63797 RISKY BUSINESS? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey November, 2010 World Bank Europe and Central Asia Region, Human Development Department Global HIV/AIDS Program, Health, Nutrition and Population Hacettepe Training, Research and Services Foundation UN HIV/AIDS Joint Programme in Turkey The Ministry of Health, Turkey RISKY BUSINESS? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey World Bank Europe and Central Asia Region, Human Development Department Global HIV/AIDS Program, Health, Nutrition and Population Hacettepe Training, Research and Services Foundation UN HIV/AIDS Joint Programme in Turkey The Ministry of Health, Turkey November, 2010 Copyright @ 2010 The International Bank for Reconstruction and Development The World Bank 1818 H Street, NW Washington, DC 20433, USA All rights reserved The World Bank enjoys copyright under protocol 2 of the Universal Copyright Convention. This material may nonetheless be copied for research, educational or scholarly purposes only in the member countries of The World Bank. Material in this report is subject to revision. ABBREVIATIONS AND ACRONYMS AIDS Acquired Immune De�ciency Syndrome HATAM Hacettepe University AIDS Treatment and Research Center HIV Human Immunode�ciency Virus IDU Injecting Drug User PLHIV People Living with HIV MARP Most-At-Risk Population MSM Men who have Sex with Men MOH Ministry of Health MTCT Mother-to-Child-Transmission STI Sexually Transmitted Infection SW Sex Workers UN United Nations UNGASS United Nations General Assembly Special Session UNAIDS United Nations HIV/AIDS Joint Programme WB The World Bank VCT Voluntary Counseling and Testing Vice President: Philippe H. Le Houerou Country Director: Ulrich Zachau Sector Director: Mamta Murthi Sector Manager: Abdo S. Yazbeck Task Team Leader: Ethan Yeh Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 4 Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey Table of Contents Executive Summary ................................................................................................................................................. 7 I. Introduction ..................................................................................................................................................... 9 II. Study Methodology .........................................................................................................................................11 III. Who are these workers? ..................................................................................................................................13 IV. HIV Risk Factors.............................................................................................................................................15 V. Knowledge and Awareness of HIV .................................................................................................................25 VI. Factors related to Risky Sex............................................................................................................................27 VII. Conclusions and Recommendations................................................................................................................29 Bibliography ............................................................................................................................................................31 Annex 1. Survey locations and organizations involved in data collection ..............................................................33 Annex 2. HIV risks among At-Risk Mobile Worker Groups...................................................................................34 Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 5 Acknowledgements This report was the result of a collaboration between Hacettepe Training, Research and Services Foundation, UN HIV/AIDS Joint Programme in Turkey, the World Bank, Hacettepe University AIDS Treatment and Research Center (HATAM), and the Turkish Ministry of Health. The World Bank team was led by Ethan Yeh, and comprised of Rekha Menon, Patricio Marquez, Jennifer Manghinang, Elif Yonca Yukseker, Joseph Green, and Christopher Reinstadtler, under the overall support and guidance of Abdo Yazbeck, Mamta Murthi, and Ulrich Zachau. The Turkish project team was comprised of Aygen Tümer (HATAM), Serhat Ünal (Hacettepe University and HATAM), Ömer Kutlu (Ankara University), Cem Babadoğan (Ankara University), Gökhan Atik (Ankara University), Nevin Topkara (HATAM) and Mustafa Topkara (HATAM). The report bene�ted greatly from comments provided by Nedim Jaganjac, Son Nam Nguyen, Ina-Marlene Ruthenberg, Aparnaa Somanathan, Harsha Thirumurthy, Benjamin Welch, and David Wilson. Our thanks in particular to our partners Ela Aktürkoğlu, UNAIDS Focal Point and UN HIV/AIDS Joint Programme Manager in Turkey, and Ercan Bal, Ministry of Health, for their contribution to the realization of this project. Funding was provided by the World Bank Global HIV/AIDS Program, Health, Nutrition and Population as part of the UNAIDS Uni�ed Budget and Workplan and the UN HIV/AIDS Joint Programme in Turkey. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 7 Executive Summary Turkey remains a low HIV prevalence (<0.2%) the actual risk of HIV infection cannot be quanti�ed country, but the number of new HIV cases has been in this study due to the lack of HIV testing, and the increasing in recent years. HIV is concentrated among role of mobile workers in HIV transmission may most-at-risk populations (MARPs) including sex be lessened by the high proportion of circumcised workers (SW), men who have sex with men (MSM), workers (99% of workers are circumcised). and injecting drug users (IDUs), but there are concerns that HIV may spread, particularly through heterosexual What factors may drive these workers to engage in transmission with sex workers. risky sex? While this study cannot determine causal relationships, we can examine factors associated with While several studies have been conducted in Turkey risky sex. First, workers spend considerable amounts on sex workers, little evidence exists on their clients, of time away from home, providing opportunities to the demand side of the commercial sex market. This engage in casual and paid sex. Over 50% of workers is the �rst study of its kind in Turkey to be conducted spent at least 5 months away from home in the past among mobile workers, a group long thought to year. Sailors spent the most time away from home on be clients of sex workers. In February and March average, whereas truckers and construction workers 2010, quantitative and qualitative data were collected were more likely to spend shorter periods of time from 1,239 mobile workers in four sectors (maritime away from home. Tourism workers, on the other transportation, road transportation, construction, and hand, showed a bimodal distribution where they either tourism) and across four provinces (Istanbul, Izmir, spent very little time away from home or were gone Antalya, and Trabzon) to answer three key research for almost the entire year. questions: Second, 44% of workers report drinking alcohol at • What HIV-related behaviors are mobile workers least occasionally, and among all workers, sailors involved in? report the highest frequencies of alcohol use. Alcohol use has been consistently associated with risky sexual • What factors are associated with these HIV-related behavior with much supporting evidence for the behaviors? theory that, under the influence of alcohol, people are more likely to make irrational or risky decisions • What are the current levels of HIV knowledge and (MacDonald et al., 2000; Steele & Josephs, 1990). awareness among these workers? In terms of their HIV knowledge, almost all workers This study �nds that the primary HIV-risk behavior have heard of HIV/AIDS, and most workers which mobile workers engage in is unprotected sex. have a basic understanding of HIV prevention and Injecting drug use is not common, with only 1.1% transmission. Some critical misconceptions do of workers reporting ever injecting drugs. Instead, remain, however. For example, only 39% of workers mobile workers, particularly sailors and truck drivers, know that HIV cannot be contracted from a mosquito engage in frequent sex with multiple partners and with bite. Knowledge of one’s own serostatus was low sex workers. 44% of sailors and 42% of truck drivers as very few workers have ever been tested for HIV. report having 2 or more sexual partners in the last 3 Sailors are the exception, with 35% having been months, compared to 23% of construction workers and tested. The majority of HIV tests among workers were 30% of tourism workers. 35% of all workers report mandatory. engaging in occasional or more frequent sex with sex workers, and only 45% report using a condom in To explore the factors associated with risky sex their last sexual encounter with a sex worker. 45% in further depth, multivariate logit regressions of sailors and 38% of truck drivers report occasional were conducted to control for other background or more frequent sex with sex workers. Sex with characteristics such as age, income, and education. sex workers occurs with equal frequencies at home, Once all characteristics are taken into account, only while traveling in Turkey, and while traveling abroad, two risk factors remain signi�cantly associated with suggesting that much of the sexual activity with sex engaging in risky sex: alcohol usage and time away workers takes place in Turkey. It is worth noting that from home. This study �nds that workers who drink Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 8 alcohol often are 2.5 times more likely to engage in • Increase behavioral change and condom use. risky sex compared with those who do not. Time away Information and knowledge are only necessary from home has a much smaller association—a worker conditions for prevention, and other interventions who spends 5 additional months away from home is are also necessary to change behavior. Behavioral 34% more likely to engage in risky sex. Traveling change programs that are comprehensive, including abroad has little effect, likely because being away HIV education as well as other interventions such from home is the more important factor. Finally, HIV as counseling, trainings, social support programs, knowledge has little to no association with risky sex, and peer groups, are critical. Increasing access suggesting that knowledge alone is not enough for and use of condoms, particularly near the point behavioral change. of sexual contact, is also important to prevent HIV among mobile workers. Condom promotion The evidence presented here on HIV knowledge, efforts should perhaps target mobile workers at attitudes, and behaviors among mobile workers should key locations such as truck stops and ports. be used to inform future policy and research. Three key recommendations can be made based on these • Continue to improve VCT and STI services. results: Very few workers had been tested for HIV, especially voluntarily. Voluntary counseling and • Better evidence and surveillance. Biomarker testing (VCT) services should continue to be data is needed to understand the true risks of expanded to provide opportunities for individuals, sexual and drug behaviors, and the �ndings and their partners, to become aware of their own here could be combined with HIV testing status. Given that heterosexual transmission in data to understand the implications for HIV Turkey accounts for the majority of identi�ed cases transmission. Integrated bio-behavioral studies in Turkey, increasing access to and utilization of should be conducted among at-risk groups, STI services can also be an effective tool in HIV particularly among MSM and IDUs. Finally, prevention. STI services and programs should HIV sentinel surveillance, targeted among at-risk continue to be expanded and developed for target groups, should be further developed in Turkey. populations (e.g., sex workers) in Turkey. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 9 I. Introduction Reported HIV prevalence in Turkey has historically As these are only registered cases, the actual number been low, but the number of new cases has been of People Living with HIV (PLHIV) is likely increasing. According to the latest UNAIDS and to be higher. Turkey has a National HIV/AIDS Turkish Ministry of Health (MOH) �gures, the adult Program Framework for 2007-2011 that guides HIV prevalence in Turkey is less than 0.2% (UNAIDS, HIV strategy and policy in the areas of prevention, 2008a; UNAIDS, 2008b; Turkish MOH, 2008). From treatment, counseling, social support, monitoring and 1985 to 2009, a total of 3,898 HIV/AIDS cases were evaluation, and intersectoral collaboration. Under identi�ed (771 AIDS cases and 3,127 HIV cases). this framework, the Turkish MOH launched the HIV/ However, as Figure 1 shows, the number of new cases AIDS Prevention and Support Programme in 2007 has been increasing in recent years. that includes expanding HIV testing and treatment Figure 1. Identified HIV and AIDS Cases in Turkey, 1985-2009 Source: Turkish Ministry of Health, 2009 services, but Turkey still lacks systematic monitoring HIV prevalence rates as high as 2.5% to 3.5%.1 Some of HIV infections and sentinel surveillance. As a evidence suggests that HIV prevalence among IDUs result, HIV cases are probably underestimated (Tumer may be as high as 2.6% (Mathers et al. 2008). & Unal, 2006). Through an estimation study carried out in 2010, the number of PLHIV was estimated at Among these target groups, heterosexual sex with 4,600 for 2009 (UNAIDS Thematic Group, 2010). SWs appears to be the main driver of HIV infections in Turkey. There are an estimated 100,000 legal Available evidence suggests that the epidemic in and illegal SWs in Turkey, and condom use during Turkey is concentrated among most-at-risk populations intercourse with SWs has generally been low, ranging (MARPs) of sex workers (SWs), men who have sex from 35.8% to 42% reporting consistent condom use with men (MSM), and injecting drug users (IDUs). (UNAIDS Thematic Group, 2010). Legal SWs must For example, a study carried out in Ankara, Izmir be women of at least 21 years of age with Turkish and Istanbul found that HIV prevalence was 0.8% citizenship, registered, and subject to routine checks among SWs, 1.2% among MSM, and 1.5% among and treatment for HIV and other sexually transmitted IDUs (UNGASS, 2007). Studies on SWs have found infections (STIs). Unfortunately, a large proportion 1 In 2009, a sex worker study found that 2.5% of sex workers were HIV positive (Pink Life, 2009); however, HIV prevalence was zero among female sex workers, and 5% among transgender sex workers, suggesting that the MSM prevalence rate may be underestimated. A study on illegal sex workers in Istanbul found that 3.5% were HIV positive (UNGASS 2007). Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 10 of SWs are illegal and unregistered, increasing the transmission modes across two periods: 1985 to risk of HIV infection and transmission (UNAIDS 2003 and 2004 to 2009.The share of cases attributed Thematic Group, 2010). Indeed, Turkish MOH to heterosexual sex was 51% in the �rst period but �gures of modes of infection show that 57% of HIV/ increased 12 percentage points to 63% in recent years. AIDS cases were through heterosexual sex, by far the Most other transmission routes declined in their shares dominant transmission pathway (see Figure 2). Of or stayed fairly even.2 The proportion attributable to the remaining cases, 9% were among MSM and 4% IDUs decreased substantially from 6% to 2%, while among IDUs. Unfortunately, 26% of the cases have the proportion due to MSM increased just 2 percentage unknown causes. points from 8% to 10%. The low transmission due to IDUs, in particular, is in contrast to many other When changes over time are examined, heterosexual countries in Eastern Europe and Central Asia, where transmission becomes an even greater concern. Table IDUs have accounted for a majority of total HIV cases 1 examines the share of cases attributable to various (IHRD, 2008). Figure 2. Modes of HIV Transmission for Identified Cases from 1985 to 2009 Source: Turkish Ministry of Health, 2009 Table 1. Change in Modes of HIV Transmission for Identified HIV/AIDS Cases in Turkey Modes of Mode of Cases between Mode of Cases between Percentage point change Transmission 1985 and 2003 2004 and 2009 Heterosexual Sex 51% 63% 12% Homosexual/Bisexual 8% 10% 2% Injecting Drug Users 6% 2% -4% Blood transfusions 2% 1% -2% MTCT 1% 2% 0% Other 1% 1% -1% Source: Turkey Ministry of Health and HATAM This study focuses on, what is believed to be, a key who are often working-age males spending signi�cant client group of sex workers: mobile workers. While time away from home, are more likely to have multiple several studies have been conducted in Turkey on sexual partners and visit sex workers (World Bank, sex workers, little evidence exists on their clients, the 2009; Kulis et al., 2009; Marck, 1999). In a review demand side of the commercial sex market. Research of 87 surveys from around the world, for example, from other countries has shown that mobile workers, mobile workers consistently reported more sex with 2 While the share of cases with unknown transmission declined from 30% to 23%, suggesting some improvements in identi�cation, 41% of the cases in 2009 (217 cases) were due to unknown causes. This high number of unknown cases is a concern. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 11 sex workers than men in the general population (Carael • What HIV-related behaviors are mobile workers et al., 2006). The study found that in South and South involved in? East Asia, a median of 30% of mobile workers had sex • What factors are associated with these HIV-related with sex workers, and in Brazil, 21% engaged with sex behaviors? workers. Although data on mobile workers is limited • What are the current levels of HIV knowledge and in Eastern European and Central Asian countries3, a awareness among these workers? recent study in Georgia found that 24% of sailors and 52% of truck drivers reported having sex with sex It is important to note two caveats of this study. First, workers at least once a week (World Bank, 2009). no biomarker data (e.g., HIV testing) was collected While not speci�c to sex with sex workers, another from these workers, so conclusions regarding HIV study found that 37.7% of truck drivers in the Baltics prevalence or concentration of the epidemic cannot be engaged in casual sex while traveling (Kulis et al., made. Second, the sample did not include other at-risk 2009). groups or the general population, so it is dif�cult to Unfortunately, evidence suggests that this risky sexual know whether the behaviors are more or less risky in behavior has translated into higher HIV infection and other population groups. Where possible, comparisons transmission.4 In Azerbaijan, 1.54% of truck drivers with other studies or countries are presented, but travelling through the country were found to be HIV there are likely important differences in sampling and positive, much higher than the 0.2% general population methodology. The major contribution of this study, prevalence rate (Botros et al., 2009). Over 50% of however, is that it characterizes the risk factors within HIV cases in Croatia were among sailors (Stulhofer the mobile worker population in Turkey. et al., 2006), and 14% of HIV cases in Montenegro were found to be among tourist workers (Godinho et al., 2005). II. Study Methodology The objective of this study is to determine the HIV 1. Provinces in the study knowledge, awareness, and behaviors among four mobile worker groups in Turkey that may have elevated This study was conducted in four provinces in risks for HIV infection and transmission: sailors, Turkey with high estimated HIV prevalence and truck drivers, construction and tourism workers. the large estimated populations of mobile workers: International transportation, construction, and tourism Istanbul, Izmir, Antalya, and Trabzon. Istanbul, are some of the fastest growing sectors of Turkey’s Izmir, and Trabzon provinces are connected to economy5, and mobile workers in these sectors are Turkey’s international highways and are also tourist believed to be one of the main HIV transmission routes for future infections (Tumer & Unal, 2006). Mobile destinations. Antalya is a major commercial region and workers are not yet included as an at-risk group in tourist destination that provides international maritime Turkey’s National HIV/AIDS Program Framework, connections. Due to tourism and transportation work, and it is important to �rst understand their behaviors these four provinces also have large numbers of and potential risks. The three key research questions are: construction projects. 3 For example, the Carael et al. (2006) review did not contain estimates from mobile worker populations in Eastern European and Central Asian countries. 4 Please see Annex 2 for a longer discussion of the literature on HIV and mobile workers in international transportation, construction and tourism. 5 Both maritime and road transportation are signi�cant sectors of the economy due to Turkey’s unique position between Europe and Asia. For example, the Euro-Asian highways and the Black Sea Ring Corridor are important links for shipping and trade. Turkey’s foreign trade volume has considerably expanded since the 1990s, and consequently, shipping and transport along national and international roads and sea routes have been increasing. The construction sector has grown considerably from $5.4 billion in 2004 to $19.5 billion in 2007 and employs 1.1 million people, roughly 5.3 percent of total employment (State Planning Organization 2006). Tourism has grown signi�cantly in recent years, with the number of foreign tourists increasing from 10.4 million people in 2000 to 21.1 million people in 2005. At the same time, overall tourism revenue increased from $7.6 billion in 2000 to $18.2 billion in 2005. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 12 2. Methods and instruments non-form questions. Throughout the report, illustrative quotes from the focus group interviews are presented The �eld work was conducted by the Hacettepe Train- under the heading, “In their own words�. ing, Research and Service Foundation, in partnership with the Turkish MOH, Hacettepe University AIDS 3. Sample Treatment and Research Center (HATAM), UNAIDS, and the World Bank. Both quantitative and qualitative Mobile workers are often dif�cult to identify, and data were collected from workers in the four provinces thus, many studies utilize convenience samples or to provide a more thorough assessment of knowledge, location-based sampling. In this study, workers were attitudes, and behavior. The quantitative data was col- identi�ed through regional professional organizations lected using an extensive questionnaire that covered or “chambers� of workers. These organizations helped background characteristics, alcohol and drug use, to establish a sample frame, from which a sample sexual behavior, HIV knowledge and awareness, HIV of workers was selected for both quantitative and testing and other health behaviors. The questionnaire qualitative data collection.6 was piloted in Ankara and revised for understanding, to maintain con�dentiality, and to meet norms of cul- Initially, 75 respondents were to be selected from each tural sensitivity. of the four industries within each province for a total Qualitative data was collected through focus group sample of 1,200 workers for the quantitative survey, interviews, which followed an “interview� method plus an addition 64 workers as alternates in case of using an “interview form approach� as described in incomplete or incorrect responses.7 The surveys were Patton (1987) (see also Karasar, 2004; Yıldırım and administered between February and March of 2010.8 Şimşek, 2006). An interview form of 13 main questions The �nal survey sample, including the alternates, is a was prepared after consultation and review of other total of 1,239 workers, and the breakdown by province studies, but interviewers still had the flexibility to ask and industry is shown in Table 2. Table 2. Composition of Survey Respondents Sailors Truck drivers Construction Tourism Total Antalya 84 78 77 78 317 Istanbul 76 77 76 77 306 Izmir 85 75 75 70 305 Trabzon 83 76 75 77 311 Total 328 306 303 302 1,239 The surveys were administered in a variety of locations From the survey sample, a sub-sample of respondents where respondents worked such as aboard ships, at were selected for the focus group interviews. The initial truck stops, at hotels and at construction sites in the sample included 20 workers from each industry for a target regions. Annex 1 describes some of the sites total of 80 workers, but due to some non-responses, the where (and organizations through which) the survey �nal focus group sample was 76 workers. The breakdown data collection took place. of the qualitative sample is shown in Table 3. 6 The survey team attempted to include both informal and formal sector workers, but it is likely that informal sector workers are still under- represented. 7 There were 16 alternates in each of the 4 worker industries. 8 Adverse seasonal effects were not expected to bias the results. Both land and road transportation are expected to be quite active during that time, with little difference in activity across seasons. Similarly, the construction sector is active during Feb-Mar as well, although perhaps even more so than during other times. The main seasonal effect may be with the tourism workers, and thus, Antalya �eld work was conducted in March to allow for warmer weather for more tourism workers to be present. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 13 Table 3. Composition of Focus Groups Respondents Sailors Truck drivers Construction Tourism Total Antalya 5 4 5 4 18 Istanbul 3 5 5 4 17 Izmir 6 4 5 6 21 Trabzon 6 5 5 4 20 Total 20 18 20 18 76 III. Who Are These Workers? Background characteristics for the sample of workers Important differences exist across the different are presented in Table 4. Across the entire sample, the sectors, however. Female respondents in the sample average worker is male, 35 years old, completed lower are mostly in the tourism sector, accounting for or upper primary education, married, and has 2.4 19.0% of tourism workers. The average age among children. Almost all survey respondents are of Turkish truck drivers is the highest in the sample at 41.0 nationality, Muslim, heterosexual, and circumcised. years, while the tourism workers are the youngest Evidence has shown that male circumcision at 29.3 years. In terms of age distribution, the truck signi�cantly reduces the risk of HIV acquisition by drivers and sailors tend to be older with a greater men during heterosexual sex by approximately 60% percentage of workers between the ages of 31-50 (Auvert et al., 2005; Bailey et al., 2007; Gray et al., years, whereas tourism workers tend to be younger 2007). With 99.4% of workers circumcised, the risk of with 63.6% of workers between the ages of 21-30 HIV infection for these workers may be relatively low. years. Table 4. Background Characteristics By Worker Type Sailors Truck drivers Construction Tourism All Gender Male 99.0% 99.0% 100.0% 81.0% 94.9% Age Average 35.2 41.0 32.7 29.3 34.6 (9.3) (8.6) (9.0) (6.5) (9.4) ≤ 20 years 0.9% 1.0% 4.3% 3.3% 2.3% 21-30 years 36.6% 13.1% 47.2% 63.6% 40.0% 31-50 years 56.4% 71.9% 46.2% 33.1% 52.1% > 50 years 6.1% 14.1% 2.3% 0.0% 5.6% Education completed Less than Primary 3.7% 4.9% 9.6% 3.0% 5.3% Primary (Lower and Upper) 25.5% 62.2% 67.9% 26.4% 45.2% Regular or Vocational Secondary 58.4% 31.6% 22.2% 54.2% 41.8% University or Graduate 12.4% 1.3% 0.3% 16.4% 7.6% Marital Status Married 55.8% 78.5% 62.4% 46.2% 60.6% Single 34.6% 11.9% 32.3% 46.5% 31.4% Living together unmarried 1.6% 1.0% 2.0% 4.0% 2.1% Other1 8.1% 8.6% 3.3% 3.3% 5.9% Children Number of children 2.1 2.7 2.7 1.8 2.4 (1.1) (1.2) (1.4) (1.0) (1.3) Nationality Turkey 99.4% 100% 99.7% 98.3% 99.4% Religion Muslim 98.5% 100% 99.7% 98.3% 99.1% Sexual Orientation Heterosexual 97.8% 98.7% 98.7% 97.7% 98.2% Circumcision Circumcised 98.8% 100.0% 99.3% 99.6% 99.4% Notes: Standard deviations in parentheses. All percentages shown are percentages within a particular worker group (i.e., within the column) 1 “Other� includes Divorced/Separated, Widowed, Living separate. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 14 Sailors and tourism workers are the most highly Construction workers are the least educated with educated with over 50% of workers completing 67.9% completing either lower or upper primary regular or vocational secondary school, and over 10% education, followed by 62.2% of truck drivers with with university or higher education (see Figure 3). primary education.9 Figure 3. Highest Level of Education Completed By Worker Type Less than Primary Primary Secondary Higher than Secondary 68% 62% 58% 54% 32% 25% 26% 22% 16% 12% 10% 4% 5% 3% 1% 0% Sailors Truck Drivers Construction Tourism On average, the majority of workers are married, similar distribution of earnings as well, with 56% except for tourism workers, where 46.2% of workers earning between 500-1000 Lira a month. Truck drivers are married and 46.5% are single (see Table 4). The and sailors have a different income distribution, with highest proportion of married workers are among truck most workers in those sectors earning between 1001- drivers, where 78.5% of workers are married, and only 2000 Lira a month. Sailors are generally the most 11.9% self-report as being single. highly paid with 19% earning between 2001-3000 Although they are among the most educated Lira, and 14% earning over 3001 Lira each month. workers in the sample, tourism workers in our The latest GNI per capita for Turkey is USD $9,020 sample are paid the least on average with 60% for 2008, which would suggest that those workers earning between 500-1000 Lira (USD $315-$630) a earning between 500-1000 Lira per month are earning month (see Figure 4).10 Construction workers have a less than the average GNI per capita. Figure 4. Average Monthly Income (in Turkish Lira) By Worker Type 500 1000 Lira 1001 2000 Lira 2001 3000 Lira 3001 4000 Lira 4001 Lira and above 60% 54% 56% 47% 40% 31% 31% 21% 19% 10% 12% 4% 2% 1% 2% 0% 1% 4% 1% 1% Sailors Truck Drivers Construction Tourism Notes: Some worker percentages may not add up to 100% because of an “Other� income category. 9 51% of construction workers have only completed lower primary education, whereas 37% of truck drivers have only completed lower primary education. 10 For income, the survey asked respondents to estimate their monthly income in different categories, so an average across all workers (or by worker type) cannot be calculated. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 15 IV. HIV Risk Factors Based on the literature, we focus on four potential in the most sex with multiple partners, with 44% of factors that may increase mobile workers’ risks for sailors reporting 2 or more sexual partners in the last HIV infection and transmission: 3 months. 42% of truck drivers, 23% of construction workers, and 30% of tourism workers report 2 or 1) Risky sexual behavior more sexual partners in the last 3 months. Since the 2) Injecting Drug Use likelihood of reporting multiple partners may vary by 3) Time away from home gender, it should be noted that the percent of female 4) Alcohol usage workers is highest in tourism, with other sectors being predominantly male, which may influence the �gures The �rst two are behaviors that are most directly related in Table 5. It is worth noting that the majority of to HIV infection and transmission, whereas the last workers self-report that they did not have sex in the two are factors that may increase the opportunities and last 3 months (48.2% for all workers). likelihood of engaging in either risky sex or injecting drug use. One may think that most of the workers who are engaging in sex with multiple partners may be single. A. Risky Sexual Behavior However, a substantial proportion of married workers 1. Multiple sexual partners are engaging in sex with other partners. Figure 5 shows that at least 20% of married workers are engaging in As shown in Section I, unprotected heterosexual sex with other partners. The situation is particularly sex is believed to be the primary HIV transmission alarming for sailors and truck drivers where 38% and route in Turkey. Mobile workers may be at risk 37%, respectively, engaged in sex with other partners. for HIV infection because a substantial proportion While there is a lower probability of having sex of workers have multiple sexual partners. 35% of with multiple partners among married workers when workers self-report that they had 2 or more sexual compared to single workers, the potential risk of HIV partners in the last 3 months. Table 5 shows the transmission may be higher if the married partner is breakdown by sector: sailors appear to engage unaware of the worker’s sexual activity. Table 5. Proportion of workers and their number of sexual partners in the last 3 months by worker type No of partners Sailors Truck drivers Construction Tourism Total 0 42.6% 42.8% 53.5% 54.5% 48.2% 1 13.1% 15.2% 23.6% 15.9% 16.9% 2 16.4% 12.5% 12.9% 17.3% 14.7% 3 8.5% 10.4% 4.9% 7.2% 7.8% 4 7.2% 9.4% 3.5% 2.2% 5.7% 5 4.6% 5.1% 1.4% 1.8% 3.3% 6+ 7.6% 4.7% 0.4% 1.1% 3.5% Note: Percentages are by worker type. In their own words… “Sex with a stranger is an experience of every young man before marriage� Construction worker Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 16 Figure 5. Proportion of workers having sex with more than 1 partner by marital status and worker type Married Single 69% 59% 43% 38% 37% 35% 22% 26% Sailors Truck Drivers Construction Tourism Condom usage with other partners is relatively low Interestingly, truck drivers appear to use condoms among all workers with only 20.8% of all workers on the most, which may be due to the greater awareness average “always� using condoms in the past 3 months, globally about HIV and land transportion. Still, these and 11.7% reporting “often� using condoms. Figure condom usage �gures are relatively low in comparison 6 shows that among all workers, construction workers to those found in other studies. In a survey conducted are the least likely to use condoms when having sex among truck drivers in the Baltics, for example, 66.6% with a partner other than one’s spouse (44% “never� of truck drivers reported “always� using condoms and 21% “very little�). While 24% of tourism workers during casual sex while traveling, and only 18.4% report using a condom always, many of them also report reported “never� using condoms (Kulis, Chawla, low condom usage (52% “never� or “very little�). Kozierkiewicz, & Subata, 2009). In their own words… “Yes I would certainly prefer using condom but it may not always be possible.� Sailor Figure 6. Frequency of Condom Use with Partner other than Spouse in the last 3 months Never Very little Occasionally Often Always Sailors 22% 25% 24% 14% 15% Truck Drivers 24% 20% 17% 10% 29% Construction 44% 21% 11% 11% 14% Tourism 31% 21% 12% 12% 24% Unfortunately, married workers are less likely to use or “never�. This �nding suggests that the spouses condoms during sex with other partners. Figure 7 of married workers may be at elevated risk for HIV presents the same condom usage frequencies as in infection, and perhaps additional programs should Figure 6, but by marital status, and shows that 58% be targeted, not only at workers, but also at their of married respondents use condoms “very little� spouses.11 11 Although not shown here, the results by sector and marital status are similar to the results in Figure 6. A greater proportion of married construction (76%) and tourism workers (65%) report using condoms “never� and “very little�, in comparison to sailors (54%) and truck drivers (51%). Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 17 Figure 7. Frequency of Condom Use with Partner other than Spouse in the last 3 months by Marital Status Never Very little Occasionally Often Always Single 24% 17% 23% 16% 20% Married 33% 25% 13% 8% 21% 2. Sex with Sex Workers Figure 8. Frequency of sex with sex workers Sex with sex workers is one of the greatest HIV risk last sexual encounter, an extremely low percentage. factors among mobile workers, based on evidence Figure 9 shows that condom usage with sex workers from other countries around the world. In addition is fairly similar across sectors, with no statistically to sex with other partners, workers were asked signi�cant difference between worker types.12 Most speci�cally about interactions with sex workers. self-reported surveys tend to overestimate condom Figure 8 shows that an extremely high number of usage, so the actual percent of workers consistently workers have ever had sex with sex workers, with the using condoms with sex workers may be even lower. lowest share among tourism workers (49%) and the highest among sailors (73%). Figure 8 also shows Low condom usage with sex workers is a serious that sailors and truck drivers tend to have more sex concern among these mobile workers because most with sex workers—45% of sailors and 38% of truck other studies have found much higher condom usage. drivers report having occasional or more frequent sex For example, a 2008 study of Georgian mobile workers with sex workers, compared to 28% and 26% among found that 77% of Georgian truck drivers and 83% of construction and tourism workers. Overall, the high sailors used condoms during their last sexual encounter proportion of workers engaging in occasional or more with a sex worker (World Bank, 2009). A similar frequent sex with sex workers (average of 35% among study in Croatia found that 85% of sailors and 75% of all workers) is alarming because self-reported surveys construction workers had used condoms in their last generally underestimate risky sexual behavior. encounter with a sex worker (Stulhofer, Brouillard, Nikolic, & Greiner, 2006). In Vietnam, 75% of When asked about condom usage with sex workers, construction workers used a condom in their last sex only 45.1% of workers report using a condom in their with a sex worker (Population Council, 2003). 12 T-tests conducted and statistically signi�cant at 5%. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 18 Married workers appear to have less sex with sex that are having sex with sex workers are less likely to workers than single workers, as only 27% of married use condoms. Figure 11 shows that 41% of married workers report at least “occasional� sex with sex workers used a condom in their last sexual encounter workers compared to 47% of single workers (see with a sex worker, in comparison to 50% among single Figure 10). Unfortunately, those married workers workers. Figure 9. Condom use during last sexual encounter with a sex worker Yes No Don't Recall Sailors 42% 49% 9% Truck Drivers 45% 42% 12% Construction 47% 47% 6% Tourism 47% 47% 5% In their own words… When asked about using a condom… “No need. I am a man. I am Turkish and we do not need condoms.� Construction worker Figure 10. Frequency of sex with sex workers by marital status Never Very little Occasionally Often Always l Single 25% 28% 32% 11% 4% 1% Married 41% 32% 22% 5% Figure 11. Condom use during last sexual encounter with a sex worker by marital status Yes No Don't Recall Single 50% 45% 6% Married 41% 48% 11% Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 19 Given their time away from home, workers may be travel abroad for work. A substantial proportion of expected to �nd sex workers while away from home, both sailors and truck drivers report having sex with either in Turkey or abroad. Figure 12 shows where sex workers while travelling in Turkey, however, with workers had sex with sex workers in the last 3 months. more truck drivers reporting sex with sex workers in Workers were allowed to report multiple responses, Turkey (38%) than abroad (30%). Once again, this such as having sex with sex workers both in Turkey is partly due to the amount of time that truck drivers and abroad, so the percentages should not be added spend traveling either in Turkey or abroad, and does together.13 The �gures also include workers who did not suggest that sex work is more likely to take place not have any sex with sex workers. in either location. Construction (29%) and tourism workers (21%) report the highest percentages for engaging with sex workers at home. Interestingly, across all workers (at the bottom of the �gure), sex with sex workers takes place at home But how do workers generally encounter or �nd sex (16%), while traveling in Turkey (15%), and while workers? Table 6 shows that bars and nightclubs are traveling abroad (17%) in roughly equal proportions.14 popular locations for meeting sex workers, particularly This �nding suggests that much of the sexual activity for sailors (47%), construction (31%), and tourism with sex workers actually takes place within Turkey, workers (27%). Workers often meet sex workers in either at home or while traveling in Turkey. one’s place of lodging as well, with 36% of truck drivers and 25% of sailors meeting sex workers this Important differences exist within particular sectors, way. Unique among all workers, 33% of truck drivers however. Over 30% of sailors and truck drivers report meet sex workers while on the road, suggesting that having sex with sex workers while traveling abroad, in any prevention efforts will not only need to be in contrast to 3% among construction workers and 0.3% cities, but also along transportation corridors. Other among tourism workers. This result seems reasonable locations that workers report �nding sex workers are given that construction and tourism workers rarely by telephone, at brothels, and through friends. Figure 12. Travel and sex with sex workers in the last 3 months by worker type Notes: Multiple answers were allowed, so the bar percentages should not be added. The specific multiple answers (e.g., “At home� and “Traveling in Turkey�) are available upon request. In their own words… “They [sex workers] �nd us, they meet us on the way.� Truck driver 13 For example, 42% of all workers reported where they had sex with a sex worker, not 48% if the percentages were added together (16% + 15% + 17%). 14 There was no statistical difference in the proportions for “all workers�. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 20 Table 6. Where Sex Workers Are Found By Worker Type Sailors Truck drivers Construction Tourism All Internet 10% 4% 9% 7% 8% Bar/nightclub 47% 20% 31% 27% 31% Lodging place 25% 36% 15% 10% 22% On the Road 11% 33% 4% 3% 13% Other 10% 9% 17% 9% 11% Notes: Percentages are by worker type. Multiple answers were allowed, so the percentages in a column should not be added. Finally, the speci�c locations where sex with sex workers and tourism workers (21%). Construction workers takes place are shown in Table 7. Once again, multiple often have sex with sex workers in their house (21%), answers are allowed. There is considerable variation also a common response for sailors (18%) and tourism in the locations where workers had sex. Hotel is a workers (17%). Thirty-nine percent of truck drivers common response for sailors (29%), truck drivers (21%), report having sex with sex workers in their vehicle. Table 7. Specific Locations of Sex with Sex Workers in the last 3 months Sailors Truck drivers Construction Tourism All House 18% 10% 21% 17% 16% Brothels 8% 6% 13% 4% 8% Appointed house 13% 6% 6% 4% 7% Vehicle or Ship 12% 39% 3% 3% 14% Hotel 29% 21% 7% 21% 20% Outdoor 2% 6% 3% 1% 3% Notes: Percentages are by worker type. Multiple answers were allowed, so the percentages in a column should not be added. 3. Multiple Concurrent Partners There has been increasing focus on the effect of Table 8 shows the proportion of workers reporting that multiple and concurrent partnerships (MCPs) on HIV they are still having sex with at least two out of their transmission in recent years, particularly in sub-Saharan last 3 partners. A higher proportion of sailors (13.3%) Africa (Halperin & Epstein, 2004; Epstein, 2007; Morris and construction workers (10.8%) report having MCPs, et al., 2008). Epidemiological studies have established and the lowest share was among tourism workers an association between concurrent partnerships and HIV (7.8%). The second row shows workers who report transmission, but the role of MCPs in driving the HIV both irregular condom usage (not using a condom at last epidemic remains unclear, largely due to inadequate sexual intercourse with one of the last 3 partners) and and inconsistent data (UNAIDS, 2009). This study having MCPs. Over ten percent of sailors report both collected data on MCPs among workers as a baseline irregular condom use and having MCPs, suggesting that for comparisons with other risk groups and over time 80% of sailors with MCPs do not use condoms regularly. in Turkey. Although the de�nition for MCPs does not The proportion of workers with MCPs using condoms strictly follow UNAIDS de�nition, the survey did ask irregularly is roughly the same across all worker types— about concurrent partnerships at the time of the survey.15 roughly 80%. 15 The survey attempted to follow the UNAIDS guidelines for 3 consistent MCP questions, but unfortunately, due to constraints, the full set of 3 questions (When was the last time you had sexual intercourse with this person? When was the �rst time you had sexual intercourse with this person? Are you still having sex with this person?) was not used (see UNAIDS Reference Group on Estimates, Modelling, and Projection: Working Group on Measuring Concurrent Sexual Partnerships, 2009). However, in following with the guidelines, the survey did ask about sexual experiences with the last three sexual partners. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 21 Table 8. Workers with multiple concurrent partners Sailors Truck drivers Construction Tourism All Proportion with MCPs 13.3% 9.5% 10.8% 7.8% 10.5% Proportion reporting irregular condom use and MCPs 10.6% 7.6% 8.0% 6.6% 8.3% Notes: Percentages are by worker type B. Drug usage Injecting drug use is another important risk account for the majority of the injecting drug use factor, particularly in many Eastern Europe and (2.6% of sailors reporting ever injecting drugs). Central Asia countries. However, only 1.1% of these workers self-reported ever injecting drugs, Overall, it appears as though injecting drug use is not providing support for the evidence presented earlier a primary risk factor among mobile workers, although on low HIV transmission in Turkey due to IDUs. these �gures may be underestimated due to reporting Although the sample of IDUs is quite small, sailors bias. In their own words… “I cannot say that drug usage is common.� Construction worker Table 9. Workers reporting ever injected drugs Sailors Truck drivers Construction Tourism All 2.6% 0.3% 0.7% 0.7% 1.1% Notes: Percentages by worker type. While workers also use alcohol and cigarettes, the C. Time spent away from home number of workers that report using other types of drugs (e.g., marijuana, cocaine, ecstasy, etc.) is One of the primary factors that increases the extremely small. In fact, roughly 90% of respondents vulnerability of mobile workers to HIV infection report that they have never used any drugs aside from is their long absences from home. Table 10 shows alcohol or cigarettes. In the focus group interviews, that these workers spent, on average, 5.7 months some workers did suggest that there was some drug away from home in the past year. Sailors spent the usage, but not particularly common. A construction most time away from home, 6.8 months on average, worker remarked that, “I met two people using drugs compared to averages of 5.0 to 5.7 months for other in a construction work branch. And I met minimum 20 workers. These absences are signi�cant, representing people using drugs within 12 years.� almost half a year away from home. Table 10. Average months away from home in the past 12 months by worker type Sailors Truck drivers Construction Tourism All 6.8 5.2 5.0 5.7 5.7 (3.0) (3.2) (3.2) (4.1) (3.4) Notes: Standard deviations in parentheses. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 22 It is important to examine the distributions of re- distributions (or curves) shown are similar to a histogram, sponses, however, to truly understand the amount except that one can think of the bars as having been of time these workers spend traveling and away smoothed to create a curve.16 This smoothed curve from home. Figure 13 shows the distributions of represents the distribution for the general population workers according to the number of months they of each type of worker, not just the results from our spent away from home in the past year. Each of the sample. Figure 13. Distribution of Time Spent Away from Home .15 .1 Density .05 0 0 3 6 9 12 Time away from home (months) Sailors Truckers Construction Tourism Notes: Kernel density graphs with Epanechnikov kernel function and bandwidth of 1.2. For example, the curve for sailors is in blue, and the appears that of all the worker types, workers results show that among the general population of sailors, in the tourism sector are both least and most the largest proportion (about 12%) spend a little over 8 likely to spend time away from home. months away from home. The distribution for sailors is skewed to the right, suggesting that more sailors tend All workers appear to travel quite frequently, either to spend more than 6 months away from home. This within or outside of Turkey. Figure 14 shows that a is in contrast to truck drivers, shown in red, where the large proportion of all worker types travel domestically distribution is skewed to the left. Fewer truck drivers to other cities in Turkey, which is important because the spend over 6 months away from home, and most tend to opportunities to engage in risky behaviors may exist spend between 2 to 6 months away. Construction workers while away from home either inside or outside of Turkey. (in green) have closer to a normal distribution, although Workers most commonly report travel to major cities or few construction workers spend more than 9 months tourist areas such as Istanbul, Izmir, Antalya, and Ankara. away from home. Finally, the distribution for tourism Many tourism (57%) and construction workers (67%) workers (in orange) is bimodal with many workers travel domestically, in contrast to the small proportion spending between 1 to 3 months away from home, that travel internationally. Truck drivers appear to and many workers spending between 9 to 12 months travel the most domestically, with 88% of truck drivers away from home. In our sample, 26.2% of tourism having traveled to other cities in Turkey in the past year. workers spent only 1 month away from home. Thus, it Sailors and truck drivers are much more likely to travel 16 The distributions are non-parametric kernel density estimations of the density of a population. Using the histogram values as essentially point samples, the kernel density estimator generates the probability density function. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 23 internationally for work with 92% of sailors and 73% of work. When addressing the risks that these truck drivers traveling abroad in the last year. Tourism workers face, it will be important to take into and construction workers travel abroad much less with account the travel patterns, both domestically and only 12% and 7%, respectively, traveling abroad for internationally, of these workers. Figure 14. Proportion of workers who traveled internationally or domestically for work Travel abroad Travel in country 92% 88% 73% 67% 67% 57% 12% 7% Sailors Truck Drivers Construction Tourism Note: Percentages are by worker type. Table 11. Proportion of workers visiting a country in the last 12 months Country All workers Sailors Truck drivers Russia 34.7% 40.1% 29.7% Italy 18.0% 24.6% 12.2% Romania 16.8% 16.2% 19.8% Germany 12.6% 8.1% 15.3% Ukraine 11.7% 16.2% 7.7% Georgia 11.3% 3.0% 24.8% Bulgaria 9.2% 4.7% 17.1% Libya 8.4% 14.8% 1.4% Greece 6.5% 6.7% 7.7% Portugal 4.2% 6.7% 1.8% Note: Percentages are by worker type. Workers that did travel abroad, again mostly sailors place them in higher risk situations for HIV infection. and truck drivers, traveled to various countries all over Europe, Asia, Africa, and the Americas. Countries that Overall, the travel results suggest that these workers are extremely mobile: 97% of all workers having spent workers traveled to in the last year ranged from Spain at least 1 month away, and 87% having spent at least to Japan to Congo to Saudi Arabia. Table 11 shows 2 months away from home in the past year. Figure 15 the proportion of workers that have traveled to certain shows the percent of workers who have spent at least foreign countries. Overall, 34.7% of workers traveled X number of months away from home. This �gure is to Russia in the past year, with 40.1% of sailors having similar to Figure 13, but rather than focusing on the traveled to Russia, and 29.7% of truck drivers. Italy and probability distribution, this �gure shows the reverse Romania were also commonly visited by both sailors cumulative distribution, to better reflect the mobility of these workers.17 Sailors spent the most time away and truck drivers. Georgia, Bulgaria, and Germany were from home with 72% spending at least 5 months away visited by many truck drivers, whereas sailors traveled from home. For almost all worker types, at least ~50% often to Ukraine and Libya. Travel to areas with higher of workers are spending 5 or more months away, a HIV prevalence by both sailors and truck drivers may considerable amount of time away from home. 17 The �gure also shows data speci�c to our sample, rather than estimated for the general population distribution. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 24 Figure 15. Reverse cumulative distributions of workers that have spent a number of months away D. Alcohol Usage Research has shown that alcohol usage is consistently that they drink alcohol at least “occasionally�. Figure associated with risky sexual behavior (WHO, 2005; 16 shows the breakdown by sectors with sailors MacDonald et al., 2000; Steele & Josephs, 1990). drinking slightly more than the average worker, For example among mobile workers, a study in India followed by truck drivers. It is worth noting that showed that long distance truck drivers who consumed many workers report “never using alcohol�, ranging alcohol were 2.71 times more likely to visit a sex from 27% for sailors to 37% for construction workers. worker than those who did not drink (Chaturvedi et The most popular alcoholic beverages consumed (in al., 2006). order) were beer, raki, and vodka based on both the survey and focus group discussions. In Section VI, the In our study, alcohol consumption is fairly common relationship between alcohol usage and risky sex will among workers with 44.4% of all workers reporting be explored in greater detail. Figure 16. Frequency of Alcohol Use Always Often Occasionally Very little Never Sailors 12% 14% 28% 19% 27% Truck Drivers 3% 12% 30% 21% 34% Construction 7% 9% 23% 25% 37% Tourism 7% 10% 22% 27% 34% Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 25 V. Knowledge and Awareness of HIV Given that these workers engage in frequent risky sex, the average scores for each industry group. Figure 17 what is their current level of HIV knowledge? Ninety- shows the slightly right-skewed distribution of HIV four percent of workers have heard of HIV/AIDS, knowledge scores for all workers, suggesting that the indicating high awareness of the virus and disease.18 majority of workers have basic knowledge of HIV HIV knowledge is at a more basic level, however, with transmission and prevention. some misconceptions among workers. The same result is evident when examining responses to Workers were asked 13 questions about HIV speci�c HIV knowledge questions (see Figure 18). Workers transmission and prevention, and the average number are generally aware that HIV or AIDS exists, and they of correct answers for all workers is 8.7 (67.2% understand the basic principles of how it is spread (through correct on average). There is little difference between unprotected sex, injecting drugs, not using a condom). In their own words… “When we have sexual relations with foreign people, we use condoms to protect ourselves, and if we do not use it, we can get the disease� Tourism worker Figure 17. Distribution of Correct Scores on HIV Knowledge Questions .03 .02 Density .01 0 0 20 40 60 80 0 100 00% on questions) HIV Knowledge (0-10 Note: Distribution for all workers. However, critical myths and misconceptions about HIV 8-13% of workers have been tested for HIV. Sailors and AIDS remain among the workers. For example, were the exception with 35% having been tested for HIV 50% of respondents believe that HIV/AIDS can be cured. (see Figure 19). The main reason for the much higher More than half believe that the disease can be contracted proportion of sailors tested for HIV is that testing was by sharing food or drink with an infected person, and mandatory for many sailors. Among truck drivers and only 37% understand that abstinence can protect against construction workers, more than half of those tested for HIV infection. Only 39% of respondents know that HIV HIV were also due to mandatory testing. The low number cannot be contracted from a mosquito bite. of workers who have been voluntarily tested may partly be the result of few Voluntary Counseling and Testing Knowledge of one’s HIV status remains low as only (VCT) centers being available in Turkey. 18 Awareness of HIV/AIDS were higher than reported among general populations of currently married men (92.9%) in the Demographic and Health Survey (DHS) of 1998, and among ever-married women (88.1%) in the DHS and 2003. Unfortunately, the DHS 2008 did not ask about HIV/AIDS. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 26 Figure 18. Proportion with correct responses to HIV awareness and knowledge questions Notes: Percentages are for all workers. Figure 19. Self-Reporting HIV Testing Among Workers Note: Percentages may not add up to 100% because of an “other� category. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 27 VI. Factors related to Risky Sex The results so far have shown that these mobile sex workers “occasionally� or more often. A worker workers in Turkey engage in risky sexual behaviors, who responds “yes� to either of these two questions is spend large amounts of time away from home, drink categorized as one who engages in risky sex.19 alcohol, and have basic HIV knowledge. In this section, we examine the relationship between risky Table 12 shows that the proportion of workers sex and speci�c risk factors, such as whether workers engaging in risky sex is quite high, ranging from spend more time away from home or drink alcohol. 53% among sailors to 34% among tourism workers. Risky sex is de�ned here as either: i) using a condom Sailors and truck drivers engage in the most risky sex, “occasionally� or less with partners other than one’s which is consistent with the results presented earlier, spouse in the last 3 months, or ii) having sex with particularly the frequency of sex with sex workers. In their own words… “I wish to use [a condom] but the atmosphere then, at the bar, drunk may prevent us from thinking rationally and therefore at the end of the night it is not generally used.� Sailor Table 12. Prevalence of risky sex by different worker characteristics Sailors Truck drivers Construction Tourism All Risky sex Engages in risky sex 53.0% 48.6% 38.5% 34.0% 44.0% Frequency of Alcohol Usage Occasional or more 60.7% 61.1% 57.7% 52.9% 58.6% Very little or never 42.9% 37.8% 26.3% 21.2% 31.6% Time away from home in last 12 months At least 5 months 55.3% 56.7% 48.4% 50.0% 53.4% Less than 5 months 50.0% 44.0% 32.2% 40.0% 41.0% Travel abroad in last 12 months Traveled abroad 53.1% 53.6% 48.6% 50.0% 52.9% Did not travel abroad 46.2% 35.4% 36.8% 32.9% 35.5% HIV Knowledge >70% correct answers 51.9% 44.8% 34.6% 26.8% 40.4% <70% correct answers 53.8% 51.5% 41.7% 38.5% 46.6% HIV Testing Tested for HIV 54.7% 41.0% 35.7% 46.4% 48.6% Not tested for HIV 51.9% 49.8% 39.1% 32.6% 43.0% More importantly, Table 12 shows how the incidence to drinking, one may fully intend to use a condom, but of risky sex varies by different factors. First, alcohol after drinking, one may forget or decide that a condom usage appears to be strongly associated with risky is not as important (as one’s prior self had thought). sex, a �nding consistent with most of the literature In our sample, 59% of workers who drink alcohol (World Bank, 2009; WHO, 2005; MacDonald et al., occasionally or more engage in risky sex, whereas 2000; Steele & Josephs, 1990). The empirical research only 31% of those who rarely or never drink alcohol suggests that, under the influence of alcohol, people are engage in risky sex.20 More time away from home and more likely to make irrational or risky decisions. Prior travel abroad are both also associated with risky sex, 20 It is important to note that these are simply tabulations and do not imply causal relationships. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 28 although it is important to keep in mind that these two economics and psychology literature that shows that factors are correlated—those workers who spend more knowledge about HIV prevention and transmission time away from home are also those who more often is not well correlated with risky sex (Dinkelman et travel abroad. al., 2006). Similarly, research has shown that HIV testing interventions generally have little or no impact The association between HIV knowledge/testing on sexual behavior or condom usage (Thornton, and risky sex is more mixed. HIV knowledge, 2008; Coates et al., 2000; Wolitski et al., 1997). measured by the percent of correct answers in the To investigate these associations in more detail, multi- survey, is associated with slightly less risky sex when variate logit regressions were conducted to control for comparing those scoring above and below 70% on other background characteristics such as age, income, the 13 question assessment.21 For HIV testing, those and education. HIV testing is not included because the who have been tested are more likely to engage in number of workers tested is small. Overall, once other risky sex, although there is much variation between characteristics are taken into account, the magnitude the worker groups. Both of these results suggest of the associations with risky sex decrease consider- that knowledge alone may not be enough to lead to ably. Table 13 presents the odds ratios from multivari- behavioral change. Other non-informational factors ate logit regressions examining determinants of risky may have a much bigger influence, such as alcohol sex, including alcohol usage, time away from home, usage and time away from home, as well as factors foreign travel, and HIV knowledge. Across all worker that we have not measured, such as access to condoms types, alcohol usage remains the most highly associated and more. There has been growing behavioral with risky sex. Workers who drink alcohol often have, Table 13. Determinants of Risky Sex (Odds ratios) (1) (2) (3) (4) (5) All Sailors Truck drivers Construction Tourism Drinks alcohol at 2.506*** 2.458*** 2.422*** 3.894*** 2.772** least occasionally (0.370) (0.726) (0.678) (1.230) (1.347) 1.060** 1.114** 1.092* 1.074 1.059 Months away from home (0.0258) (0.0610) (0.0535) (0.0550) (0.0654) 1.451* 1.250 1.594 0.616 1.692 Traveled abroad (0.312) (0.611) (0.512) (0.323) (1.276) 0.995 1.006 0.998 0.996 0.974** HIV Knowledge Score (0.00406) (0.00812) (0.00759) (0.00995) (0.0124) 0.972*** 0.954** 0.964* 0.996 0.961 Age (0.0101) (0.0193) (0.0193) (0.0219) (0.0431) 0.556*** 0.424** 0.340*** 0.429** 1.791 Married (0.102) (0.143) (0.130) (0.182) (0.948) 1.148 1.470 0.744 0.533 2.369 Income 500-1000 TL (0.306) (0.749) (0.324) (0.802) (2.315) 1.300 1.312 0.860 0.580 2.575 Income 1001-2000 TL (0.292) (0.495) (0.320) (0.879) (2.128) 1.638 27.55*** 4.083 Education: Less than Primary (0.760) (30.31) (5.746) 1.525 2.289 7.345 0.874 0.577 Education: Primary (0.541) (1.344) (9.545) (0.427) (0.475) 1.120 1.510 6.951 0.730 0.397 Education: Secondary (0.370) (0.780) (9.089) (0.443) (0.264) Province controls Y Y Y Y Y Sector controls Y N N N N Observations 885 263 262 238 117 Notes: Coefficients presented are odds ratios. Robust standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1. Less than primary education variable not included in Columns 4 and 5 due to collinearity with other variables. 21 70% correct answers is approximately the median for the overall sample. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 29 In their own words… “From time to time you use a condom or not. If you are drunk you may ignore using it.� Tourism worker on average, 2.5 times the odds of engaging in risky sex risky sex. Alcohol usage, which is common among compared with those who do not. Construction work- workers, is the most signi�cantly correlated—workers ers who drink alcohol often (Column 4), in particular, who drink alcohol often are 2.5 times more likely to appear to be substantially associated with risky sex. engage in risky sex compared with those who do not. Further investigations should be conducted on the rela- Time spent away from home is also correlated with tionship between alcohol usage and risky sex,22 among risky sex. After controlling for other factors, a worker mobile worker and other at-risk populations. spending 5 additional months away from home is 34% more likely to engage in risky sex, which is of concern Time away from home is also signi�cantly associated because over 50% of workers spent at least 5 months with risky sex, with each additional month away away from home in the past year. Workers have a basic increasing the odds of risky sex by 1.06 times. For a understanding of HIV prevention and transmission, worker who traveled for 5 more months than another although some misconceptions do remain. worker, he or she would be associated with 1.34 times greater odds of engaging in risky sex.23 Other An important �nding of this study is that these mobile background characteristics that are signi�cantly workers are a diverse, heterogeneous group. Critical associated with risky sex include age (older workers differences exist in their background characteristics, are also less likely to engage in risky sex) and marital behaviors, and knowledge, and they should not be status (married men have almost half the odds of treated as a single at-risk group. For example, tourism engaging in risky sex). workers show a bimodal distribution for time spent away from home with some workers traveling for After controlling for these other factors, traveling almost the whole year, while others only spend 1 or abroad and HIV knowledge are no longer signi�cantly 2 months away from home. The majority of sailors associated with risky sex.24 More HIV knowledge and truck drivers travel internationally for work, appears to have little to no effect on risky sex, providing whereas very few construction and tourism workers support to the hypothesis that knowledge alone is not travel internationally. Truck drivers report the enough for behavioral change. highest condom usage, whereas construction workers report substantially lower condom usage. Policies, VII.Conclusions and programs, and future research should be mindful of Recommendations the differences between these worker groups. This study examines the behavioral risk factors and As mentioned earlier, there are important HIV knowledge of mobile workers, a population limitations of this study. This study describes group long thought to be key clients of sex workers. only the knowledge, attitudes, and behaviors of Using quantitative and qualitative data collected from mobile workers; it does not have biomarker data to four provinces, we �nd that unprotected heterosexual show how these risky behaviors are related to HIV sex is the primary HIV-risk behavior among these prevalence. This study is also unable to compare workers. Only 1.1% of workers report any injecting knowledge or behaviors from these workers with drug use, whereas in contrast, 35% of workers report those of other at-risk groups or with the general engaging in occasional or more frequent sex with population in Turkey. The sample is exclusively of sex workers, and only 45% report using a condom in mobile workers. Mobile workers are unlikely to be their last sexual encounter with a sex worker. Two the sole clients of sex workers, and their proportion of risk factors, in particular, are strongly correlated with the client pool cannot be determined from this data. 22 It is important to note that this association should not be interpreted as causation, however. This study is not suggesting that alcohol use is driving risky sex. High alcohol usage may be correlated with other unobservable factors that are causing risky sex. 23 1.34 odds ratio calculated from the coef�cient (0.0586) times the number of months (5), and then converting to an odds ratio (by raising e to the power of the resulting number). 24 In Column 1, the coef�cient on traveling abroad is statistically signi�cant at the 10% level, although the coef�cient is not signi�cant in the other columns with smaller sample sizes. Similarly, the coef�cient on HIV knowledge is signi�cant at the 5% level among tourism workers only. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 30 Risky sex may be just as prevalent in other population unprotected sex, and behavioral change programs groups, so one should be careful about drawing may want to address these behaviors as well. conclusions from this study on the epidemic drivers in Turkey. Another approach is harm reduction by increasing access and use of condoms. This study �nds These knowledge, attitude, and behavioral results that condom usage is low among these workers should still be used to inform future policy and and can be substantially improved. Awareness research, however, and several recommendations can and knowledge of condoms is vital, and so is be made: access to condoms, particularly near the point of sexual contact. Ensuring that condoms are 1. Better evidence and surveillance widely available on ships, at truck stops, along Rigorous evidence on HIV and related risk transportation corridors, at construction sites, behaviors remains scarce in Turkey. This study and in tourist destinations (e.g., hotels, bars) are is the �rst on mobile workers, and more evidence all critical steps to increase the practice of safe is needed for this population group and others. sex. Joint interventions that address alcohol over- Biomarker data is critical to understand the HIV consumption and HIV prevention, near points of risks of sexual and drug behaviors. Only with sexual contact with sex workers, should also be HIV prevalence and infection data is it possible explored. to make evidence-based decisions on how much attention is required and where resources should 3. Continue to improve VCT and STI services be directed. This evidence can be gathered at Finally, very few workers in the sample had two levels. First, further research and studies been tested for HIV, and those that had often had are necessary among at-risk groups in Turkey mandatory tests. While VCT services themselves that combine biomarker and behavioral data. are unlikely to be an effective intervention for Methodologies that are consistent across different HIV prevention, HIV testing presents a critical at-risk groups will enable better comparisons to opportunity for increasing awareness of one’s own identify which groups are driving the epidemic risks and for one-on-one education. As part of a in Turkey. The relationship between alcohol use comprehensive prevention program, VCT services and risky sex among mobile workers, in particular, can: i) help increase knowledge and awareness warrants further research. Second, enhanced HIV among HIV-positive persons (particularly for surveillance is necessary. In Turkey, there is no treatment), ii) provide testing and counseling to HIV/AIDS sentinel surveillance, so HIV serologic partners of HIV-positive persons, and iii) provide data currently come from inconsistent testing of HIV data for a national surveillance system. different groups—blood donors, sex workers, military recruits, pre-surgical patients, pre- Given that heterosexual transmission in Turkey marital couples, antenatal females, etc. Instead, accounts for the majority of identi�ed cases in HIV sentinel sites can be established (e.g., drug Turkey, increasing access to and utilization of STI treatment facilities, STI clinics), and data from services can be an effective tool in HIV prevention. sentinel groups of at-risk populations can be Studies have consistently shown that testing and collected over time at selected sites. treatment for STIs, particularly targeted at key at-risk population groups (e.g., sex workers) can 2. Increase behavioral change and condom use reduce HIV infection transmission. STI services Evidence presented here (and in other literature) and programs should continue to be expanded and suggests that increasing HIV knowledge and developed for target populations in Turkey. awareness is not enough to prevent HIV infection. Information and knowledge are only necessary conditions for prevention, and other interventions are also necessary to change behavior. Behavioral It is important to build on the information presented change programs that are comprehensive, including here with more rigorous evidence among MSM, IDUs, HIV education as well as other interventions and other risk groups. With better evidence, policies such as counseling, trainings, social support and programs can be targeted to the speci�c behavior programs, and peer groups, are critical. This study patterns and risks in Turkey. Only by remaining suggests that reducing time away from home vigilant and evidence-based will Turkey ensure that and alcohol usage may be effective in reducing HIV prevalence remains low. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 31 Bibliography Auvert, B., Taljaard, D., Lagarde, E., Sobngwi- El-Sayyed, N., Kabbash, I., & El-Gueniedy, M. (2008). Tambekou, J., Sitta, R., & Puren, A. (2005). Knowledge, attitude and practices of Egyptian Randomized, controlled intervention trial of male industrial and tourist workers towards HIV/AIDS. circumcision for reduction of HIV infection risk: East Mediterr Health Journal,14(5):1126-35. the ANRS 1265 Trial. PLoS Med, 2(11):e298. Erratum in: PLoS Med. 2006, 3(5):e298. Epstein, H. (2007). The Invisible Cure: Africa, the West, and the Fight Against AIDS. New York: Bailey, R.C., Moses, S., Parker, C.B., Agot, K., Farrar, Strauss and Giroux. Maclean, I., Krieger, J.N., Williams, C.F., Campbell, R.T., & Ndinya-Achola, J.O. (2007). Godinho, J., Jaganjac, N., Eckertz, D., Renton, A., & Male circumcision for HIV prevention in young Novotny, T. (2005). HIV /AIDS in the Western men in Kisumu, Kenya: a randomised controlled Balkins: Priorities for Early Prevention in a High trial. Lancet, 369(9562):643-56. Risk Environment. Washington: The World Bank Botros, B., Aliyev, Q., Saad, M., Michael, A., Sanchez, Group. J., Carr, J., & Earhart, K. (2009). HIV infection and Halperin, D.T. & Epstein, H. (2004). Concurrent associated risk factors among long-distance truck sexual partnerships help to explain Africa’s high drivers travelling through Azerbaijan. International HIV prevalence: implications for prevention. Journal of STD & AIDS, 20:477-482. Lancet, 364(9428): p. 4-6. Carael M., E. Slaymaker, R. Lyerla, & S. Sarkar. (2006). Clients of sex workers in different regions IHRD (2008). Harm Reduction Developments of the world: hard to count. Sexually Transmitted 2008. New York: International Harm Reduction Infections (82) (Suppl. III): iii26-iii33. Development Program of the Open Society Institute. Chaturvedi, S., Singh, Z., Banerjee, A., Khera, A., Joshi, R., & Dhrubajyoti, D. (2006). Sexual Kanda, K., Obayashi, Y., Ditangco, R., Matibag, Behavior Among Long Distance Truck Drivers. G., Yamashina, H., Okumura, S., et al. (2009). Indian Journal of Community Medicine , 31 (3), Knowledge, Attitude, and Practice Assessment of 153-156. Construction Workers for HIV/AIDS in Sri Lanka. Journal of Infection in Developing Countries , 3 Coates, T. J., Grinstead, O. A., Gregorich, S. E., (8), 611-619. Heilbron, D. C., Wolf, W. P., Choi, K. H., Schachter, J., Scheirer, P., van der Straten, A., Kamenga, M. Kulis, M., Chawla, M., Kozierkiewicz, E., & Subata, C., Sweat, M. D., De Zoysa, I., Dallabetta, G., E. (2009). Truck Drivers and Casual Sex: An O’Reilly, K. R., van Praag, E., Miller, D., Ruiz, M., Inquiry into the Potential Spread of HIV/AIDS in Kalibala, S., Nkowane, B., Balmer, D., Kihuho, the Baltic Region. Washington: The World Bank F., Moses, S., Plummer, F., Sangiwa, G., Hogan, Group. M., Killewo, J., Mwakigile, D. and Furlonge, C. (2000). “Ef�cacy of Voluntary HIV-1 Counselling MacDonald, T. K., Fong, G. T., Zanna, M. P., & and Testing in Individuals and Couples in Kenya, Martineau, A. M. (2000). Alcohol myopia and Tanzania, and Trinidad: A Randomised Trial,� condom use: can alcohol intoxication be associated Lancet, 356 (9224), 103–112. with more prudent behavior? Journal of Personality Currie, R. R. (1997). A Pleasure-Tourism Behaviors and Social Psychology, 78, 605–619. Framework. Annals of Tourism Research , 24 (4), 884-897. Marck, J. (1999). Long-Distance Truck Drivers’ Sexual Cultures and Attempts to Reduce HIV Risk DPT. (2006). State Planning Organization Development Behavior Amongst Them: A Review of the African Plan 2007-2013. Ankara: Turkish Department of and Asian Literature. In J. Caldwell, P. Caldwell, Public Transport. J. Anar�, K. Awusabo-Asare, J. Ntozi, & I. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 32 Morris, M., S. Goodreau, & Moody, J. (2008). Tumer, A., & Unal, S. (2006). HIV/AIDS Epidemiology Sexual Networks, Concurrency, and STD/HIV, in in Light of Current Information. 9-27. Sexually Transmitted Diseases, K.K. Holmes, et al., Editors. McGraw Hill Medical: New York. p. Turkish Ministry of Health. (2009). Mode of HIV 109-126. Tranmissions by Year. Ankara: Turkish Ministry of Health. Nzyuko, S., Lurie, P., McFarland, W., Leyden, W., Nyamwaya, D., & Mandel, J. (1997). AIDS , 11 UNAIDS. (2009). M&E for MCP Programmes, July (Supp 1), S21-26. 2009. Orubuloye, Resistance to Behavioral Change to UNAIDS. (2008a). Epidemiological Fact Sheet on Reduce HIV/AIDS Infection in Predominantly HIV and AIDS: Turkey 2008 Update. Ankara: Heterosexual Epidemics in Third World Countries United Nations. (pp. 91-100). Canberra: The Australian National UNAIDS. (2008b). 2008 Report on the global AIDS University. epidemic, July. Pink Life. (2009). HIV/AIDS Research and Testing Among Sex Workers. Ankara: Pink Life LGBTT UNAIDS Reference Group on Estimates, Modelling, Solidarity Association. and Projection: Working Group on Measuring Concurrent Sexual Partnerships. (2009). Podhisita, C., Wawer, M., Pramualratana, A., “Recommended De�nition, Measures, and Kanungsukkasem, U., & McNamara, R. (1996). Research on Concurrent Sexual Partnerships Multiple Sex Partners and Condom Use Among and HIV Transmission in Sub-Saharan Africa�, Long-Distance Truck Drivers in Thailand. AIDS unpublished manuscript. Education and Prevention , 8 (6), 490-498. UNAIDS Thematic Group. (2010). HIV/AIDS in Population Council.a (2003). Expanding Workplace Turkey: Situation Analysis. Ankara: United HIV/AIDS Prevention Activities for a Highly Nations. Mobile Population: Construction Workers in Ho Chi Minh City. August. WHO (2005). Alcohol Use and Sexual Risk Behavior: A Cross-Cultural Study in Eight Countries. Steele, C. M., & Josephs, R. A. (1990). Alcohol Department of Mental Health and Substance myopia: its prized and dangerous effects. American Abuse. Geneva. Psychologist, 45, 921–933. Wolitski, R. J., MacGowan, R. J., Higgins, D. L., & Stulhofer, A. (2004). HIV/AIDS related knowledge, Jorgensen, C. M. (1997). “The effects of HIV attitudes and sexual behavior in Croatian migrant counseling and testing on risk-related practices workers. Zagreb: International Organization for and help-seeking behavior,� AIDS Education and Migration Mission Croatia. Prevention, 9 (3 Suppl), 52–67. Stulhofer, A., Brouillard, P., Nikolic, N., & Greiner, N. World Bank. (2009). Knowledge, Attitudes, and (2006). HIV/AIDS and Croatian Migrant Workers. Behavior Related to HIV/AIDS Among Transport Collegium Antropologicum. Sector Workers. Washington: The World Bank Group. Thornton, R. L. (2008). “The Demand for, and Impact of, Learning HIV Status.� American Economic Yıldırım, A. and Şimşek, H. (2006). Nitel araştırma Review, 98(5): 1829–63. yöntemleri. Ankara: Seçkin Yayınevi. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 33 Annex 1. Survey locations and organizations involved in data collection Survey location/Organization Industry Sample Trabzon Türkiye Denizcilik İşletmeleri A.Ş. Sailors 83 Trabzon Liman İşletme Müdürlüğü (3 Ro Ro Ships) Trabzon Yeteroğlu Turizm Nakliyat Tic. Ltd. Ştd. Truckers 76 Trabzon Aydemirler Denizcilik Nakliyat Trabzon Özdoğan İnşaat Taahhüt Ticaret Limited Şirketi Construction 75 Construction Job Site Trabzon Kormas İnşaat İthalat ve İhracat Limited Şirketi Construction Job Site Trabzon Usta Park Hotel Tourism 77 TRABZON TOTAL 311 İstanbul Zeytinburnu Liman İşletmeleri San. ve Tic. A.Ş. (Zeyport) Sailors 76 İstanbul Zeytinburnu Liman İşletmeleri San. ve Tic. A.Ş. (Zeyport) Truckers 77 İstanbul Edem İnşaat A.Ş. Construction 76 Sabiha Gökçen Hava Limanı Construction Job Site İstanbul Nippon Hotel Tourism 77 İstanbul Point Hotel İSTANBUL TOTAL 306 Antalya Port - Akdeniz Antalya Limanı Antalya Denizcilik Sailors 84 Antalya City Municipality Marina Antalya Büyük Liman Free Zone TIR parking zone Truckers 78 Antalya Minor Port Customs TIR parking zone Antalya Ekol İnşaat Sanayi Ticaret Ltd. Şrt. ( at 3 different Construction 77 construction sites) Antalya Crowne Plaza Tourism 78 Antalya Rixos Hotel ANTALYA TOTAL 317 İzmir TCDD İzmir Port Port Operation Directorate Sailors 85 İzmir Atlas Gemi İşletmeciliği İzmir Kaşif Denizcilik İzmir Canbaz Gemicilik ve Ticaret İzmir International Transporters Association Truckers 75 U.N.D. Gürpınar Customs Operations TIR Parking Zone İzmir UKAT International Truckers İzmir Selamoğlu International Transport TIR Parking Zone İzmir Folkart Yapı Sanayi Ticaret A.Ş Construction 75 İzmir Kuryap İnşaat Sanayi Ltd. Şti. İzmir Best Western Konak Hotel Tourism 70 İzmir Hotel Marla İzmir Beyond Hotel İZMİR TOTAL 305 GRAND TOTAL 1,239 Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 34 Annex 2. HIV risks among At-Risk Mobile Worker Groups Sailors HIV/AIDS Commission, 2007). A recent study in Georgia also found that 50% of sailors reported having Research has found that sailors often engage in sex with sex workers at least 2 or 3 times a month, and risky behaviors that can lead to HIV infection and 83% reported using a condom during their last sexual transmission.25 Long periods at sea, extensive travel, encounter (World Bank, 2009). and the ready availability of sex workers in port areas can create opportunities for sailors to engage in casual and paid sex. In East Asia, seafarers are one of the Truck drivers largest mobile population groups, and HIV prevalence has been estimated to be as high as 22 percent in Perhaps the most extensive literature on mobile workers some areas of the Mekong region including Vietnam, and HIV concerns truck drivers, where research from Thailand, and Cambodia (World Bank, 2009). A study countries around the world has shown that road of Croatian sailors found that 90% of Croatian HIV transport workers are a key at-risk group. Like sailors, infections resulted from heterosexual sex abroad and truck drivers travel long distances, spend considerable that over half of HIV/AIDS cases in Croatia were time away from home, and often come into contact among seafarers (Stulhofer et al., 2006). The study with sex workers and other partners at truck stops and also found that 33% of Croatian sailors had multiple border crossings (World Bank, 2009). A review of partners over the last year, 27% of whom did not use various global country reports has found that 30% of condoms regularly. In Montenegro, HIV prevalence truck drivers reported having sex with sex workers in among sailors has been found to be as high as 6%, the last year (Carael et al., 2006). Unfortunately, this and 15% of identi�ed HIV/AIDS cases are among risky behavior has led to much higher HIV prevalence sailors. (Godinho et al., 2005; Montenegro National among truck drivers (see Table 14). Table 14. Transport Sector and General Population HIV Prevalence in Several Countries Country General Population HIV Prevalence Transport Sector HIV Prevalence (2008 (UNAIDS 2007) UNGASS Reports) Benin 1.8 5 Cameroon 5.4 16.3 Congo, Dem. Rep. of 3.2 3.3 Eritrea 2.4 7 Ghana 2.3 4 Guinea 1.5 7.3 Malawi 14.1 14.7 Mali 1.7 2.5 Niger 1.1 1.7 Rwanda 3.1 16.1 Nepal 0.5 1 Malaysia 0.5 3.7 Croatia <0.1 .2-.6 Source: Adapted from World Bank, 2009 25 There is also an extensive literature on �shermen, particularly in sub-Saharan Africa, and the sex-for-�sh market. Risky Business? HIV Knowledge, Attitudes, and Behavior Among At-Risk Mobile Workers in Turkey 35 Speci�c country examples include: migrant and mobile workers, perhaps leading to localized hot spots for risky sexual behaviors. The introduction of a large group of men with disposable • In Azerbaijan, 1.54% of truck drivers traveling income can motivate local residents to engage in sex through the country were found to be HIV positive, for direct or indirect payment (Nzyuko et al., 1997). much higher than the 0.2% general population prevalence rate (Botros et al., 2009). A Sri Lankan study, for example, found that two- • A study of truck drivers in the Baltics found that thirds of construction workers lived away from home 37.7% engaged in casual sex while traveling, one- (Kanda, et al., 2009). HIV prevention knowledge third of whom had inconsistent condom use (Kulis was low as only 37% knew how to use a condom et al., 2009). and 55% understood that condom use could prevent HIV transmission. Workers who did not travel were • In Thailand, 23% of male truck drivers currently more likely to be sexually active as they tended to living with a spouse had sex with a sex worker, travel home less often to see their spouses. In Croatia, 13% had sex with a non-commercial sex partner, construction workers were found to have the lowest and 8% reported both types of extramarital HIV knowledge and also the lowest rates of condom encounters (Podhisita, Wawer, Pramualratana, usage (Stulhofer, 2004). Kanungsukkasem, & McNamara, 1996). Only 58.5% of truck drivers who reported encounters with a sex worker used condoms consistently. Tourism • A study of long distance truck drivers in India found that 57% had sex with sex workers, and Like construction workers, tourism workers frequently only 7% used a condom consistently (Chaturvedi, move to cities or destinations where work is available. Singh, Banerjee, Khera, Joshi, & Dhrubajyoti, Tourist work is often seasonal causing workers to 2006). Married drivers were as likely as single spend time away from their families. Moreover, tourist drivers to frequent sex workers. destinations are often locations where more casual sex takes place and sex workers congregate (Currie, 1997). Unfortunately, less research has focused on Construction Workers tourism workers speci�cally.26 A study among tourism workers in Egypt found fairly high HIV knowledge Although not as likely to travel long distances, but low belief in using condoms to prevent HIV construction workers also spend signi�cant time away transmission (El-Sayyed et al., 2008). In Montenegro, from their homes and families. Large construction 14% of HIV cases were found to be among tourist projects can create employment opportunities for workers (Godinho et al., 2005). 26 There is some literature on sex tourism that focuses more on sex workers and the risks facing foreign and domestic tourists, rather than the tourism workers. Notes: