Cross-border Trafficking In Human Beings: Prevention and Intervention Strategies for Reducing Sexual Exploitation

5 Background 5 Objectives 5 Selection Criteria 5 Data collection 6 Results and Conclusion 6 Summary 6

Abstract BACKGROUND Over the years, growing attention has been given to the phenomenon of trafficking in human beings (THB). Sexual exploitation was until recently by far the most commonly identified feature of THB, followed by forced labour. Many activities to combat trafficking for the purpose of sexual exploitation have been initiated by numerous supranational, international as well as national organizations. Much is written about these initiatives, but some areas have been neglected. Knowledge on 'what works' is in particular limited. The growing attention to THB entails a demand for more information. The severity of the crime and the impact on its victims makes it of utmost importance to gain more insight into the working and effectiveness of anti-trafficking strategies and interventions.

OBJECTIVES
The main objective of this review was to assess the presently available evidence on the effects of interventions that aim to prevent and suppress trafficking in human beings.
The following questions were central to the systematic review: • What types of anti-THB strategies and interventions can be identified that have been accompanied by some form of empirical analysis? • Which of these studies incorporate (quasi-)experimental evaluations that are rigorous enough to determine the effect of these anti-THB strategies on preventing and suppressing THB? • What are the outcomes of these (quasi-)experimental studies?

SELECTION CRITERIA
In this review only studies that focused on cross-border trafficking for the purpose of sexual exploitation were included. The focus is on cross-border trafficking, because internal trafficking is not (yet) or just recently recognized in many countries. Furthermore, this review was limited to trafficking for prostitution or sexual exploitation, in any form there is. Trafficking for forced labour or services, slavery or practices similar to slavery, servitude or the removal of organs were kept out of consideration. In addition, included were only studies on anti-trafficking measures that involved evaluations of strategies -policies and interventions -to prevent or suppress cross border trafficking for the purpose of sexual exploitation with at least a level 3 research method of the Maryland Scientific Methods Scale (SMS).

DATA COLLECTION
An extensive search strategy was used to identify studies for inclusion, consisting of a search in electronic databases, a search in the library of the Bureau of the (Dutch) National Rapporteur on Trafficking in Human Beings (BNRM), and an Internet search. Furthermore, we contacted relevant members of the professional network. Among the over 19.000 titles we came across, 144 studies were found, of which around 20 studies seemed possibly eligible and were coded by the authors.

RESULTS AND CONCLUSION
No studies were found that met all criteria (prevention and suppression strategies, cross-border trafficking, sexual exploitation and a design of at least level 3 of the Maryland Scientific Methods Scale (SMS), i.e. a controlled design with both pretest and posttest measures and comparable control conditions.). Consequently, no conclusions could be drawn on the effectiveness of anti-THB intervention strategies for preventing and reducing sexual exploitation. Some studies had employed other evaluation designs. We decided to include a narrative review presenting these studies since they sketch the landscape of anti-THB interventions, and are informative of the current state of evaluations of these interventions. However, since these are non-controlled studies, they do not provide a basis for drawing conclusions about actual outcomes and impacts.

Introduction
Over the years, growing attention has been given to the phenomenon of trafficking in human beings (THB). Human trafficking generally involves the use of force, fraud, or coercion to enslave people in situations that are exploitative and in many cases illegal and dangerous (GAO, 2007). The use of fraud, force, or coercion typically distinguishes THB from smuggling people. Victims of trafficking may be forced to work in prostitution, strip clubs, sweatshops, agricultural businesses, armed conflict, service industries and people's homes (Boonpala & Kane, 2002;GAO, 2007;Gervais, 2005;Kelly, 2005). Underage victims are forced to beg for money in cities, work in carpet shops, and participate in pornography and sexual acts with adults (BNRM, 2009;Friesendorf, 2007). Furthermore, trafficking in children in the sports business, especially in soccer, is not uncommon (de Boer, 2002;David, 2005). In summary: trafficking occurs in many different sectors and industries.
Not all governments recognize trafficking as a problem. They may treat foreign trafficking victims as illegal immigrants and deport them back to their home countries, rather than protect them against their perpetrators or against possible future trafficking (GAO, 2007). However, in most countries trafficking has been recognized as a national and international problem, and there is a widespread agreement for the need of a multilateral response on both levels (Laczko, 2005). The United Nations (UN) Protocol to Prevent, Suppress and Punish Trafficking in Persons, especially Women and Children, supplementing the UN convention against Transnational Organized Crime (2000) stimulated anti-THB legislation and interventions in many countries all over the world. Furthermore, there have been a growing number of (inter)national conventions on human trafficking in the last decade as well as a vast increase in the number of publications on THB and on ways to prevent and suppress it (Kelly, 2005;Laczko, 2005). The anti-trafficking protocol is used in combating trafficking through judicial and law enforcement means. It has been signed by 117 countries 1

RESEARCH ON CROSS-BORDER TRAFFICKING
, and the definition in the protocol is used by governmental and non-governmental organizations (De Sas Kropiwnicki, 2007), making it the most widespread trafficking definition (e.g. Gallagher, 2006).
Until recently, sexual exploitation was by far the most commonly identified feature of THB, accounting for 79% of all cases (Chawla, Me & Pichon, 2009). According to a working paper of the International Labour Organization (ILO) 75% of the reported cases were identified as sexual exploitation (Belser, 2005). 2 Traditionally labour exploitation is not a priority for law enforcement activities and often remains undetected (Europol, 2009). Legal definitions may fail to include labour trafficking (Smit, 2011), and victims of forced labour may not come forward because they do not consider themselves victims of trafficking (Andrees & Van der Linden, 2005). Still, even though sexual exploitation is the most documented type of THB, this does not imply an overall consistency in recognizing sexual exploitation. For example, governments disagree on whether there is a difference between 'forced ' and 'voluntary' prostitution (GAO, 2007). 3 Furthermore, organizations continue to face difficulties as they work on global and regional initiatives to combat trafficking (Barnitz, 2001). It proves to be difficult to develop initiatives for under-reported forms of exploitation such as forced or bonded labour, domestic servitude and forced marriage, trafficking for the purpose of organ removal and the exploitation of children in begging.
Much research has focused on the prevalence of THB, but no consensus has been reached on its incidence (Friesendorf, 2007;Goodey, 2008). In 2006 the United States Government estimated 600,000 to 800,000 persons being trafficked across international borders annually. However, according to UN Children's Fund (UNICEF) and ILO the number is much higher with approximately 2 million people being trafficked every year worldwide (Boonpala & Kane, 2002). These estimates are questionable due to methodological weaknesses, gaps in data, the interpretations made and the numerous discrepancies found in studies (Cwikel & Hoban, 2005;GAO, 2006;World Congress against CSEC, 2001). With regard to registered victims, the numbers vary greatly as well. Italy, for example, accounts for a high number of reported victims compared with other countries in Europe. In 2006, Italy reported 2,143 victims of trafficking. This relatively high number 4 Research on THB has also focused on trends and developments, and on facts about traffickers in order to understand the phenomenon. This understanding of the depth, the breadth and the scope of the problem is needed before developing any initiatives towards prevention and suppression (Chawla et al., 2009;Surtees, 2008;Weiner & Hala, 2008). Research on conviction has also been done. Statistics have shown an increase in the number of convictions, but most convictions still take place in only a few countries. According to Chawla et al. (2009), who studied THB in Europe, Asia, Africa and America, two out of every five countries (41%) do not report any convictions of THB and 14% of all countries have no data available at all. Possible explanations for the lack of data and/or (reporting of) THB convictions are that some countries pay little or no attention towards the phenomenon or are poorly prepared and equipped for dealing with THB (Chawla et al., 2009). More attention given to THB seems to lead to higher statistics. As more is known and/or reported about THB, more awareness to the phenomenon is given by the criminal justice system. Consequently, more cases of THB are investigated which in turn leads to an increase in the number of trafficking cases reported. A rise in prevalence is, therefore, not necessarily (only) the result of an increase in THB cases.
is possibly related to large numbers of illegal immigrants in general, and also to both an actually high number of victims and the extensive care for victims. This can lead to a higher motivation for victims to denounce. In all probability only a small percentage of human trafficking is registered leaving a high dark number in its prevalence (BNRM, 2007;Goodey, 2008).

LIMITATIONS OF RESEARCH ON THB
Much is written about these initiatives, but some areas have been neglected.
Knowledge on 'what works' is particularly limited (Tyldum, Tveit & Brunovskis, 2005). Most publications are primarily descriptive, for example describing the hardship of, or social assistance to victims. Others mainly aim at collecting trustworthy numbers to define the scale of the phenomenon (Kelly, 2005). Few, if any, interventions are accompanied by evaluation research and few systematic data collections exist (Larsen, Lindley & Putt, 2009). There has been relatively little independent 5 The individuals involved in human trafficking, whether they are the perpetrators or victims, are considered hidden populations of which the size and boundaries remain unidentified (Laczko, 2005;Weiner & Hala, 2008). As human trafficking is illegal and mainly underground (BNRM, 2007;Boak, Boldosser & Biu, 2003) trustworthy data is difficult to collect, resulting in conclusions that may be far from the truth. Deficient data tracking systems and methodologies for estimation and evaluation could increase the risk that programmes targeting this specific phenomenon are not accurate (Clawson, Williamson & Garrett, 2008;Kelly, 2002). Evaluation is important as evaluative knowledge on trafficking can be used to develop adequate prevention techniques or policies (Dottridge, 2007a;Kelly, 2005;Larsen, Lindley & Putt, 2009). The United States Government Accountability Office reviewed documents of 23 U.S. Government funded anti-trafficking projects in Indonesia, Thailand and Mexico. It revealed that 21 of the 23 projects included one or more monitoring elements but only 10 state how performance was measured. The majority lacked a logic framework of monitoring that linked activities to goals, indicators and targets (GAO, 2007). And even though the number of monitoring and evaluation studies seems to be rising, more scientific evaluation research is needed to ensure the effectiveness of the interventions developed, to help agencies achieve the results desired, and to enable more effective oversight and accountability (GAO, 2006).
research that evaluates and assesses the effectiveness of counter trafficking policies, programs, and various interventions (Chase & Statham, 2005;Laczko, 2005). This can be problematic since organizations evaluating their own initiative may have explicit or implicit political or practical agendas that could influence their conclusions, for instance by selecting or disposing of certain information (Vandekerckhove, 2003). It is possible that more or less weight is given to certain data due to the authority of certain key actors .
Another frequently encountered restriction in the research on THB is the availability of only small samples. Often, this is due to the limited access to victims and perpetrators (Ennew, 2008;Laczko, 2005). Intrinsically, there is nothing wrong with small samples. If treated sensibly a variety of methods can produce illuminating results. But often preference is given to a single method such as interviews only, or questionnaires only. Furthermore, small amounts of data are sometimes pushed through numerical hoops that produce misleading results (Ennew, 2008). As a result, little (reliable) knowledge on effectiveness of interventions in this area is available (BNRM, 2007).
While knowledge on the effectiveness of anti-trafficking interventions is scarce, interventions can cause collateral damage (Boermans, 2009;Dottridge, 2007b;Limanowska, 2003). Anti-trafficking programs can cause damage by putting restrictions (sometimes unlawful) on young women's decisions to migrate, to travel or even leave the parental house (Friesendorf, 2007). For instance, refusal to issue a visa or to allow a young woman to enter the country is presented as an antitrafficking measure (Limanowska, 2003). In 2002 the UN High Commissioner for Human Rights issued a series of recommended Principles and Guidelines on Human Rights and Human Trafficking. The third principle states that anti-trafficking measures "shall not adversely affect the human rights and dignity of persons, in particular the rights of those who have been trafficked…" (p. 3) thereby aiming to protect victims of trafficking against (more) damage by interventions. Guideline 3 provides more detail on this point and urges governments, intergovernmental organizations and NGOs to monitor and evaluate "the relationship between the intention of anti-trafficking laws, policies and interventions, and their real impact" and to distinguish "between measures which actually reduce trafficking and measures which may have the effect of transferring the problem from one place or group to another" (p. 7). According to Dottridge (2007c), the High Commissioner's Guidelines do not appear to have been heeded by many government agencies and NGO's involved in anti-trafficking activities. Many years have passed since the endorsement of the UN Trafficking Protocol but little is known about which interventions have been the most effective in preventing human trafficking, protecting victims and prosecuting traffickers (Friesendorf, 2007;GAO, 2007).

Objectives of the Review
The background information yields a number of justifications for a Campbell Systematic Review on this topic. The growing attention to THB entails a demand for more information (Goodey, 2008). The number of interventions to counter THB is rising, and many initiatives have been described. At the same time the effectiveness of these programs remains unknown. 6 However, the severity of the crime and the impact on its victims makes it of utmost importance to gain more insight into the working and effectiveness of anti-trafficking strategies 7 Consequentially, this review aims at contributing to a more evidence-based approach in the prevention and suppression of cross border THB for the purpose of sexual exploitation. and interventions. Hence, the objectives of this systematic review are twofold. Firstly, this review will bring together and assess the available evidence on the effects of anti-THB strategies and interventions. This will contribute to the knowledge on anti-THB interventions and their effectiveness and will therefore provide information to policy makers. Secondly, this review will increase the knowledge about the strengths and weaknesses of research concerning anti-trafficking strategies and interventions that can stimulate future research and improve evaluations. Awareness will be raised on the need for more rigorous research designs that also shed light on possible collateral damage of interventions that otherwise may remain unknown (Dottridge, 2007c).
The questions to be answered are the following: a) What types of anti-THB strategies and interventions that have been accompanied by some form of empirical analysis can be distinguished? b) Which of these studies incorporate (quasi-)experimental evaluations that are rigorous enough to determine the effect of these anti-THB strategies on preventing and suppressing THB? c) What are the outcomes of these (quasi-)experimental studies?

Methods
To meet the objectives of this systematic review, the following methodology was proposed. This section is a summary of the methodology detailed in the protocol which was approved by the Campbell Collaboration in 2010. 8

Types of interventions
As discussed in the introduction, the definition of human trafficking is wideranging. In this review only studies focusing on cross-border trafficking for the purpose of sexual exploitation were collected. The focus is on cross-border trafficking, because internal trafficking is not (yet) or just recently recognized in many countries. Furthermore, this review was limited to trafficking for prostitution or sexual exploitation, in any form there is. Trafficking for forced labour or services, slavery or practices similar to slavery, servitude or the removal of organs were kept out of consideration. Until recently, there was little awareness of these types of trafficking, and for quite some time no legislation against these types of trafficking existed. Today, many countries have included into their laws trafficking for labour exploitation and organ removal. But often, labour exploitation has not been strictly defined, leaving the burden of interpreting that part of the law to the judge (ILO, 2005). This resulted in different definitions and many discussions about when bad labour conditions turn into trafficking (Smit, 2011), not making it a very transparent field to study. Furthermore, trafficking for sexual exploitation, trafficking for labour exploitation and especially trafficking for organ removal represent crimes in different 'domaines', which are not necessarily committed, nor investigated, prosecuted, or studied by the same people or organizations, resulting in a field considered too broad to include entirely in the present review.
In addition, studies on anti-trafficking measures were included that involved evaluations of strategies -policies and interventions -to prevent or suppress cross border trafficking for the purpose of sexual exploitation. This meant that evaluation studies focusing exclusively on sexual exploitation were incorporated in the analysis.
Furthermore, studies focusing on trafficking in general, but where data on trafficking for sexual exploitation could be extracted were included. Studies evaluating general interventions were excluded from the systematic review as it was impossible to isolate the effectiveness of interventions focusing on trafficking for the purpose of sexual exploitation due to the absence of transparent data.

Types of studies
All quantitative evaluation studies with respect to anti-trafficking interventions and strategies that focused on sexual exploitation were included. They were categorized according to the Maryland Scientific Methods Scale (SMS) (Sherman et al., 1998) in order to identify and assess the strength of the evidence (Sherman, Farrington, Welsh & MacKenzie, 2002). Accordingly, only studies with a design of level 3 or higher were included in the review, meaning that both pretest and posttest measures should be administered and that control conditions are comparable to experimental conditions. 9

Participants
Studies considering best-practices or evidence-based practices on how to counteract THB were included. Descriptions of good practices based on field experiences but lacking any numerical data were excluded.
Evaluation research dealing with initiatives targeting individual perpetrators and criminal groups as well as vulnerable individuals and professionals in the field of anti-THB were included. Policies or initiatives exclusively regarding victim assistance, the reintegration of victims and arrangements for legal status, were not included.

Publication
Eligible studies were either published -electronically or in print -or unpublished. Studies from the year 2000 and onwards were incorporated in the review. This year was chosen since it marks the UN Trafficking Protocol. Our search included publications up to the end of June 2009. Publications after this date were not included in this review, but will be examined in the update.

Country of origin
Studies from any country all over the world were included. However, due to limited resources and knowledge of a few foreign languages within the research team, only those studies that were written in Danish, Dutch, English, French, German, Italian, Norwegian, Spanish and Swedish were reviewed. Eastern European languages were not included; however, inquiry at La Strada International yielded no additional studies in these languages. 10 Of all the possibly eligible studies (144), only 18 were not in English. Of these, 13 studies were written in Dutch, 1 in German, 1 in French, 1 in Danish and 1 in Spanish. One study was in Finnish but had an English summary.

SEARCH STRATEGY FOR IDENTIFICATION OF RELEVANT STUDIES
This review employed a broad search to identify all studies -published or otherwisethat meet the criteria described above. This included a search through electronic databases, searching the BNRM library, contacting professional networks, searching through the reference list of all documents and an Internet search. Special attention was given to searching and collecting relevant studies captured in the so-called grey literature. Table 1 provides a summary of the search results. To find relevant studies, keywords and wild card terms were used as showing in Figure 1. These keywords have been constructed in a preliminary review of the literature. Only studies with these keywords in the title, subtitle and/or abstract were eligible for this review. A list of key words in the different languages used can be found in appendix I. Keywords that were used involved words that described that: -the subject of study (e.g. human trafficking, sexual exploitation, sex*, anti-trafficking) AND -the interventions and programs of interest (intervention, prevent*, program) AND -research effectiveness (effect*, evaluat*, best practice) From the 19,398 citations found on sex trafficking, we found 144 studies that seemed eligible for this review. These studies contained the keywords in the title and/or abstract. A detailed scan of these 144 studies resulted in 20 studies that were relevant for this review (for example, they contained empirical research). The authors read these 20 studies with scrutiny and coded them. Of these 20 studies only four studies seemed to have some form of evaluation (level 1 or 2 of the SMS) of an intervention that prevents trafficking with sexual exploitation and were therefore included in the narrative review (see 3.6 for more information on the relevance of a narrative review in this systematic review). None of these studies contained an evaluation with a level 3 or higher of the SMS. Studies included with level 3 SMS or higher 0

Electronic Databases
Computerized bibliography searches were conducted using different electronic databases. The databases were selected due to a number of reasons. Some we knew in advance to contain relevant literature, others came forward during the internet searches, or because they were mentioned in relevant literature or by people from our professional network. A detailed list of the searched databases can be found in appendix II. In total, 66 electronic databases were searched. Of these databases, 973 citations were reviewed. However, none of the citations were found to have a design according to level 3 or higher of the SMS.

BNRM library
The Bureau of the (Dutch) National Reporter on Trafficking in Human Beings (BNRM) has a library with over 3,000 articles, books and reports related to human trafficking. Using their search engine (Reference Manager), all articles were scanned for applicability by using the keywords. In total, 3080 citations were reviewed and of these citations 10 documents seemed to fit the initial criteria. After careful examination, these documents were removed from the final inclusion as they did not have a design according to level 3 or higher of the SMS.

Professional Networks (existing contacts)
Staff of relevant organizations was contacted to provide the authors with any published or unpublished study. Initially, the network of BNRM was contacted. Furthermore, other personal networks of the authors were consulted such as policymakers, researchers and contacts at international and supranational organizations. We sent 954 letters, by mail and e-mail. Of these, 91 individuals emailed or wrote back with information or suggestions. This search strategy yielded much information; however, many citations did not deal with evaluations of initiatives. Instead, many National Action Plans, program descriptions and general information were received.
Furthermore, Terre des Hommes, a non-governmental organization located in the Netherlands, invited the researchers to examine their effect studies on anti-THB interventions. These studies were not published and therefore fell under the grey literature. In total 15 studies were reviewed of which 6 seemed to meet our initial criteria. A closer examination of these studies, however, resulted in an exclusion for the systematic review due to their research design and/or for focusing on THB in general and not specifically on sexual exploitation.

Cross-referencing of bibliographies
During the process, references in relevant review articles, primary studies and metaanalyses were scanned for new leads. Only articles that contained the words listed in appendix II in the title or subtitle were reviewed. To locate any grey literature, government reports as well as nongovernmental organization's reports, book chapters, policy documents, personal networks, and research organizations' websites were consulted. Another 39 unduplicated references seemed useful and were examined. Furthermore, many relevant websites were found in bibliographies, which were all visited. None of the references or websites resulted in an inclusion of studies. Two studies found by cross-referencing were included in the narrative review.

Internet searches
Websites of NGOs (e.g. Terre des Hommes, UNICEF and CATW), professional associations and organizations involved in applied and policy research on human trafficking were searched. A list of related NGOs (and their websites) available at BNRM was used and supplemented. In addition, keyword searches were conducted using search engines such as google.com. In total 110 websites were visited. These websites are listed in appendix III. Furthermore, the authors and organizations of studies that seemed eligible, but lacked (empirical) data, were contacted by e-mail to request additional data. The Internet search merely led to 2 studies that were included in the narrative review.

DESCRIPTION OF METHODS USED IN PRIMARY RESEARCH
The authors searched for evaluation studies individually by applying the inclusion criteria. The abstracts of these articles were assessed on their relevance for the systematic review. Initially, all studies that made any reference to an evaluation of a program were included, regardless of the research design or quality. An overview was made of all 144 studies containing any (combination) of the relevant key words in their title, subtitle and/or abstract. When a final overview of all studies was compiled, the second stage started by examining full-texts using the coding sheet. 11

CRITERIA FOR DETERMINATION OF INDEPENDENT FINDINGS
The third and final stage consisted of selecting studies that evaluated interventions using an evaluation design of level 3 or higher of the SMS.
This systematic review examined outcomes related to the effectiveness of anti-THB interventions. It was expected that studies might include different types of strategies or different study samples. In a few studies several independent samples were presented, these samples and their outcomes were treated as separate findings.

DETAILS OF CODING CATEGORIES
A preliminary coding sheet was developed for this review. This sheet covered a series of key aspects of the study, such as research design, methodological quality and outcome measures. Furthermore, characteristics of the intervention were coded such as the type of intervention, the population it is aimed at and the procedures used. All studies from the initial screening were coded independently by at least two authors. Coding differences did not appear.

QUALITATIVE DATA
A narrative review was done on studies that fit the criteria except for the criterion of studies that involved a comparison or reasonable counterfactual condition (SMS levels 3-5). Narrative reviews are used to summarize results from qualitative and quantitative but non-controlled studies (Lipsey & Wilson, 2001). 12 A narrative review can be useful to summarize findings which are excluded from meta-analysis and is conducted to highlight the current landscape of evidence on evaluations of anti-trafficking interventions. This narrative review gives insight in the critical research areas and the gaps in the evaluation studies on cross-border trafficking for the purpose of sexual exploitation.

Results
The search resulted in 144 potentially eligible studies. Many of these studies (38), concerned interventions of a primarily preventive nature, 23 concerned suppressive interventions (mostly regarding legislation and prosecution), and 8 were of a combined preventive/suppressive nature. In 48 studies no specific initiative was described. Most studies were in the field of sexual exploitation (68), or concerned several types of exploitation (55). The target group varied. It most often consisted of both adults and children (57), focusing on women and children, but studies on programs that were especially developed for minors (24) were found as well.
After a more extensive review no more than 20 studies seemed eligible. These studies were entirely read and coded independently by two authors, using the coding sheet. Findings were discussed with all four authors. No studies were found to be applicable for this review. There were no disagreements between authors regarding study inclusion or exclusion. Two studies appeared potentially relevant but a closer look at the website of the organization indicated that they needed to be excluded from analysis as they did not address cross-border trafficking but internal trafficking. 13 All 20 studies regarded anti-trafficking interventions that were evaluated. In some cases it was unclear whether the focus was on sexual exploitation or on trafficking in general; 11 interventions focused solely on sexual exploitation, 4 on both sexual as labour exploitation and 5 concentrated on trafficking in general. Some interventions did not specify whether they focused on internal or cross-border trafficking. Separate results for cross-border trafficking were not available.
None qualified as a (quasi-)experimental study and therefore all 20 studies were excluded from analysis. Although these studies are excluded, they are listed in Appendix IV to give an impression of existing evaluations in this area. Additionally, some information about the nature of these interventions is attached (appendix V).

Narrative Review
Four studies fit the criteria for a narrative review. Included in this narrative review are studies that regard: Some studies evaluated more than one intervention. In those cases only interventions that fit the selection criteria were included. The following results were found.

TYPE OF INTERVENTIONS
The first study (Association for Emancipation, Solidarity and Equality of Women, 2006) is a qualitative evaluation of the economic and social stabilization program for potential victims of trafficking in the border regions of Former Yugoslavic Republic Macedonia. The main goal of this program was empowering these potential victims and providing employment opportunities through skills training. Basic business training courses were organized aiming to develop knowledge on how to start and manage a small business, and how to generate income. The intervention focused on vulnerable groups, especially women.
The second study (Balanon & Barrameda, 2007) evaluated the impact of the Young Men's Camp in the Philippines, which was developed to change the sexual attitudes and practices of boys and young men. It is believed that these attitudes and practices contribute to trafficking for sexual exploitation by encouraging (illicit) prostitution. This preventive program aimed at transforming socially constructed notions held by men about gender issues.
The third study (Centre for Research on Environment Health and Population Activities, 2004) included an evaluation of an awareness raising program in rural Nepal. The primary target group of the program was adolescent girls, and their peers, parents and the community were the secondary focus group. This intervention educated the girls and aimed to prevent them from becoming victims of trafficking.
The fourth study  evaluated an Israelian prevention program aiming at raising awareness among the wider public and creating attitudinal and institutional change among decision makers and law enforcement personnel. Education on trafficking was achieved by distributing information, providing training, publicizing reports, using media advocacy and giving lectures. The program intended to influence public policy and opinion that would lead to a decrease in trafficking.

METHODOLOGY AND TYPE OF EVALUATION
The first study assessed the impact of the project on the beneficiaries, their families and local stakeholders. The main dimensions were the socio-economic position, quality of life as well as potential vulnerability towards human trafficking. The evaluators used questionnaires and interviews. For the target groups, different interview discussion guides were designed. The study consisted of a post-test measure only (SMS level 1). There was no control or comparison group and no empirical outcomes were reported.
The second study conducted pre-and posttests (SMS level 2) to assess changes in knowledge and perceptions among the boys and young men. Direct outputs, such as changes in knowledge and attitudes, were measured as well as the immediate impact of the camp on the graduates. Next to these assessments there were follow-up activities, consisting of one day fora and workshops. It remained unknown whether the intervention focused only on cross-border trafficking.
In the third study the effects were measured using interviews and surveys. It contained a pre-and posttest measure without a comparison or control group (SMS level 2). Program participants were randomly picked from the target population. Outcome measures were the increase of awareness and knowledge on trafficking.
The fourth study ) evaluated a hotline for migrant workers in Israel. The researchers conducted a pre-and posttest by carrying out interviews and questionnaires. However, there was no control or comparison group (SMS level 2).
None of the four studies met level 3 or higher of the SMS. Three studies were level 2 studies including pre-and post-test measurements, one study was only level 1. All four studies failed in generating control or comparison groups. Furthermore, data available were only of a qualitative nature, empirical data was lacking in the reports and could not be obtained from the researchers.

FINDINGS
The first three studies reported positive outcomes, referring to "immediate effects [such] as changes in awareness and perceptions" (AESEW, 2006). The fourth study addressed that there were no differences between pre-and posttest regarding awareness . No negative or side effects were reported. In none of the four studies long-term effects were assessed. Only measurements during the intervention and shortly after were reported. The Young Men's Camp had follow-up activities, however these activities consisted of workshops, and not a follow-up evaluation. The authors recommended an improved monitoring and evaluation system indicating (some) awareness about the weaknesses in their research.
It should be noted that even though some outcomes were reported, the methodological quality of the evaluation studies do not allow for firm conclusions regarding the effectiveness of these interventions.

Conclusion
The current landscape of anti-trafficking initiatives shows that many activities to combat trafficking have been initiated by numerous supranational 14 Literature suggests that most interventions focus on prevention through awareness raising activities. Some of the interventions aim directly at the at-risk population, victims or perpetrators, others focus more indirectly on populations, trying to create more awareness among social workers and policy makers. Furthermore, initiatives can combine several goals and activities, such as focusing on legislation, policy development and awareness-raising. Different methods are used in awarenessraising initiatives. In most cases, awareness-raising initiatives provide skills training to the targeted population, such as victims and the at-risk population. A popular instrument is the media (television, radio and posters), mainly used to raise awareness among the public.
, international as well as national organizations. In general, these organizations assess the prevalence and seriousness of the phenomenon and develop initiatives based on their findings. These activities can be distinguished in preventive, suppressive and victim support interventions. The specific goals of these interventions vary from raising awareness, providing education and employment to changing the legislation, prosecution of perpetrators and empowerment of victims.
The aim of this systematic review was to explore what is known about the effectiveness of interventions that prevent or suppress cross-border trafficking for the purpose of sexual exploitation, in order to contribute to a more evidence-based approach in the prevention and suppression of cross-border trafficking for the purpose of sexual exploitation. The main conclusion is that there is no evaluation research on anti-trafficking initiatives that meets the methodological criterion of level 3 or higher of the SMS. Consequently, no substantive conclusion about the effectiveness can be made, resulting in neither support nor rejection of the present anti-trafficking initiatives.

Discussion
Even though reliable conclusions about effectiveness are lacking, some observations are worth mentioning. 15 First, the number of evaluation studies seems to increase. Researchers more often include some form of evaluation or monitoring of a program in their report. 16 Second, outcomes are not always presented comprehensively. Weaknesses in the research design are not acknowledged and empirical evidence is lacking, making it difficult to fully understand or to replicate the research.
This may indicate a growing interest in an evidence-based approach and a willingness to evaluate anti-THB initiatives. However, researchers face many methodological challenges and ethical concerns (Brennan, 2005), such as attaining the vulnerable population, and the secrecy of the crime (Gozdziak & Bump, 2008). Although some serious attempts have been made to evaluate anti-trafficking initiatives, improvements in the research design are needed in order to draw substantive conclusions on the effectiveness of these initiatives. Improving research designs is perhaps not that difficult. For example, the study of the Young Men's Camp, which we described briefly in the narrative review used a pre-and post test measure, but a control group was lacking. Adding a control group would strengthen the design. Young men who do not participate or who are on the waiting list can be treated as the control group.
Third, a certain inconsistency seems to arise. Whereas existing programs have not been evaluated properly, new anti-THB interventions arise frequently. This is not only costly, but initiatives can also cause collateral damage (Boermans, 2009;Dottridge, 2007b;Limanowska, 2003). It is therefore crucial to know more about the effectiveness of anti-THB initiatives to understand where money needs to go and to understand which support the (potential) victims of trafficking need.
Fourth, goals of evaluation and monitoring activities are diverse. For example, evaluating the implementation of an intervention, the satisfaction of received services, increased knowledge about the topic and increased collaboration between organizations. Program evaluation operates under the idea that social programs should have explicit goals by which success or failure can be measured (Berk & Rossi, 1999), but several studies do not focus solely on trafficking, but on more general social-economic problems or child protection as well (e.g. Bühler, Barron, Thy & Sovanny, 2004). Some evaluation researchers use questionnaires, others conduct interviews or observations and focus group discussions. In other words, the methods varied. The variety in objectives and methods makes comparison difficult.
Fifth, numerous reports did not make a distinction between prostitution and trafficking for the purpose of sexual exploitation. Instead, these terms were used as synonyms. Neither was a clear distinction made between trafficking for forced prostitution and voluntary migration for sex work (see also Gozdziak & Bump, 2008). This makes reviewing these studies difficult.
Sixth, as many interventions concern awareness raising, it should be noted that: "by definition, [awareness raising] rests on a presumption of possible future entrapment and can therefore never be fully empirically tested. The programs therefore depend on something of a leap of faith, supported by knowledge of the general vulnerability of children/families/communities in a given town or region, and a belief in certain strategies to influence their behavior and lifestyle. This undoubtedly has some validity, even if it cannot be empirically measured over the short life span of a project, but disconcertingly, the realities and modalities of the clients and exploiters are rarely factored into the equation. Given so many intangibles, it is extremely important that prevention programs have built into them medium-term monitoring and/or follow-up of the children and families." (ILO-IPEC, 2001 p.22).
A possible problem with information campaigns can be the broad target group. A mixed population asks for a variety of methods, or at least various materials (posters, TV commercials) that are attractive and can reach the subgroups. It is not always clear whether the instrument fits the target group.
Finally, this review only dealt with trafficking for sexual exploitation. Other forms of trafficking, may be relevant to study as well. Relevant evaluation studies on internal trafficking seem to exist. And in our search for eligible studies for this review also studies on labour exploitation were found. Therefore, it seems worthwhile to explore the possibility of systematic reviews on internal trafficking and on labour exploitation.

Implications
Methodological weaknesses of available studies preclude any conclusions concerning program effectiveness, ineffectiveness or collateral damage. The absence of appropriate studies makes it difficult to advice about policy implications. The designs were not rigorous enough and empirical data was lacking. More detailed information on working mechanisms was missing as well. It goes without saying that there is a need for rigorous evaluation research. Researchers, funding institutions, and policy makers, they should all demand higher standards of evaluation studies in order to gain knowledge on outcomes. This would enable them to better succeed in preventing and suppressing trafficking in human beings for the purpose of sexual exploitation. To bring about more rigorous research designs, partnerships can be sought in the area of politics and funding agencies. Funding agencies and other external parties are able to influence researchers to make use of a more rigorous research design (Lum and Yang, 2005). It should be standard practice that evaluations are not conducted by the agency funding or offering the program.
Also, more attention should be paid to the methodology. Information on methodology is crucial in order to be able to rate a study and the results at its true value, this is crucial information. In addition, to interpret the results, a transparent presentation of the qualitative as well as quantitative data is required. This includes adding examples of questionnaires and interview questions used. Such information is also a necessary requirement for conducting a meta-analysis.
Doing evaluation research and publicizing the conclusions allows others to see what is effective and what not. This may lead to the development of new and more effective anti-THB initiatives. Such an evidence-based approach is needed to reduce and eliminate human trafficking. As Dottridge (2007a) puts it: "In the case of every substantial initiative to prevent trafficking, therefore, it is important that it should be evaluated. Whenever possible, the lessons from the evaluation should be published or made available to others conducting similar work. The fastest way of circulating information of this sort at the moment is to put it on a website, ensuring the title mentions 'evaluation' or 'impact assessment' (of efforts to prevent child trafficking), so that this is picked up by Internet search engines" (Dottridge, 2007a, p.66).

Acknowledgements
We would like to thank the Campbell Collaboration (C2) in Norway and the Dutch Ministry of Justice for their financial support. We are also grateful for the support given by the BNRM by allowing us to search through the library. Furthermore, we would like to thank Peter van der Voort from the Netherlands Institute for the Study of Crime and Law Enforcement (NSCR) for his expertise in searching relevant studies. And finally, we would like to thank all the people and organizations for responding to our request by sending their information on trafficking. Finally, we are grateful for the comments of our anonymous reviewers and the Campbell reviewers for their comments on earlier versions of this review.

Plans for updating the Review
The next update of this review will be in five years. Introduction: This project compares the achievement of prevention and intervention programs in the Netherlands and the United Kingdom. The programs focus on social care for commercial sexual exploited children.
Methodology: Pre and posttest study with comparison group, in which a specially designed monitoring tool and case study model are used to collect qualitative and quantitative data of at risk youth.
Outcome measures: Monitoring tool; lifestyle risks and level of sexual exploitation.
Results: In accordance to specific outcomes due to contact with social workers, small differences were discovered. However, the general risk for sexual exploitation was decreased by the intervention, as well in the Netherlands as in the UK. Methodology: Data was collected from professionals (pre and posttest questionnaires, focus group discussions and interviews), young people in schools (questionnaires, focus group discussions and interviews) and young people in Pupil Referral Units (interviews). Pre and posttests were accomplished, as well as followup measures.

Reference
Outcome measures: There were items on knowledge of sexual exploitation, awareness of sexual exploitation, understanding of rights and on the assessment of the training.

Results:
In general, the program has positive effects: there is an increase in awareness and knowledge for the professionals and young people in PRUs. For the young people in schools the growth in knowledge and awareness remains unknown, as they only conducted a posttest questionnaire.

ADDITIONAL INFORMATION ON 20 STUDIES THAT WERE EXCLUDED AFTER CODING
All 20 studies that were coded are reports, of which 16 come from an NGO and 4 are the result of cooperation between an NGO and a government. It were also mainly NGOs that were responsible for operating and maintaining the interventions (11), or NGOs in collaboration with a government (7). Half of the interventions (10) were evaluated by an independent evaluator. In one case the evaluator had delivered the intervention and had a direct influence on operations. Furthermore, all reports were written in English, although country of origin varied from Southeast European to Asian and Southeast Asian. Besides, not only the typical countries that are mostly cited in the literature were represented. Studies also regarded interventions in Israel, Austria, Kazachstan and Costa Rica.
Explicit goals of the interventions consisted mainly of 'awareness raising', 'capacity building' and 'employment/empowerment'. Of the interventions that had 'awareness raising' as goal (16), 6 used a (information) campaign as primary method to reach their target group. Other interventions made use of trainings, shelter home visits, media advocacy and lectures. As most interventions attempted to raise awareness, their target group mainly consisted of a mixed group, including juveniles and adults and as well males and females. Some interventions focused in particular on juveniles (3), for example by life skills training, others exclusively on adults (3). Some interventions concentrated only on males or females. Introduction: The purpose of the programme Prevention and Capacity Building Activities is to establish a functional government and NGO infrastructure capable of raising awareness, and to facilitate and build regional NGO network and also to effectively train the law enforcement personnel.
Methodology: Posttest, consisting of desk work (report analysis) and interviews.
Outcome measures: Level of awareness.

Reasons for exclusion:
The evaluation did not focus on the effectiveness of the intervention, but on the expenditure.

Introduction:
The aim of the 'Children in need of a special protection program' is to enable social and legal systems to undertake effective preventive and rehabilitive measures for children at risk and in distress, including combating child abuse, trafficking and sexual exploitation.
Methodology: Posttest, using interviews, a desk review, field visits, workshops and reports.

Outcome measures:
The role and relevance of the intervention, its design & focus, its effectiveness and efficiency, sustainability, replicability and mainstreaming of results.

Reasons for exclusion:
The intervention did not focus on sexual exploitation, but on trafficking in general. Introduction: This reports looks at the good practices of 4 programs, the only program with a description of evaluation: The Young Men's Camp. Five key components: 1)development of teaching materials, 2) the actual conduct of the Camp, 3) Training of Trainers, 4) the Camp assessment, 5) support to follow-up activities.
Methodology: Pre-and posttest, by reviewing documents and using questionnaires, interviews and text messaging.

Introduction:
The 100-day campaign is the largest multi-regional, multi media public education anti trafficking campaign in Russia. It aims to raise awareness on THB.

Outcome measures:
The number of materials used and number of phone calls.

Reasons for exclusion:
There was no evaluation, only monitoring activities.

Introduction:
The Program activities aimed to strengthen the human resources development capabilities of social development personnel in the participating countries to assist young victims and potential victims of sexual abuse and sexual exploitation ultimately to be reintegrated in communities and families, have access to food and health services.
Outcome measures: Program output and impact/content.

Reasons for exclusion:
The evaluation was not about effectiveness.
9. Garcia, F. E. B. (2006). Prevention, protection and rehabilitation of girl survivors of sexual abuse, prostituted girls and girls at risk. The Netherlands: Terre des Hommes.

Introduction:
The project aims to prevent, protect and rehabilitate girl survivors of sexual abuse, prostituted girls and girls at risk through the residential shelter and drop-in center, outreach and mothers and babies program. Relevant for this systematic review: girls at risk.

Methodology:
Posttest, using interviews with stakeholder, observations, review of relevant documents and an existing database.
Outcome measures: Number of girls in residential setting and drop-in centre, extent of services, assessment of program and activities, assessment of services, long-term outcomes, positive-negative contributions to the project, organization capacities.

Reasons for exclusion:
The program focused not specifically on cross-border trafficking. Methodology: Posttest, using interviews.
Outcome measures: Employment rates, computer access, training program: the interviewees felt more positive about the future due to the training program. They had more friends, more confidence at their jobs and more information.

Reasons for exclusion:
The program focused on trafficking in general, not specifically on sexual exploitation. Introduction: This report looks at THB and prostitution in Israel where an overview is given of the actual context, legal dispositions, measures taken by national, local authorities and NGO's.

Introduction:
This evaluation looks at Cambodia, Indonesia, Philippines and Thailand and the projects implemented during 2001-2005 to account for the initiatives undertaken and to provide basis for future strategic planning. This report is a summary of the reports from the different countries.
Methodology: Posttest, using structured interviews and focus group discussions.

Outcome measures: Not available
Reasons for exclusion: The conducted study was not an evaluation.
16. IOM. (2008). Final report to the Ministry of Interior of the Czech Republic: Counter-trafficking campaign targeting clients of prostitution in the Czech Republic. Czech Republic: IOM.

Introduction:
The IOM carried out an information campaign. The aim was to raise awareness about human trafficking and to present more information, such as information about non-profit organizations that offer support to victims of human trafficking. The campaign involved part of the Czech republic, with a focus on Prague.
Outcome measures: Registration of responses to hotline, email, website.

Reasons for exclusion:
The study was not an evaluation.

Introduction:
The project Life Skills Education for Prevention of Trafficking in Human Beings was implemented in 11 boarding schools in Moldova. The goal was to facilitate the psychosocial integration of students from institutions, through life skills development, in order to reduce the risk of unemployment and trafficking in human beings.
Methodology: Posttest, through interviews, focus group discussions and questionnaires.
Outcome measures: Educational approach bases on life skills development through interactive methods, peer-to-peer training in life skills, change in children's knowledge, attitudes and skills, improvement of educational process from the point of view of life skills development for prevention of THB.

Reasons for exclusion:
The program did not focus specifically on sexual exploitation.
Introduction: This report has evaluated SIDA-financed counter trafficking initiatives in 2 countries: Kazakhstan and Kyrgyzstan. It concerns mainly the impact of the SIDA financed support.

Outcome measures: Not applicable
Reasons for exclusion: The study was not an evaluation and the program did not focus specifically on sexual exploitation. Methodology: Posttest, using interviews and questionnaires.
Outcome measures: Level of information and awareness.

Reasons for exclusion:
The program did not focus specifically on trafficking for the purpose of sexual exploitation. Introduction: This project evaluates the awareness raising materials from the UK, the Gambia, Costa Rica and Thailand by looking at campaign posters of child sex tourism.
Methodology: Posttest/market research, consisting of interviews.

Outcome measures: Not applicable
Reasons for exclusion: The study was not an evaluation on effectiveness.

PUBLICATION YEAR OF EVALUATION STUDIES
The search strategy identified 144 potentially eligible studies based when reading title, subtitle and abstract. The distribution of these studies over the last ten years is presented in table 2.