Conceptualizing sexual and gender-based violence in European asylum reception centers

Background Sexual and gender based violence (SGBV) is a major public health problem and a violation of human rights. Refugees, asylum seekers and migrants are exposed to a constant risk for both victimization and perpetration. Yet, in the context of European asylum reception centers (EARF) professionals are also considered to be at risk. Our study explores the conceptualization of SGBV that residents and professionals have in this specific context. Further, we intent to identify key socio-demographic characteristics that are associated with SGBV conceptualization for both groups. Methods We developed a cross-sectional study using the Senperforto project database. Semi-structured interviews were conducted with residents (n = 398) and professionals (n = 202) at EARF. A principal component analysis (PCA) was conducted to variables related with knowledge on SGBV. Chi-square test and Fisher’s exact test were applied to understand if significant statistical association exists with socio-demographic characteristics (significant level 0.5%). Results The majority of residents were male (64.6%), aged from 19 to 29 years (41.4%) and single (66.8%); for professionals the majority were women (56.2%), aged from 30 to 39 years (42.3%) and married (56.8%). PCA for residents resulted in 14 dimensions of SGBV representing 83.56% of the total variance of the data, while for professionals it resulted in 17 dimensions that represent 86.92% of the total variance of the data. For both groups differences in SGBV conceptualization were found according to host country, sex, age and marital status. Specific for residents we found differences according to the time of arrival to Europe/host country and type of accommodation, while for professionals differences were found according to legal status and education skills. Conclusion Residents and professionals described different conceptualization of SGBV, with specific types of SGBV not being recognized as a violent act. Primary preventive strategies in EARF should focus on reducing SGBV conceptualization discrepancies, taking into account socio-demographic characteristics.


Background
Sexual and Gender-based Violence (SGBV) is a major public health problem and a violation of human rights [1,2]. SGBV encompasses gender-stereotyped acts of violence, based on unequal power relations and denying human dignity, rights and development [1,3].
Considering the global challenge of (forced) migration [4], United Nations High Commissioner for Refugees (UNHCR) defines SGBV as "(…) violence that is directed against a person on the basis of gender or sex. It includes acts that inflict physical, mental or sexual harm or suffering, threats of such acts, coercion and other deprivations of liberty (….)" [1]. SGBV comprises five categories, namely, physical, psychological, sexual, socio-economic violence and harmful cultural practices [1]. SGBV conceptualization is a matter of judgement, affected by cultural beliefs, social norms and values [5]. What is considered a violent behavior varies according to specific determinants such as sociocultural and historical conditions [6].
Referring to SGBV conceptualization within the context of European Union (EU) policy documents, SGBV in general, and more specifically in migrants, has been framed as violence against women [7,8]. Yet, literature has demonstrated that female, male and transgender refugees, asylum seekers (AS) and migrants are vulnerable to SGBV [4,[9][10][11][12][13]. In a study on SGBV among refugees, AS and undocumented migrants in European asylum reception centers (EARF) a high percentage of multiple types of SGBV was reported in all sexes [14]. A study done in Belgium and the Netherlands found a high prevalence of direct or indirect SGBV exposure among migrant: 87/223 respondents had been personally victimized since their arrival in Europe. The majority of perpetrators were male (74.0%), and 69.3% of victims were female (male victims were 28.6%). Also, asylum related professionals were found to be assailants in one fifth of the reported violence [13].
A socio-ecological approach is described in the literature as an understanding model for SGBV [1,2,15]. The model assumes SGBV as the result of a permanent and dynamic interaction between health determinants at four levels: individual, relational, community and society [1,2,15]. A combination of these levels triggers the patterns of SGBV [16,17]. At individual level, research has shown that women and girls, especially the impoverished are more prone to victimization [13,18]. Recent evidence demonstrates that boys and men are also exposed to sexual violence [19]. In the context of EARF, both females and males have a tendency to be victims and/or perpetrators [14]. Furthermore, age [20], attained education and cultural beliefs appears to be important determinants when addressing SGBV [10]. At a relational level, children exposed to a violent context are more susceptible of becoming victims and/or perpetrators [21]. Further, a systematic review highlights that immigrant adolescents are exposed to high rates of violence [22]. From a community and societal perspective, studies have shown that an important determinant for sexual violence among refugees, AS and undocumented migrants is their restricted legal status [7] and the migration process itself [3].
In the context of migration, it becomes relevant to engage with affected communities [23] and to understand legal power relations triggered by society constructed knowledge, beliefs and norms that undermine refugees, AS and migrants, threatening their human rights and putting them at higher risk of SGBV [24]. Primary prevention of SGBV should focus on measures ensuring 'basic condition for sustainable and effective change' [25]. In this sense, a wide conceptualization of SGBV from an individual, relational, community and societal perspective is needed to promote a comprehensive prevention approach to violence [26].
Moreover, the intersectional nature of SGBV should be acknowledged while addressing preventive measure [8,27].
Our study aims to expand the understanding of SGBV conceptualization, in a vulnerable population of refugees, AS and migrants on the one hand and in professionals working with these communities in EARF on the other. Further, we identify sociodemographic characteristics of both groups that can be associated with SGBV conceptualization.

Study design
A cross-sectional study was conducted using data from the Senperforto Project developed in eight European countries (Belgium, Greece, Hungary, Ireland, Malta, The Netherland, Portugal and Spain). The main objective of Senperforto was to explore what knowledge, attitude, practice (KAP), and needs of professionals and residents from EARF were, in order to develop a gender-balanced European Frame of Reference for both beneficiaries [28]. Senperforto applied a community based participatory research methodology, mobilizing stakeholders -AS and refugees, asylum reception professionals, policy makers, civil society (…)from the participating countries in the community advisory boards. Further, community researchersprofessionals and/or residents that showed good social and communication skillswere trained (standardized training course) to conduct semi-structured interviews. Finally, a KAP survey was conducted.
For a detailed description of Senperforto project and methodology we refer to the article Sexual and Genderbased violence in the European asylum and reception sector: a perpetuum mobile [14].

Participants, sample and data collection
The Senperforto Project participant sample included 600 residents and professionals living and working in EARF. Residents refer to refugees, AS, and undocumented migrants. Professionals refer to services and health care providers working in the facilities. The inclusion criteria for the residents (n = 398) implied being member of the most numerous groups of asylum seeking and unaccompanied minor communities in the host country of research. They had to be staying at, or just having left, an asylum reception facility in the country of research. For professionals (n = 202), they had to work, or just stopped working at asylum reception facilities. Regarding the selection of facilities (open or closed (detention) facilities, reception or return centres, private accommodation, urban/rural, unaccompanied minors facilities, AS centres and refugee centres) all official reception facilities were listed; and facilities were selected in order to have at least one category of facility represented among the respondents. If more than one centre was available for a certain type of facility, centres were chosen randomly.
Also, a geographical distribution over the country of research was conducted and taken into account to the feasibility of the study. Considering that the situation of the asylum reception sector in each partner country differs, the sampling strategy was adapted to the local situation. In all countries random sampling was used except for Spain and the Netherlands where convenient sampling was applied due to political constraints [14].
.Data was obtained through semi-structured interviews that were conducted by well-trained community researchers. The questionnaire included data on sociodemographic characteristics of the participants and continued with three dimensions of research [1]: knowledge of the respondent on types of SGBV, on occurrence of violence and existence of prevention measures [2]; attitudes regarding SGBV and its prevention within EARF [3]; and a part on their evaluation of effectiveness of existing SGBV prevention and response measures and suggestions. Our study focuses on the first part of the questionnaire, which consisted of 82 closed questions coded with a Likert scale (I fully agree, I agree, Neutral, I do not agree, I fully disagree). Questions described the different acts of SGBV as put forward in the UNHCR guidelines on SGBV prevention and response [1] and inquired on a gender conceptualization, i.e. did they perceive the described behaviour as a violent act when it was done to girls and women, and subsequently if the same act happened to boys and men? Finally, the questionnaire was translated and back translated into the languages of the main groups of AS in the 8 participating countries, as well as the official language of each participating country (Arabic, Dari, Dutch, English, French, Greek, Hungarian, Portuguese, Romanes, Somali, Spanish, Russian, Maltese, Amharic and Tigrigna). A pilot test was done with members of the community advisory board. Prior to the interview respondents had agreed with the community researcher on the chosen language and sex of the interviewer. The interviews were conducted one-to-one at a private place in or near the asylum reception facility.
The Senperforto project applied the ethical and safety guidelines in researching violence recommended by WHO and UNHCR. Furthermore, it complied with the local ethical requirements and received ethical approval from Ghent University Hospital Ethical Committee [B67020096667].

Statistical methods
The questionnaires from Senperforto project included quantitative and qualitative data. For qualitative data, a framework analysis technique was used, further categorization and introduction into IBM® SPSS software. Quantitative data was introduced directly in IBM® SPSS software database. For our study we used a factor analysis approach using Principal Component Analysis (PCA) [29] for a factor extraction and Varimax rotation, to reduce the volume of the data. We conducted a multivariate analysis of 82 variables regarding SGBV knowledge. PCA analyses data representing observations described by dependent but inter-correlated variables. The goal is to extract the most important information from the original data and to convert this new information as a set of new variables, i.e. principal components (PC) [29]. These PC's were analyzed and named dimensions of SGBV, according to the questions with higher loading result from PCA output. The next step consisted of the recodification of the PC'sdimensions of SGBVinto nominal variables, each of them with three categories (negative, neutral and positive) according to the crosscut values for lower and upper barrier outliers. The lower fence outliers matched with the group of people that fully agreed with the dimension of violence in analysis while the upper fence outliers matched with the ones that fully disagreed.
Subsequently, we selected specific socio-demographic characteristics for residents and professionals. Commonly analyzed socio-demographic characteristics included: country of research (from here called host country), sex, age, marital status, religion, status according to immigration law and type of facility living/working (detention center, open reception center, local reception initiative, return center). Specifically for residents, we included the variables: having children, year of arrival to Europe and to hosting country, kind of accommodation (house, apartment, container, room, home-less…), attained education, daily activity in the country of origin and hosting country. For professionals we included: number of languages speaking and number of languages actually needed at work (here interpreted as language skills), to be working in a reception center by the time of questionnaires and the current occupation. Statistical tests were applied as the Chi-square Test and Fisher's exact test, to understand if significant statistical association exist at the 5% significance level.

SGBV conceptualization Residents
When analyzing the results of the multivariate analysis of principal components, we found 14 new variables, representing 83.56% of the total variance of the data. These new variables were analyzed according to the questions with higher PCA output loading, labeled as dimensions of SGBV according to UNHCR definition [1] and represents residents SGBV conceptualization. The questions that correspond to each dimension are described in Table 2.

Professionals
The multivariate analyze for the group of professionals resulted in 17 new variables representing 86.92% of the total variance of collected data. These new variables were analyzed and labelled dimensions of SGBV [1] representing professionals SGBV conceptualization. The representative questions of each dimension of SGBV are described in Table 3. Table 4 shows the conceptualization of SGBV for residents and professionals from EARF grouped according to UNHCR SGBV definition [1].
The association between each dimension of SGBV conceptualization and resident's socio-demographic characteristics or professionals' characteristics are presented in Tables 5 and 6, respectively. Our results describe whether, or not, what is considered a specific behavior/sexual act as violence is different according to sex, age, kind of accommodation (…). We will now describe the significant results first for residents and subsequently for professionals.

Sexual violence
For residents, sexual innuendo conceptualization was associated with the host country (p = 0.010), kind of accommodation (p = 0.026), the level of education of residents (p = 0.016) and daily activity in the host country (p = 0.037). This mean that residents living in Belgium and Ireland, in a container, studio or room, with an education (primary, secondary or higher), or do not have a job in the host country tend to disagree that sexual innuendo is a type of SGBV.
Marital rape was associated with the age of residents (p = 0.001), and the kind of accommodation where they were living in (p = 0.001). Youth and adults' residents (0-39 years old) or residents living in containers, room or studio tend to disagree that marital rape is a form of violence. Abuse, rape and trafficking was associated with

Psychological violence
The concept of confinement was significantly associated with age (p = 0.032), meaning that residents aged until 18 years old tend to disagree with confinement as a form of violence.

Harmful cultural practices
Denial of education for girls was associated with marital status (p = 0.033) and the fact of having (or not) a religion (p = 0.019). Single residents tend to fully agree with this as a form of violence. The conceptualization of genital mutilation as a form of violence was associated with attained education (p = 0.033).    Honor killing and maiming conceptualization was associated with the country of research (p = 0.001), sex (male or female) (p = 0.004) and age (p = 0.042) of residents. Residents hosted in Belgium and Greece, male or aged from 19 to 39 years old tend to disagree with this concept as a form of violence.

Socio-economic violence
The concept of social exclusion based on sexual orientation was associated with the time of arrival to Europe or hosting country (p = 0.018 and 0.007), and daily activity in the country of origin (p = 0.046). Residents that arrived recently to the host country or Europe (less than 5 years) or used to have a job in the country of origin tends to fully disagree that social exclusion based on sexual orientation is a form of violence.

Sexual violence
For professionals, sexual innuendo was associated with language skills (p = 0.012). Professionals with good language skills (at least 2 EU languages) tend to fully disagree. Visual sexual harassment conceptualization was associated with language skills (p = 0.038) and status immigration law (p = 0.037). The tendency to disagree was found in professionals without the national citizenship and basic language skills (1 EU language). Denudement was associated with the hosting country and language skills (p = 0.030, p = 0.000, respectively). Professionals from Portugal or with basic language skills (1 EU language) tend to fully disagree. Abuse, rape and trafficking conceptualization was different according to the age of professionals (p = 0.021). Older professionals (> then 40 years old) tend to fully disagree. Further, sexual exploitation as a form of violence was associated with hosting country, marital status and language skills (p = 0.002, p = 0.014 and p = 0.031). Tendency to fully disagree were found in professionals from Malta, Netherlands and Portugal, married or with good language skills (1 EU and 1 non-EU language).

Physical violence
The concept of physical assault without permanent consequences as form of violence was significantly associated with hosting country (p = 0.015). Professionals working in Hungary tend to fully disagree.

Psychological violence
Verbal violence was associated with marital status (p = 0.042), with single professionals disagreeing more than the average of respondents. Confinement (individual level) as a form of violence was associated with host country (p = 0.004), status of immigration (p = 0.001), language skills (p = 0.040) and the fact of being working (p = 0.005). Professionals that have a tendency to fully disagree were from Belgium and the Netherland, without the national citizenship, with good language skills (2 EU languages) or with a current job at the time of the questionnaire.

Harmful cultural practices
Genital mutilation was associated with professionals' sex (p = 0.043), meaning that male professionals tend to fully disagree with it as an act of violence. Early marriage as form of violence was different according to the hosting country, type of reception facility, language skills or the fact of being working (p = 0.001, p = 0.027, p = 0.047 and p = 0.031). Professionals working in Belgium, in open     reception facilities or with good language skills tend to fully disagree.

Socio-economic violence
Denial of opportunities and services as a form of SGBV was associated with sex (p = 0.049) and female professionals tend to fully disagree that it represents a form of violence. In sum, our results suggest that professionals from EARF considered more behaviors as violence than residents.

Discussion
The scientific understanding of violence and more specifically SGBV is primordial [6] to enhance primary preventive measures. In this sense, if we want to prevent violence in the EARF, understanding the knowledge that residents and professionals have regarding SGBV conceptualization is needed. Our study explored SGBV conceptualization according to residents and professionals from EARF, covering a myriad of countries of origin of the refugees, AS and migrants. Our results show a disparity between what is, or what is not considered a violent behavior. Professionals have shown to have a wider knowledge then residents, considering more acts as violence. We believe this can be related to residentsrefugees, AS and undocumented migrantsbeing described as more vulnerable to SGBV and professionals assuming a privileged position and control towards residents [24].
Conceptualization is a process of development and clarification of concepts; it shapes the field in which a concept is understood, measured and evaluated [30]. Different SGBV conceptualization can be found in the literature. Walby [8] refers that different definitions are used for assault and for rape, which are inconsistent and out of alignment with international law. Also, different SGBV conceptualizations were found in our results for residents and professionals. To consider that definitions of violence have evolved through multiple variations according to the field and the range of forms of violence [30]. A consistent and coherent measurement of violence against women and men will benefit accuracy while measuring changes in society and effectiveness of public services [8]. Given this, we believe a common SGBV conceptualization should be considered while addressing preventive measures. The requisite for developing information, education and communication (IEC) interventions addressing SGBV has already been acknowledged by UNHCR (2003). We believe our results call for the urgent need for IEC interventions, addressing what is, or what is not an SGBV act.
For both groups differences in SGBV conceptualization were found based on specific socio-demographic characteristics. As for gender, our results evoke no differences in SGBV conceptualization. Moreover, the fact that a violent act is directed to a girl/woman or a boy/man is equally considered violence, even though the majority of victims continue to be women [13] However, moving from SGBV conceptualization to specific types of SGBV differences arise. When conducting association tests between types of SGBV and the gender of our respondents we found significant associations. A more in-depth analysis suggests male residents tend to disagree that honor killing and maiming is an SGBV act when compared with the mean average of our respondents. Moreover, male professionals disagree with genital mutilation as a form of SGBV, and female professionals tend to disagree with the denial of opportunities as a form of SGBV.
Another relevant association was found between age and specific forms of SGBV. Results from our study, found that professionals aged above 40 tended to disagree that "abuse, rape and trafficking" is a form of SGBV. This association is particularly screaming for action, once we assist to professionals working with persons, already in a vulnerable situation, and assuming that a behavior legally punishable by law is acceptable. Considering that professionals play an important role in SGBV prevention, and the fact that they are in a privileged position to mitigate SGBV, we believe that our results are screaming for action. From one side we assist to professionals having a broader SGBV conceptualization when compared with residents. However, professionals aged above 40, do not consider abuse, rape and trafficking as a form of SGBV. In this sense, we believe there is a need for a strict screening when engaging professionals to work in EARF and continuous sensitization and training on SGBV. Our results are aligned with previous evidence reporting the requirement for healthcare workers' regular training [25], integrated and widespread preventive and response measures [14]. Furthermore, professionals and persons in power working with migrants and refugees have been identified as potential perpetrators of SGBV [12,24,25]. In EARF context, professionals have been identified as potential perpetrators of SGBV, especially socio-economic violence [14].
Specific types of SGBV not being recognized as a violent act is of major importance while addressing preventive measures in asylum centers. Residents and professionals must have a complete and equal knowledge regarding all types of SGBV to avoid being victims and/or aggressors. Placing SGBV in a public health perspective, we can assume SGBV conceptualization is the baseline for primary prevention [5]. Furthermore, significant association with socio-demographic characteristics have arisen from our results. This fact shows the importance of recognizing the intersectionality of SGBV concepts [8,27] with characteristics, such as gender, age, social status. We call for an urgent action from all stakeholders to increase the knowledge on SGBV of residents and professionals, based on IEC interventions, as the baseline to prevent violence before it occurs.
Future pertinent research regards the potential association between SGBV conceptualization and case disclosure.
Moreover, it is of utmost importance to have a clear and in-depth understanding of professionals' SGBV conceptualization. The fact that professionals might perpetuate SGBV acts, and exercise a higher power relation towards residents, represents a call for intervention. We challenge researchers to go beyond the understanding of professionals' SGBV conceptualization and to consider the influence of it with the potential perpetuation of violence. Another relevant aspect to consider in the future regards the evaluation of primary preventive measures, and specifically the focus on promoting and implementing a widespread SGBV conceptualization among residents, professionals and host population. If we reach a level where professionals and residents have similar SGBV conceptualization, will we still witness such high levels of SGBV?
Even though relevant findings were described it is important to acknowledge potential limitations. The Senperforto project applied multi-types of sampling methods, as random and representative sampling were not possible in all countries. However, even though our results cannot be generalized, we believe it can be transferable to similar populations in comparable contexts, in a sense that a broad SGBV conceptualization is presented in our researchunderstanding refugees, AS and undocumented migrants' perspective and also professional's perspective. Specifically related with SGBV conceptualization, we cannot exclude that community researchers conducting the interviews during the implementation of Senperforto project, could have had a different SGBV conceptualization, even with the implementation of a standardized training.
Stepping out of EARF, it would be pertinent to compare SGBV conceptualization between migrants and hosting population, once public health policies should be adapted to the cultural and structural context. Moreover, it is important to consider the challenge of having refugees, AS and undocumented migrants with different SGBV conceptualization "integrated" in European countries, especially if they have a narrow concept. Accordingly, we believe migrants might be exposed to higher vulnerability to both victimization and perpetration. Considering the recent migration wave to European countries, it urges to address this issue. SGBV conceptualization needs to be addressed equally, not only for migrants and professionals, but also for hosting populations. What is or what is not an SGBV act should not differ according to a migration status. By not doing it, we believe European countries and its representatives might be increasing migrants' vulnerability and inducing obstacles to their integration.

Conclusion
Residents and professionals from European asylum centers have a different concept of what SGBV entails with professionals considering more acts as violence then residents. However, types of SGBV were considered equally violent if afflicted upon female or male. Some acts that were not considered violence by the professionals are legally a crime, increasing the perpetration risk.
The Socio-Ecological Model as an explanatory model of SGBV helps moving from the individual conceptualization of SGBV to a societal conceptualization considering the influences of relational, community and societal factors [31].
SGBV conceptualization is the core to primary prevention of SGBV and it should focus on harmonizing the concept, IEC activities, training and "collegiate" discussion/participatory activities towards consensus and European policies. What is considered (or not) a violent behavior should be taken into consideration if we want to mitigate SGBV.
We call for the development, implementation and monitoring of European-wide SGBV prevention programs in EARC context, aligned with SGBV conceptualization of the target population.