6 These studies indicate that perceived benefit and costs of unprotected sex are important factors that may influence condom use among FSWs.
Environmental-structural support for condom use has been identified as an important predictor of consistent condom use among FSWs.3,4,13 These environmental-structural factors include access to condoms in the establishment, communication between establishment owners/managers and FSWs about the importance of condom use, an affirmative policy of condom use, and on-going reminders regarding condom use. Morisky et al.13 reported that FSWs in establishments where a supportive condom use policy exists were 2.6 times more likely to consistently use condoms compared with establishments that do not have such a policy in place. Perceived gatekeeper support for condom use was found to be positively associated with condom use communication with sexual partners and condom use frequency among Chinese FSWs.14 These findings highlight the importance of integrating environment-structural factors into conceptual frameworks of HIV prevention.
China’s burgeoning AIDS epidemic is fueled by widespread commercial sex work. It is estimated that there are approximately 4 to 10 million active FSWs in China.15,16 Sentinel surveillance data reveals that the proportion of heterosexual transmission among cumulative HIV infections increased from 10.7% in 2005 to 37.9% in 2007.17 HIV prevalence among FSWs in some surveillance sites was high, with 11% HIV seroprevalence among FSWs in Guangxi province and 4.6% in Yunnan province.18 Prevalence rates of STDs are considerably higher; nearly half of FSWs tested in 2 recent studies were currently infected with a STD.19,20 Behavioral-risk assessment data reveal that nearly half of FSWs never used a condom, and only about one tenth of FSWs reported always using condoms with clients in China.21
Consistent condom use is an effective method to control STD/HIV transmission among FSWs. Comprehensive interventions with a focus of condom promotion have been launched in many entertainment establishments in China with the support of the Chinese government in recent years.17 Several intervention studies have been conducted among FSWs and found to be effective in reducing sexual risk behaviors.22–25 Evaluations of these studies have shown that theory-based interventions significantly increased consistent condom use among FSWs with their clients, and reduced the infection rate of STDs among FSWs.24,25 To inform future efforts in this regard, it would be helpful to have a greater understanding of the context and dynamics of condom use in commercial sex.
Drawing on data gathered through the baseline survey of a community-based HIV intervention project among a group of establishment-based FSWs in China, this study attempted to enhance the understanding of dynamics of condom use in the context of commercial sex work. We simultaneously examined the relationship of several important determinants of condom use using a hypothesized theoretical model (Fig. 1). Specifically, we address the following 2 questions (1): in what ways are HIV/STD knowledge, condom use skill, and susceptibility/severity related to self-efficacy, perceived benefits, and condom use? (2) What are the relationships between self-efficacy, perceived benefit, perceived establishment-level support, and condom use?
The study was carried out in 2004 in a suburban area (H County) of Nanning, the capital city of Guangxi Zhuang Autonomous Region (“Guangxi”). The original study was a community-based voluntary counseling and testing intervention study.24 This analysis is based on the baseline data of the original voluntary counseling and testing intervention study.
Guangxi has witnessed an alarming rise in HIV prevalence in the past decade. A total of 8602 HIV-infected cases were officially reported at the end of 2003, which placed Guangxi third among Chinese provinces in terms of reported HIV seropositive cases.26 A prosperous economy, increased international contact, and growing tourism in Guangxi have created a market for commercial sex. According to the statistics from the Guangxi public security agency, there were at least 50,000 FSWs in Guangxi. There were an estimated 200 entertainment establishments, with more than 2000 women engaging in commercial sex service in the H County.
Participants and Survey Procedure
Participants were recruited from restaurants, barbershops, and hair-washing rooms in 3 geographic locations in H County: the county seat, a recently established development zone in rural-urban conjunction, and 1 rural township. These locations were selected based on recommendations from local health agencies (e.g., the county health department, the county antiepidemic station). The detailed recruitment procedure has been described elsewhere.24 Briefly, the research team and local health workers conducted an ethnographic mapping of entertainment establishments providing commercial sexual services and identified 85 establishments in the 3 targeted areas (53 in the county seat, 12 in the development zone, and 20 in the township). The owners/managers of these establishments were contacted for their permission to conduct research in their premises. We received permission from owners/managers of 57 (67%) establishments. Trained outreach health workers from the county antiepidemic station and local hospitals requested women in these establishments to participate in this study. Among the 582 women contacted, 454 (78%) women agreed to participate in this study and completed a self-administered questionnaire. Participants were paid the equivalence of US $3 for their participation.
A structured questionnaire was administered to all women who provided written informed consent. The survey was confidential and conducted in separate rooms or private spaces where participants were recruited (e.g., workplace, living quarters). The questionnaire was pretested in 2 waves among 22 women (7 in wave 1 and 15 in wave 2) to determine that the content and language were appropriate for the study population. No person was allowed to stay with the participant during the survey except the interviewer who provided the participant with necessary assistance. The questionnaire took about 1 hour to complete. The study protocol was approved by the institutional review boards at Wayne State University in the United States, Beijing Normal University, and Guangxi CDC in China.
In the questionnaire, there were 22 items addressing knowledge of STD symptoms (10 items), HIV transmission modes (6 items), and misconception of HIV transmission through routine daily contact (6 items). STD and HIV knowledge was measured by adding the number of correct answers to 10 questions on STD symptoms and 6 questions on HIV transmission. Conversely, HIV misconception was measured by adding the number of incorrect answers to 6 questions regarding the possibility of HIV transmission. The internal consistency estimates (Cronbach α) for the 3 knowledge subscales were 0.82, 0.90, and 0.79, respectively. In path analysis, the sum of the number of correct answers to the 22 knowledge questions was retained as a single score of STD/HIV knowledge, with higher scores reflecting increased knowledge about the transmission and symptoms of HIV/STD.
Condom Use Skill.
Participants were asked to sort 6 preprinted pictorial cards describing major steps of condom use. Using procedures developed by Wright et al.,27 a picture sequencing score was calculated for each participants based on 2 criteria: closeness of each picture was to its correct absolute position and number of pictures that were sequenced correctly, regardless of their absolute position. To calculate these scores, each picture was initially numbered from 1 to 6 indicating its correct order. Then 1 point was given for every picture with a lower number that was placed to its left (maximum possible score of 15), and 1 point was given for each correctly adjacent pairs of pictures (maximum possible score of 5). These 2 scores were summed, yielding a maximum score of 20, with higher scores indicating increased condom use skills.
Susceptibility and Severity.
This composite measure assessed perceived personal susceptibility to and perception of negative consequences resulting from risky behavior. Participants were asked to rate their perceptions regarding the likelihood of acquiring HIV and STD infection (e.g., “how likely do you think it is that you would get an STD/HIV in the future?”). This 2-item scale had a Cronbach α of 0.66. Participants were also asked to assess their perceptions regarding negative consequences resulting from being infected with HIV (e.g., “one will lose his/her friends if he/she becomes infected with HIV”). The Cronbach α for this 3-item scale was 0.71. In path analysis, a composite score was created by summing the numbers of positive responses (e.g., likely or agree) across the 5 items. The composite score ranged from 0 to 5 with a higher score indicating a higher level of perceived susceptibility/severity.
The self-efficacy scale assessed personal belief about own ability to use a condom. There were 5 items measuring self-efficacy (e.g., “I can persuade my partner to use a condom if he is unwilling to use it;” “I will refuse to have sex if my partner does not want to use a condom”). The Cronbach α for the 5 items was 0.55. A composite score was created by summing the numbers of positive responses across the 5 items. The composite score ranged from 0 to 5 with a higher score indicating a higher level of self-efficacy.
Five items were employed to measure perceived benefits of unprotected sex (e.g., “if I do not use condoms, my clients will pay me more;” “if I do not use condoms, my clients will come back in the future”). The Cronbach α was 0.75. In path analysis, a composite score was created by summing the numbers of positive responses across the 5 items. The composite score ranged from 0 to 5 with a higher score reflecting a higher level of perceived benefits of unprotected sex.
Perceived Establishment-Level Support.
Four items were included in the final composite measure of establishment-level support for condom use. These items included women’s perceptions of their boss’/mommy’s attitude toward condom use, access to condoms in the establishment, and discussion of condom use with boss/mommy. The Cronbach α for the 4 items was 0.66. A composite score of perceived establishment-level support was created by indexing positive condom use attitudes and establishment practice (i.e., requiring/reminding FSWs to use condoms, allowing FSWs’ refusal of sexual intercourse if a client does not use a condom, discussing condom use with FSWs, and providing FSWs with condoms). The composite score ranged from 0 to 5 with a higher score indicating a higher level of support for condom use.
Participants were asked about the overall frequency of condom use with clients and the number of times (0–3) they had used a condom during their last 3 episodes of commercial sex. In addition, proper use of condom was measured by asking women how often they put on a condom before penetration when having sex with a client. Overall frequency of condom use and proper use of condom were assessed along a 5-point Likert scale (0 = “never” -4 = “always”). The Cronbach α for this 3-item scale was 0.60. In path analysis, a composite score was created by summing overall frequency of condom use, number of times using condoms during last 3 sexual encounters, and frequency of proper condom use. The composite score ranged from 0 to 11 with a higher score reflecting a higher level of consistent and proper condom use.
First, descriptive statistics were calculated for all variables in the analysis. Bivariate analysis (Student t test) was performed to examine the association of condom use with determinant variables (e.g., HIV/STD knowledge, condom use skill, susceptibility and severity, perceived establishment-level support, self-efficacy, and perceived benefits of unprotected sex). Second, correlation analysis using Pearson correlation coefficients was conducted to examine the strength of associations between all observed variables.
Third, path analysis was performed to test the hypothesized model that specifies relationships between all observed variable (presented in Fig. 1), using the SAS System’s CALIS procedure. These analyses used the maximum likelihood method of parameter estimation and were performed on the covariance matrix. Standardized coefficients for all paths were estimated. The fit of the path model to the data was assessed using absolute and comparative fit indices. Absolute fit indices are based on the discrepancies between the covariance matrix of the data and the covariance matrix implied by the model. The absolute fit indices used in this study included the χ2 and root mean square error of approximation (RMSEA). Comparative fit indices are based on the comparison of the hypothesized model to a null model where no paths are included. The comparative fit index (CFI) and non-normed Fit Index (NNFI) were used to evaluate model fit in this study. Acceptable model fit is determined by a χ2 value close to zero, a probability value greater than 0.05, an RMSEA less than 0.08, and values of CFI and NNFI greater than 0.90.28,29 All the statistical analyses were performed using SAS 9.2 statistical software package (SAS Institute Inc., Cary, NC).
Characteristics of Study Sample
Data on 454 FSWs were collected through baseline survey administered between March and May 2004. The mean age of the women was 23.5 years. The women received an average of 5.7 years of formal education. More than half of the women were of Han ethnicity; Zhuang ethnicity constituted about one third of the sample. Three quarters of the women lived with other FSWs, 14% lived alone, and 12% lived with their family members or other relatives. Most (80%) of the women grew up in rural areas. Forty percent of the women were married. The women had worked as sex workers for an average of 12.2 months. On average, each woman had 2 sexual clients per week, and a maximum of 1.2 clients per day. There were 16 FSWs on average at each establishment.
As shown in Table 1, demographic characteristics and sexual behavior were compared between women who used condoms frequently (e.g., with high condom use score) and women who did not (e.g., with low condom use score). There was no association of condom use with age, ethnicity, living arrangement, marital status, duration of prostitution, and size of establishment. Women who used condoms frequently received a relatively higher education, were more likely to grow up in urban areas, and had significantly more sexual clients per week.
Descriptive statistics of all observed variables are displayed in Table 2. Bivariate analyses were conducted to test associations between these variables and condom use. Self-efficacy of condom use, perceived benefit of unprotected sex, and perceived establishment-level support for condom use were significantly associated with condom use. STD/HIV knowledge was also related to condom use. However, condom use skills and susceptibility/severity were not significantly related to the frequency of condom use.
The strength of associations between all observed variables was examined using Pearson correlation coefficients (Table 3). STD/HIV knowledge was significantly associated with condom use skills (r = 0.17, P <0.001) and susceptibility and severity (r = 0.14, P <0.01). STD/HIV knowledge and condom use skills were significantly related to self-efficacy (r = 0.35, P <0.001; r = 0.21, P <0.001) but not to perceived benefits of unprotected sex and condom use. Susceptibility/severity was significantly related to self-efficacy (r = 0.11, P <0.05) and perceived benefits (r = 0.19, P <0.001) but not to condom use. Self-efficacy was not significantly associated with perceived benefits. Condom use was positively associated with self-efficacy (r = 0.29, P <0.001) and perceived establishment-level support (r = 0.16, P <0.01); it was negatively associated with perceived benefits (r = −0.21, P <0.001). In addition, perceived establishment-level support was significant associated with STD/HIV knowledge (r = 0.27, P <0.001) and self-efficacy (r = 0.22, P <0.001).
Path Analyses of the Hypothesized Relationships
Structural equation model with latent constructs was initially used to test the specific hypotheses outlined in our conceptual model. The measurement model was tested using confirmatory factor analysis. However, the fit of the measurement model was poor. Poor fit may have occurred because indicator variables of some latent constructs (e.g., self-efficacy, measured by 5 yes/no questions) have less than 4 values.29 Therefore, composite variables were subsequently constructed and evaluated using path analysis.
The initial hypothesized model is displayed in Figure 1. There were 3 endogenous variables (e.g., self-efficacy, perceived benefits of unprotected sex, and condom use) and 3 exogenous variables (e.g., STD/HIV knowledge, condom use skills, and susceptibility and severity) in the model. The exogenous were allowed to covary. Estimation of this model revealed a significant model χ2 value (χ2 = 21.36, P = 0.0003). Bentler’s CFI was 0.85, Bentler-Bonett’s NNFI was 0.64, and the RMSEA value was 0.11 for this model, indicating that the fit of the model to the data could be substantially improved. In addition, several path coefficients were found not statistically significant, which also suggests the need of modification of the initial model.
In modifying the initial path model, we first reviewed the path coefficients and eliminated nonsignificant paths from the initial model. Evaluation of this preliminarily revised model revealed a significant model χ2 value (χ2 = 16.82, P = 0.019), although 3 other model goodness-of-fit indices were greatly improved (CFI = 0.92, NNFI = 0.82; RMSEA = 0.06). As a significant model χ2 can suggest a poor model fit, we attempted to further improve the model’s fit by incorporating 1 additional variable (e.g., perceived establishment-level support) into the model. Adding the path from perceived establishment-level support to condom use was consistent with previous bivariate correlation analysis that both variables were significantly correlated. The overall fit of this revised model was excellent (χ2 = 13.37, P = 0.147; CFI = 0.97; NNFI = 0.93; RMSEA = 0.038). The analysis revealed R2 value of 0.16 for condom use. This revised model was retained as the final model.
The standardized path coefficients for final model are presented in Figure 2. Significant pathways depict associations between observed variables. STD/HIV knowledge and condom use skills had significant direct effects on self-efficacy, with higher level of STD/HIV knowledge and condom use skills resulting in greater self-efficacy. Susceptibility/severity had a significant effect on perceived benefits, with higher level of susceptibility/severity being related to greater perceived benefits. Self-efficacy and perceived benefits had significant direct effects on condom use, with greater self-efficacy and less perceived benefits resulting in more frequent condom use. Perceived establishment-level support was positively related to condom use. Self-efficacy was not associated with perceived benefits of unprotected sex. The model accounted for 17% of the variance in condom use.
Data in the current study confirm and expand the findings from earlier studies. Consistent with the literature, self-efficacy has a direct positive effect on condom use.1,2 Self-efficacy also has a significant mediating effect on the relationship between STD/HIV knowledge, condom use skills, and condom use. Perceived benefits of unprotected sex are significantly related to condom use, with greater perceived benefits of unprotected sex resulting in less frequent condom use. In addition, perceived benefit mediates the effect of susceptibility and severity on condom use. These findings suggest that self-efficacy and perceived benefit are the most proximate determinants of condom use. They mediate the potential effects of STD/HIV knowledge, condom use skills, and perceived susceptibility and severity of STD and HIV infection on condom use behaviors.
The finding that perceived susceptibility and severity had a negative indirect effect on condom use through perceived benefits is partly consistent with a previous study that reported that women with a greater perceived susceptibility to STD and HIV infection was less likely to use condoms.1 The current study did not find a significant direct effect of STD/HIV knowledge, condom use skills, perceived susceptibility and severity of STD and HIV infection on condom use, which is inconsistent with findings from previous studies.1,4 However, the path analysis suggests that these 3 factors have indirect effects on condom use through self-efficacy and perceived benefits. This finding is important because it suggests that the knowledge and skill to exercise self-protective behavior are necessary but not sufficient. Future HIV prevention interventions should continue to promote self-efficacy, and change the actual benefit structure.
The current study did not find a significant relationship between self-efficacy and perceived benefits of unprotected sex, which is inconsistent with a previous study by Melkote et al.30 They found that perception of outcome expectancies including cost and benefits of condom use was related to condom use self-efficacy. Perhaps, this discrepancy was due to the differences of measurement in 2 studies. Previous research has also shown that perceived benefits of unprotected sex were better determinants of sexual risk-taking than were perceived benefits associated with condom use.10 Further study is needed to determine the complex relationship of perceived benefits of unprotected sex with self-efficacy and condom use among FSWs.
The positive relationship of perceived establishment-level support to condom use in the current study is consistent with findings from previous studies.3,13 Two factors might contribute to this positive relationship. First, a perceived supportive working environment may empower FSWs to negotiate condom use with their clients. Second, easy access to condoms in the establishment may increase the likelihood that FSWs will use condoms. Therefore, interventions should seek not only to increase access to quality condoms in the establishments but also to promote positive attitudes towards condom use among establishment owners and managers.
There are several potential limitations in the current study. First, despite the efforts to recruit participants across a broad range of the FSW population using ethnographic mapping, the study sample remains a convenience sample. Second, the use of self-report data raises questions about response bias, in that women may give socially desirable answers. Third, the refusal rates for establishments (67%) and the participants (78%) were both relatively high. Some owners/managers and FSWs might choose not to participate because commercial sex is illegal and stigmatized in China. Prior research has shown that Chinese FSWs are a heterogeneous population with considerable variation in their demographic characteristics and sexual behavior.31 Generalization of findings from this study to other FSWs is limited. Finally, some constructs of condom use determinants in this study have relatively low reliability estimates (e.g., 0.55 for self-efficacy, 0.66 for perceived establishment support). Further attention is needed to the development of more reliable and culturally appropriate measurement of these constructs.
Despite these methodological limitations, findings from the current study have important implications for the development of STD/HIV prevention interventions in the context of commercial sex work. First, our study has demonstrated that multiple levels of factors that are incorporated from 2 conceptual models (e.g., the health belief model and social cognitive theory) are related to condom use among FSWs. This may imply that multiple conceptual models may be needed to incorporate factors that affect condom use among FSWs. Therefore, future behavioral theories are needed to guide STD/HIV prevention research and interventions to integrate factors at different levels into their conceptual frameworks. Second, self-efficacy of condom use, perceived benefits of unprotected sex, and perceived establishment-level support for condom use are found to be proximate determinants of condom use among establishment-based FSWs. Future HIV prevention intervention efforts among FSWs should focus on self-efficacy training, increasing the awareness and recognition of the severity and long-term negative health and financial consequences of unprotected sex, and promoting supportive working environment.
1. Ford K, Wirawan DN, Reed BD, et al. AIDS and STD knowledge, condom use and HIV/STD infection among female sex workers in Bali, Indonesia. AIDS Care 2000; 12:523–534.
2. Ao T, Sam N, Manongi R, et al. Social and behavioural determinants of consistent condom use among hotel and bar workers in Northern Tanzania. Int J STD AIDS 2003; 14:688–696.
3. Kerrigan D, Ellen JM, Moreno L, et al. Environmental-structural factors significantly associated with consistent condom use among female sex workers in the Dominican Republic. AIDS 2003; 17:415–423.
4. Oladosu M. Consistent condom use dynamics among sex workers in Central America: 1997–2000. J Biosoc Sci 2005; 37:435–457.
5. Sedyaningsih-Mamahit ER, Gortmaker SL. Determinants of safer-sex behaviors of brothel female commercial sex workers in Jakarta, Indonesia. J Sex Res 1999; 36:190–197.
6. Thorpe L, Ford K, Fajans P, et al. Correlates of condom use among female prostitutes and tourist clients in Bali, Indonesia. AIDS Care 1997; 9:181–197.
7. Rosenstock IM, Strecher VJ, Becker MH. The health belief model and HIV behavior change. In: DiClemente RJ and Peterson JL, eds. Preventing AIDS: Theories and Methods of Behavioral Interventions. New York: Plenum Press, 1994.
8. Bandura A. Social cognitive theory and exercise of control over HIV Infection. In: DiClemente RJ, Peterson JL, eds. Preventing AIDS: Theories and Methods of Behavioral Interventions. New York: Plenum Press, 1994.
9. Pleck JH, Sonenstein FL, Ku LC. Adolescent males’ condom use: Relationships between perceived cost-benefits and consistency. J Marriage Fam 1991; 53:733–745.
10. Parsons JT, Halkitis PN, Bimbi D, et al. Perceptions of the benefits and costs associated with condom use and unprotected sex among late adolescent college students. J Adolesc 2000; 23:377–391.
11. Ntumbanzondo M, Dubrow R, Niccolai LM, et al. Unprotected intercourse for extra money among commercial sex workers in Kinshasa, Democratic Republic of Congo. AIDS Care 2006; 18:777–785.
12. Varga CA. The condom conundrum: Barriers to condom use among commercial sex workers in Durban, South Africa. Afr J Reprod Health 1997; 1:74–88.
13. Morisky DE, Peña M, Tiglao TV, et al. The impact of the work environment on condom use among female bar workers in the Philippines. Health Educ Behav 2002; 29:461–472.
14. Yang H, Li X, Stanton B, et al. Condom use among female sex workers in China: Role of gatekeepers. Sex Transm Dis 2005; 32:572–580.
15. Tucker JD, Henderson GE, Wang TF, et al. Surplus men, sex work, and the spread of HIV in China. AIDS 2005; 19:539–547.
16. Hong Y, Li X. Behavioral studies of female sex workers in China: A literature review and recommendation for future research. AIDS Behav
18. The UN Theme Group on AIDS in China. HIV/AIDS: China’s titanic peril, 2001 update of the AIDS situation and needs assessment report. Geneva, Switzerland: UNAIDS, 2002.
19. van den Hoek A, Fu Y, Dukers NH, et al. High prevalence of syphilis and other sexually transmitted diseases among sex workers in China: Potential for fast spread of HIV. AIDS 2001; 15:753–759.
20. Wang B, Li X, Stanton B, et al. Vaginal douching, condom use, and sexually transmitted infections among Chinese female sex workers. Sex Transm Dis 2005; 32:696–702.
21. World Health Organization (WHO). HIV/AIDS in Asia and the Pacific Region 2001. Geneva, Switzerland: WHO, 2001.
22. Ma S, Dukers NH, van den Hoek A, et al. Decreasing STD incidence and increasing condom use among Chinese sex workers following a short term intervention: A prospective cohort study. Sex Transm Infect 2002; 78:110–114.
23. Liao SS, He QY, Choi KH, et al. Working to prevent HIV/STIs among women in the sex industry in a rural town of Hainan, China. AIDS Behav 2006; 10:S35–S45.
24. Li X, Wang B, Fang X, et al. Short-term effect of a cultural adaptation of voluntary counseling and testing among female sex workers in China: A quasi-experimental trial. AIDS Educ Prev 2006; 18:406–419.
25. Rou K, Wu Z, Sullivan SG, et al. A five-city trial of a behavioral intervention to reduce sexually transmitted disease/HIV risk among sex workers in China. AIDS 2007; 21(suppl 8):S95–S101.
26. Guangxi Centers for Disease Control and Prevention. Guangxi HIV/AIDS sentinel surveillance report in 2003. In: China Centers for Disease Control and Prevention and National Sentinel Surveillance Group, eds. National HIV/AIDS Surveillance. Beijing, China, 2004.
27. Wright JC, Huston AC, Ross R, et al. Pace and continuity of television programs: Effects on children’s attention and comprehension. Dev Psychol 1984; 20:653–666.
28. Hu L, Bentler PM. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling 1999; 6:1–55.
29. Hatcher L, ed. A step-by-step approach to using the SAS system for factor analysis and structural equation modeling. Cary, NC: SAS Institute Inc, 2005;141–198.
30. Melkote SR, Skinner EC, Velu S. Factors associated with condom use among female commercial sex workers in India: A case study. Paper presented at: The annual meeting of the International Communication Association; Sheraton New York, New York City. Available at: http://www.allacademic.com/meta/p11935_index.html
31. Huang Y, Henderson GE, Pan S, et al. HIV/AIDS risk among brothel-based female sex workers in China: Assessing the terms, content, and knowledge of sex work. Sex Transm Dis 2004; 31:695–700.