Changes in attitudes, risky practices, and HIV and syphilis prevalence among female sex workers in Brazil from 2009 to 2016

Abstract Background: The present study aims at investigating the progress made toward controlling the human immunodeficiency virus (HIV)/AIDS epidemic among female sex workers (FSW) from 2009 to 2016. Methods: The baseline of respondent-driven sampling (RDS) study among FSW was carried out in 2009, in 10 Brazilian municipalities. In 2016, information on FSW were collected in 12 municipalities. The analyses took into account the dependence among observations, resulting from the recruitment chains, and the unequal probabilities of selection, resulting from the different network sizes. We analyzed changes in attitudes and risky behavior practices as well as variations in HIV and syphilis prevalence based on the comparison of 95% confidence intervals for each estimate. Results: Information on 2523 (2009) and 4245 (2016) FSW were analyzed. Commercial sex debut shifted to younger ages: while in 2009 the proportion of women who started sex work under 18 years old was 28.3%, in 2016 this percentage rose to 38.3%. The proportion of FSW affiliated to a nongovernmental organization (NGO) in defense of their rights (14.0%), in 2009, decreased to 7.8%, in 2016, as well as the proportion of FSW who received counseling on sexually transmitted infections (STI) in the past 6 months, from 47.5% to 24.4%. Relevant improvements were found for HIV testing in the last 12 months (from 20.3% to 39.3%). The proportions of those who were never tested for syphilis dropped from 57.9% to 48.5%. However, an opposite decreasing trend was found for the Pap smear examination in the last 12 months, decreasing from 43.6% to 31.5%. Regular condom use with clients significantly increased in the period. Regarding HIV prevalence, the 5% level was sustained and no significant differences were found, but syphilis prevalence was found to be more than 3 times higher in 2016 (8.5%) than in 2009 (2.4%). Discussion: Many are the challenges to be faced in attempting to reverse the upward trend of syphilis among FSW in Brazil. Despite the progress in condom distribution free of charge, it is necessary to increase awareness campaigns, emphasize the use, reaffirm STI counseling, and reiterate the need of regular syphilis screening in this key population group.


Introduction
Since the beginning of the human immunodeficiency virus (HIV)/ acquired immune deficiency syndrome (AIDS) epidemic in Brazil, in the early 1980s, HIV prevalence has remained at levels lower than 1% in the general population, and it has been largely "concentrated," with higher rates among most-at-risk population groups. [1] In this concentrated scenario, groups at higher risk for HIV infection play a key role in the dynamics of the epidemics. The spread of HIV infection is influenced by the nature and intensity of interactions between the most-at-risk groups and the general population. [2,3] Through mathematical modeling, it has been shown that in epidemics with low potential reproduction rates, modest interventions in high risk groups can significantly reduce HIV incidence. [4,5] In Brazil, injecting drug users (IDUs), men who have sex with men (MSM) and female sex workers (FSW) are considered the most-at-risk groups for HIV infection. [1] Other population groups considered key elements in the spread of HIV infection are those that serve as bridges between the general population and vulnerable groups, such as clients of FSW. [6][7][8] Conducting studies on subpopulations of HIV risk with conventional sampling strategies is complicated. [9] In addition to the small population size, these groups are in general linked to stigmatized behaviors or illegal activities and are considered to be hard to reach populations. [10] Carrying out population-based studies in order to generate representative estimates of HIV prevalence among FSW, broken down by variables of interest, such as age, educational level, and place of work, would require very large sample sizes and would not be feasible due to operational and cost difficulties. [11,12] It is estimated that FSW represent 0.8% of the Brazilian female population from 15 to 49 years of age, accounting for a half million women, approximately. [13] Prostitution is not considered a crime under the National Constitution, except if minors are involved. [14] In Brazil, efforts are being made to develop a series of crosssectional studies to monitor risk behavior practices of most-atrisk populations for HIV infection at the national level. In 2006, the Department of Prevention, Surveillance and Control of Sexually Transmitted Infections, HIV/AIDS and Viral Hepatitis (acronym in Portuguese-DDAHV) of the Brazilian Ministry of Health (MoH), promoted, in partnership with the Centers for Disease Control and Prevention (CDC), University of California San Francisco, University of Tulane, and Oswaldo Cruz Foundation (FIOCRUZ), the transfer of sampling methodology to access hard to reach population groups. [1] In 2009, research projects were carried out in 10 Brazilian cities among 3 population subgroups: MSM, [15] FSW, [16] and drug users (DU) [17] ; using respondent-driven sampling (RDS) as the recruitment method. [18] Following the DDAHV commitment of HIV surveillance among FSW, a second cross-sectional survey using the same methodology was carried out in 2016 and built on the 2009 survey, defined as the baseline for future comparisons.
To analyze data collected by RDS, specific methods of analysis were developed, considering the structure of dependence of the observations and the unequal probabilities of participant selection. [19][20][21] The present study aims at investigating the progress made toward controlling the HIV/AIDS epidemic and changes in attitudes and risk behavior practices for sexual transmitted infections (STI) among FSW from 2009 to 2016.

The baseline study, 2009
The baseline RDS study among FSW was carried out in 2009, aimed at estimating prevalence of HIV and syphilis and establishing knowledge, attitudes, and practices related to HIV infection and other STI in this population group.
To conduct the research, the DDAHV selected 10 Brazilian municipalities based on the importance of the local AIDS epidemic. The sample size (2500 women) was calculated to estimate a 6% HIV prevalence, with a 95% confidence interval and 2-tailed error of 1.5%, considering a design effect of 1.5. In each of the municipalities, the attempt was made to distribute the sample proportionally to the municipality population, while setting a minimum sample of 100 women ( Table 1).
The eligibility criteria were the following: women (biologically determined at birth) aged 18 years or older; who reported commercial sex at least once in the previous four months in each municipality, to assure recent sex work activity in the city; and accepted to participate in the study and signed the informed consent form. In addition, they should present a valid coupon and should not have participated in the study previously. The study consisted of a preliminary interview-to verify the eligibility criteria and to characterize the network size and work venue; a self-completed questionnaire; and rapid tests for syphilis and HIV. In all 10 municipalities, the fieldwork was carried out in health services. Before data collection, a qualitative preparatory investigation was performed in each municipality to facilitate the study implementation taking into account their respective prostitution characteristics.
In each municipality, 5 to 10 initial participants, named seeds, were selected with different sociodemographic and sex work characteristics. The initial focus was on well-connected FSW in the community and who reported extensive social networks during the preparatory survey.
Each seed received 3 coupons to distribute to other sex workers within her social network. The recruits of the seeds in the survey were considered the first wave of the study. After participating in the interview, each participant received 3 additional coupons to distribute to their peers and this process was repeated until the sample size was achieved in each site. RDS requires a system of primary and secondary incentives. The primary incentive in this study was a gift (makeup product), payment for lunch and transportation in addition to a reimbursement for their time lost from work (approximately US$15.00). The secondary incentive was a payment of US$10.00 for each recruited person who participated in the study.
The questionnaire was completed directly on a computer using ACASI (audio computer-assisted self-interview) and included the following aspects: sociodemographic and sex work characteristics; access to preventive activities; access and utilization of health services; knowledge, attitudes, and risky practices for HIV and other STI; history of HIV and syphilis testing; sexual behavior with steady partners and clients; violence and discrimination; and illicit drugs and alcohol use.
Rapid tests were performed using capillary blood samples drawn from participants according to the MoH protocols, including pre and post-test counseling. Participants who tested positive were referred to public health services for appropriate follow-up according to the MoH guidelines.
As for the statistical analysis, the data were weighted using the following question to measure the network size: "How many female sex workers do you know personally?." In each municipality, the expansion factors were inversely proportional to the network size. [22] The research project was approved by the Ethics Committee of the FIOCRUZ (Case no. 395/07).

The RDS study among FSW, 2016
Information on FSW aged 18 years or over were collected in 2016 in 12 Brazilian cities, a priori chosen by the DDAHV according to both, geographical criteria and the epidemiologic relevance of the HIV/AIDS epidemic in the country. The minimum sample size was set at 350 FSW in each city (Table 1). Women were eligible to participate in the study with the same inclusion criteria as the baseline study.
Fieldwork was conducted in health services located in the 12 cities. For each site, 6 to 8 seeds were chosen purposively, following formative qualitative research. Each seed received 3 coupons to distribute to other sex workers within her social network. After participating in the interview, each participant received 3 additional coupons to distribute to their peers and this process was repeated until the sample size was achieved in each site. Similar to 2009, primary (makeup products and reimburse-ment) and secondary incentives (approximately US$10.00 for recruitment) were given to each participant.
The questionnaire included the same modules applied in the baseline study, with some updates related to new prevention strategies, such as pre-and postexposure prophylaxis (PrEP and PEP). The questionnaire was designed for tablets and could be self-administered according to the willingness and capacity of the participant. Tests for HIV, syphilis, and hepatitis B and C were conducted by standard rapid tests using peripheral venous blood collection, according to protocols recommended by the Brazilian MoH. All tests occurred before the interview and all participants received pre-and posttest counseling. Participants who tested positive in any one of the rapid tests received additional post-test counseling, both to address the psychological impact and to encourage partner notification, and were also referred to public health services for follow-up.
The research project was approved by the Ethics Committee of the FIOCRUZ (Protocol 1.338.989).

Data analysis
The rationale for the data analysis in both studies was to use statistical methods appropriate for data collected with complex sampling design. The analyses took into account the dependence among observations, resulting from the recruitment chains, and the unequal probabilities of selection, resulting from the different network sizes of each participant. The samples were treated as stratified, with clusters, and with different selection probabilities. Each one of the municipalities composed a stratum and, in each one, the weighting was inversely proportional to the size of the network totaling the size of the stratum. [19] In order to take into account homophily-that is, the tendency of a participant to recruit peers with similar characteristics [23] and a potential overrepresentation of individuals with certain characteristics in the study population, the 3 participants invited by the same recruiter were considered a cluster. [19] In the present study, we compared the sample distributions obtained in the 2009 and 2016 RDS based on the comparison 95% confidence intervals for each estimate. The following characteristics were analyzed: Sociodemographic variables (age group; educational level; and race); commercial sex characteristics, such as type of work venue, price of the sexual encounter, starting age of sex work; frequency of drug use; participation in prevention activities (affiliated to/or participated in a nongovernmental organization (NGO) in defense of FSW rights; received condoms free of charge, and STI counseling in the past 6 months); history of HIV and syphilis testing; frequency of Pap smear examination; discrimination in health services; sexual behavior with steady partners and clients. Additionally, changes in HIV and syphilis prevalence were also analyzed.

Results
In Table 1 Results of Table 2 show similar sample distributions by race and age group, with a slightly higher proportion of FSW 40 years old or over in 2016, while the proportion of FSW with better educational level was higher in 2016.  (Table 2). There were no differences in the proportions of drug use between the periods, and the use of illicit drugs at least once a week also remained at approximately 25%.
In Table 3, results indicate that the already low proportion of FSW affiliated to an NGO in defense of their rights (14.0%), in 2009, further decreased to 7.8%, in 2016. Although, there was some evidence of an increase in free condom distribution from 77.2%, in 2009, to 81.8%, in 2016, the proportion of FSW who received counseling on STI in the past 6 months was significantly reduced from 47.5%, in 2009, to 24.4%, in 2016. However, in 2016, the quantity was considered sufficient for more than 70% among those who received free condoms (Table 3). Some progress was found for HIV testing. As shown in Table 3, the proportion of FSW who had been tested in the last 12 months increased from 20.3% to 39.3%, and the proportion of FSW who had never been tested decreased from 35.6% to 22.7%. Also, the proportions of those who were never tested for syphilis dropped from 57.9%, in 2009, to 48.5%, in 2016 (Table 3). However, an opposite decreasing trend was found for the Pap smear examination. The proportion of a Pap smear examination in the last 12 months decreased from 43.6%, in 2009, to 31.5%, in 2016. At the same time, perception of discrimination in health services remained at similar levels. In both studies, more than one-fifth of the participants felt discriminated against in health services for being a FSW. Table 4 results show improvements in safe sex with clients. Regular condom use significantly increased from 2009 to 2016. For example, the proportion of regular condom use with clients in vaginal sex increased from 69.9% to 80.5%. With steady partners, however, the improvement is unclear: although regular condom use rose from 2009 to 2016, the proportion of never using condoms also increased.
As to HIV prevalence, the 5% level was sustained and no significant differences were found from 2009 to 2016. However, syphilis prevalence was found to be more than 3 times higher in  (Fig. 1).

Discussion
In this article, we compare the results obtained from an RDS study among FSW carried out in 2016 with the findings of the baseline investigation conducted in 2009. Analysis for the overall aggregated data for Brazil shows consistency of the 2 RDS samples. Despite the different set of municipalities and distinct sample sizes by municipality, after weighting, the samples were similarly distributed by age group and race. Furthermore, the higher educational level in 2016 is certainly reflecting the recent improvements in education in Brazil. As to commercial sex characteristics, it is of concern the increase in the proportion of adolescents who started commercial sex under 18 years old, with 13% starting at age 14 years old or younger. Although prostitution is not an illegal occupation in Brazil, sexual exploitation of minors is criminalized by the Brazilian legislation. [14] Consequences of sexual exploitation of minors are reflected in the marginalization of female adolescents from health policies. Previous studies in Brazil have already emphasized the need of ongoing STI prevention activities among sexually active female adolescents. [24,25] Some advances have been achieved in distributing free condoms. The study showed not only an increase in proportion of FSW receiving condoms free of charge, but also a substantial increase in the proportion that considered the amount satisfactory. The expansion in condom distribution has already shown impact on safe sex practices. The regular use of condoms with clients rose considerably from 2009 to 2016.
On the other hand, the percentage of women who have participated in NGOs in defense of FSW rights is decreasing in Brazil and is currently at a low 8% level. Admittedly, the participation of the civil society was a key element to promote human rights, contributing to the Brazilian response to the HIV/ AIDS epidemic. [26] Nevertheless, with the reduction in funding, there has been a significant decrease in the number of NGOs and activists dedicated to most at risk populations. Findings of this   comparison analysis already reflect the consequences, such as the reduction of community-based counseling and health promotion. In the last 30 years, Brazil has undergone several changes in terms of health care. The country has evolved to a unified health system with profound changes in health care policies and a marked expansion of primary health care. [27] More recently, several initiatives were undertaken focused on increasing cervical cancer screening as well as HIV, syphilis and hepatitis testing in primary care services. [28,29] In fact, the results of this study confirm the improvements in HIV testing and the use of primary care units as the main source of HIV testing among FSW, with a decreasing proportion seeking voluntary counseling and testing (VCT) units and private labs.
Other initiatives to encourage HIV testing have been introduced in the country. These include mobile unities (trailers) and collaboration with NGOs for stimulating HIV testing, known as "Keep track" ("Fique sabendo") program. [30] Comparison with the baseline study already shows the impact of these initiatives. HIV testing in the last 12 months increased almost 20%. As the MoH currently has in place the policy of providing antiretroviral therapy to all individuals diagnosed with HIV, the improvement in the frequency of HIV testing is particularly relevant. [31] However, although cervical cancer screening is also available free of charge in primary healthcare units, the frequency of Pap smear screening did not increase from 2009 to 2016 and even showed a decreasing trend in this period. Perhaps because the importance of taking a gynecologic examination is not being sufficiently emphasized in health services. A recent study among FSW in China shows that only 15.3% had a lifetime Pap smear and specifically emphasizes the need of cervical cancer and human papillomavirus (HPV) awareness programs in addition to regular screening for these vulnerable women. [32] At the same time, the results of this study indicate a large and stable discrimination perception in health services for being a sex worker without any further improvement in this aspect. As has been pointed out before, stigma and discrimination within health services constitute one of the main barriers to STI control. [33] Moreover, the anticipation of stigma related to their sexual work, and possibly discrimination by health staff, may affect the disclosure of FSW status in health services. [34] However, if FSW do not reveal their status, primary care physicians would not be aware that they fall into the category of most-at-risk group and would not offer the available interventions. [35] Although a more satisfactory result was found for HIV prevalence, which was sustained at the previous level of 5%, the rise in syphilis prevalence, increasing from 2.5% to 8.4% in 7 years, is alarming. Additionally, the numbers of diagnosed cases of congenital syphilis are on the rise as well. [36] During the last 10 years, a resurgence of syphilis has occurred in many other countries. [37] In the United States, 2014 data from the CDC indicate a sharp increase in the rate of new cases of syphilis, predominantly in the most vulnerable groups. [38] Many are the challenges to be faced in attempting to reverse the upward trend of syphilis among FSW in Brazil. Despite the progress in condom distribution free of charge, it is necessary to increase awareness campaigns, emphasize the use of condom, reaffirm the importance of STI counseling, and reiterate the need of regular syphilis screening in this key population group. Due to the difficulties of discussing sexual behavior or disclosing their FSW status in health services, risky practices may not be appropriately addressed. In this context, there is a need to expand access to public health services and provide preventive screening for all sexually active women, regardless of age or occupation.