What Social Impact Does Exposure to Domestic Violence Have on Adolescent Males? A Systematic Review of Literature

Domestic violence is widespread in the United Kingdom (UK) and accounts for 14 per cent of all violent crimes. It is a significant health concern, as it damages physical and emotional health and can have long lasting negative impact across a wide range of health, social and economic outcomes for families as well as having a major impact on the social development of young people. A recent report shows that in 2002 there were 750,000 children in Britain estimated to have witnessed domestic violence. However, studies demonstrated that children’s exposure to domestic violence between parents tend to seal its “inexpugnably prints” on “their impressionable minds”, they are thought to express more anger, antisocial behaviour, as well as fear, anxiety and depression have greater risk of behavioural, emotional, and psychological problems. It shows that exposure to domestic violence in adolescence is thought to be having more of an impact than exposure in younger children, with earlier exposure having little or no impact without adolescent exposure. This systemic review focuses on the exposure of male adolescents to domestic violence and the impact of domestic violence on them. The review highlights that there were a number of disparate ways in which domestic violence impacted on adolescent males. This ranged from animal cruelty, drug taking, violence and aggression, through to feelings of depression, suicidal ideations, and feelings of sadness and ambivalence. However, the results suggest that while the exposure to domestic violence does indeed have an impact on adolescent males, there appears to be no agreement on any single way in which this impact may be experienced. However, we recommend that combating the scourge of domestic violence requires skilled practitioners to intervene at any given stage. As such, practice educators must ensure that staff under their remit are adequately equipped with the necessary knowledge and skills of detection, to impact on the root causes of domestic violence. This must include increasing awareness among staff of the knowledge of the latest legislation and policies. In addition, training should emphasise the need to maintain liaison with external agencies in developing pathways into services, and timely responses by stakeholder and associate institutions to adequately address this issue that is impacting negatively on male adolescents.


Introduction
Domestic violence refers to any incident or pattern of incidents of controlling, coercive threatening, violent or abusive psychological, physical, sexual, financial or emotional behaviours between those aged 16 or over who are, or have been intimate partners or family members regardless of gender or sexuality and includes issues of concern to Black and Minority Ethnic (BME) Domestic violence impacts all of society and is widespread in the UK, and accounts for 14% of all violent crimes in the UK [8]. In 2011 the Director of public prosecutions stated in one of his key note speeches that "domestic violence is serious and pernicious. It ruins lives, breaks up families and has a lasting impact. It is criminal and it has been with us for a very long time". Yet, it is only in the last ten years that it has been taken seriously as a criminal justice issue. Before that, the vast majority of cases were brushed under the carpet with the term refrain "it's just a domestic [9].
Domestic violence is a significant health concern, as it damages physical and emotional health and can have long lasting negative impact across a wide range of health, social and economic outcomes for families as well as having a major impact on the social development of young people [1]. For example, more than one in four women has experienced at least one incident of domestic violence in England and Wales since age 16, which equates to 4.8 million women, while two women per week are killed by their current or former partner [1].
Even more remarkable is the fact that by virtue of the fact that domestic violence occurs within the family home, millions of children are therefore exposed to domestic violence. According to Women's aid (2015), in 2002 there were 750,000 children in Britain estimated to have witnessed domestic violence [2]. Utilising retrospective adult reports, Henning et al. found that there is a 20-40% exposure rate in childhood to domestic violence [10]. In defining exposure to domestic violence, Mohr et al. indicates that there has been no consensus among researchers regarding the definition of children's exposure to domestic violence [11]. However most describes it as seeing hearing, or directly involving the young person. Jouriles et al. and Edleson goes further to suggest that while some researchers describes it as children's witnessing or observing domestic violence, it goes further and involves all that as well as the experiencing of the aftermath of domestic violence such as physical or sexual assault that occurs between care givers [12,13].
It is clear that there has been many research explorations into domestic violence and its impact on children, with some researchers proposing that there are links between childhood exposure to domestic violence and subsequent social behaviours, such as bullying, offending, drug abuse, low self-esteem, running away, with adolescents 6 to 9 times more likely to have attempted suicide or have suicide ideations [3,[14][15][16]. In addition, Edelson proposed that children's exposure to domestic violence between parents tend to seal its "inexpungible prints" on "their impressionable minds", they are thought to express more anger, antisocial behaviour, as well as fear, anxiety and depression [13]. Similarly other studies have reported children to have greater risk of behavioural, emotional, and psychological problems if they were exposed to domestic violence [5,17]. Conversely, other researchers proposed that the impact of domestic violence on children may vary with age and they may be less symptomatic of its impact with maturity [18,19]. In contrast Thornberry et al. reported exposure to domestic violence in adolescence as having more of an impact than exposure in younger children, with earlier exposure having little or no impact without adolescent exposure [6].
From a gender, perspective, research suggest that there are gender differences in the externalization behaviour problems as a result of domestic violence, boys are thought to have more externalisation while girls internalise more [20]. However adolescent boys are thought to express more sadness while girls express more anger [21,22]. Other research studies caution against these findings of these findings as they postulated that exposure to domestic violence is not a "homogenous onedimensional phenomenon and some research has been beset with many methodological complexities [23,24].
There is a significant body of research relating to domestic violence and its impact on women and children with disparate findings [4,25-28] A systematic review is warranted whenever there is a substantive question, several primary studies-perhaps with disparate findings-and substantial uncertainty [29]. Using recognised tools and methods, this systematic review synthesised the findings of current research on the impact of exposure to domestic violence on adolescent males and made recommendations for future practice within health and social care.

Review Question
These authors have utilised a PEO research tool in the formulation of the research aims in this process as it is thought to be better suited to the research question and will structure the search process. This research study will therefore aim to ascertain 'What impact does domestic violence have on the social wellbeing of adolescent males? For the purposes of this systematic review, exposure will utilise Edleson's definition, of witnessing, observing and or experiencing of the aftermath of domestic violence for any period between the care-givers of adolescent males [13]. The definition will exclude actual abuse of the young persons' themselves, but instead will refer to domestic violence inflicted on their care givers which in term affects the young person. The actual experience of abuse is beyond the remit of this paper as it would require an entirely different focus in relation to experiences of victims.

Aim
The aims of this review is to systematically identify studies related to the impact of exposure to domestic violence on adolescent males and to describe the key features of these studies as it relates to this issue. In particular the exposure to domestic violence will be explored across the domain of prevalence and consequences on social wellbeing as it relates to adolescent males.

Objectives
The objectives of the review are to address two questions: To identify any, prevalence of domestic violence exposure on adolescent males.
To identify consequences of domestic violence including effects on social wellbeing of adolescent males.
To make recommendations for practice education based on findings.

Rationale for this review
Both national and local data show that domestic violence is highly prevalent in the UK. The data analysis shows a growing concern about the impact of domestic violence on children and young people, which affects them mostly through exposure to violence perpetrated against their mothers [31]. Children are often the hidden or silent victims of domestic violence [32] and to help the children, all agencies such as Social services, National Health Services (NHS), schools, the courts; national and local authorities must put child protection policies and guidelines in place. All professionals who serve children must increase efforts to identify children who are exposed to domestic violence and apply the child protection guidelines where necessary. This systematic literature review seeks to add to the body of knowledge by synthesising the findings of current research on the impact of exposure to domestic violence on adolescent males and made recommendations for future practice within health and social care.

Methodology
In recent times systematic reviews have been hailed as a robust method of identifying, selecting, appraising and synthesising all high quality research evidence relevant to a single question" [30]. This systematic review method engenders a thorough and all-encompassing summary of relevant literature to the question posed and provides a most reliable source of evidence to guide practice. The method of research allows for a precise structured process of systematic review that aims to minimise bias, improving the quality and accuracy of the conclusions that was drawn [33]. The process for producing a systematic review involves these clear stages: framing of the review question, identifying relevant studies with a clear search strategy, specifying inclusion and exclusion criteria for studies, search strategy, sifting process, critical appraisal and process of data extraction; summarising data and synthesizing data [34].
According to Aveyard, once the research question has been developed, then a systematic review of the literature should be undertaken to ensure that the widest range of published material is obtained to answer the research question comprehensively, using all available published literature [35].
These researchers had undertaken a mixed method systematic review of the literature over the last ten years (2005-2015) on domestic violence exposure and social disadvantage relating to adolescent boys, through a qualitative and quantitative appraisal of published research. Furthermore, in order to maintain validity and reliability of the research process these authors will maintain the use of the P.E.O framework to govern the search terms, paper selection, and the data extraction process. The Cardwell et al. tool will be employed to critique the paper quality. All results of the study will be presented as well as the limitations of the study [36].
However, in our attempts to undertake this research study, we were confronted with an ever increasing amount of information, with significant time limitations together with inadequacies of traditional sources of information to meet the expectations and criteria of modern academic studies, all presents substantial challenges. From a historical perspective, for instance, it has been suggested that expert opinion has often been presented in narrative reviews which are not necessarily evidence-based, and as such is prone to limitations [37]. Likewise, Montori et al. argued that unsystematic narrative reviews are more likely to include only research selected by the authors, and as a result introduces bias and frequently lag behind and contradict available evidence [38]. In Clarke view not sure what this was originally doesn't read right, systematic assembly, critical appraisal and synthesis of all relevant studies on a specific topic [33]. In view of these realities, we endeavoured to undertake a critical review of the literature using search strategy from a number of sources.
There is a significant amount of published research relating to exposure to domestic violence with varying methodologies and outcomes as significant impact on children and young people [4,25-28,39] Many of which according to Holt et al. have been beset with methodological complications and concerns [24]. We envisaged that the careful and meticulous approaches we adopted would ensure robust procedures are utilised thus increasing validity and reliability of the outcomes of this research. According to Denzim and Lincoln the combination of approaches to the study of the same phenomenon enables cross validation thus facilitating greater reliability of the study [40].

Data sources and search strategy
Cumulative Index to Nursing and Allied Health Literature (CINAHL) recognised as a comprehensive and authoritative source for professional literature related to nursing, biomedicine and allied health, Medline the main source for bibliographic literature, psycarticles, psycInfo both recording professional articles relevant to psychology and sociology, Soc Index, child development and adolescent studies, education research complete, Cochrane library, BNI, Psychology and behavioural science collection, Scopus, social care online, ERIC and TRIP were also searched. These data bases were selected as they were considered to be the most comprehensive scholarly research data bases that were relevant and applicable to the question. three elements. It began with a systematic search of the above data bases. Passer proposed that a broad and comprehensive search directly underlies the quality of the research [41]. In addition, Bruce proposed that not all published research can be found on data bases so an attempt should be made to search all grey literature [42]. In response to this view, these researchers search internet sites such as Google scholar, some governmental sites and sites of other reputable Organisations such as the NSPCC and women's aide sites were also searched and finally bibliographies and references were also checked. In addition, Fonseca stressed that appropriate search terms or key words are fundamental to an effective literature search [43]. These authors therefore began with the research question listing the key elements of the question into search terms with synonym's relating to the P.E.O framework (See Appendix A for the complete list, with Boolean operatives).

Results of the search strategy
The search yielded 2,489 papers on domestic violence or interparental violence and adolescents. Other search terms included for example domestic violence, or inter-parental violence and adolescent and males or boys which yielded 539 papers, while domestic violence and gender and prevalence yielded 381 papers. Boolean search terms and operators were used to limit and widen the search as appropriate and the literature relating to domestic violence and male children were also searched when they were found to include any of the searched terms with reference specifically to adolescent males. Some of the data bases were also combined in searches reducing duplication. From the search of the Grey literature 28 papers were retrieved 22 of which were excluded as they were mainly reports, editorial discussions papers and policy documents and as such did not meet the inclusion and exclusion criteria.
From the combined searches of all the data bases, web pages, and the hand searches of the grey literature, reference list and bibliographies 94 papers were extracted for analysis.
Of the 94 papers extracted 40 were duplicates and therefore 54 extracts were read, of these 35 were removed as they did not meet the criteria for inclusion see the inclusion and exclusion criteria. 18 full text studies were retrieved and another 6 removed as they did not meet all the inclusion criteria. Therefore eligibility criteria yielded 12 studies as the primary research papers and deemed suitable to answer this research question based on reading the titles, abstracts, and full text of the papers. Selecting the optimal tool requires effort to obtain the necessary functionality (See the list of included papers below and Appendix B for the entire list of included and excluded paper details).

Selection criteria
As these authors were undertaking this systematic review, it was felt that collecting primary research papers would be important [30]. Papers written in English from any country were collected as it was felt that this will allow for an international perspective, applicability and appreciation of the findings of the review. While it may be argued that research specific to a particular population, country or culture will reveal findings specific to the area from which data is being collected and possibly reduce generalizability, it was felt that to exclude some papers based on country of origin could equally result in presentation of bias findings, and therefore all articles including the key search terms were included regardless of country of origin. Conversely, these authors acknowledged that to exclude papers not written in English will add bias to the study but to have done so would be beyond the scope of this review. Articles were selected according to the predefined inclusion and exclusion criteria in order to reduce selection bias [37]. All the papers generated from the search have been managed using 'Refworks' which saved and managed references, and generated bibliographies.

Data extraction and quality appraisal
These authors' intention was to capture all relevant references covering the topic as wide as possible in order to identify any gaps in the evidence where available. Hausner et al. suggest that a concept-based, subjective approach is the traditional way of undertaking a search strategy, but cautioned that this method relies heavily on the researcher's knowledge of the topic under investigation [44]. All the above data bases were therefore searched, and abstracts of research based papers, peer reviewed papers both published and unpublished relating to the key concepts of adolescent males, domestic violence, prevalence, and social impact, using the P.E.O framework were extracted. The literature relating to domestic violence and its impact on children were also examined as it transpired that there were some documented evidence relating to adolescents within this literature. Additional searches, grey literature, reference list and bibliographies revealed another six papers that were relevant to the questions of this research paper.
Initial results from the data bases revealed one hundred and fourteen published papers having used the key search terms, after application of Boolean operators these were reduced to ninety four papers. Forty of which were removed due to duplicates. Fifty four abstracts were then read and full text papers retrieved for eighteen of these thought to be relevant to the question. Of those papers extracted, six of them were not relevant to this review as they either were not gender specific or adolescent specific, or not relevant to the key outcomes being researched. Final papers selected equalled twelve papers (See Appendix B for the summary of the papers selected and criteria for paper selection).
The screening of the papers was conducted by the first reviewer but was supervised and checked by the second reviewer. Similarly, data extraction was conducted by the first reviewer using established data extraction forms and independently checked by a second reviewer. Bethany Saltikov outlines that appraising the quality of articles is fundamental in the judgement of the quality of the study results, and ultimately the conclusions drawn and its subsequent applicability to practice [30]. As such it is important to outline who will be doing the quality assessment how many people as more than one person increases cross referencing and reduces bias.
The methodological quality of the included studies was therefore appraised by the two reviewers using established data extraction forms. Bettany-Saltikov proposed that in order to conduct an effective quality assessment a robust quality assessment tool is required, and a tool suitable to meet the study design should be selected. Twelve primary research papers were extracted for this systematic review, eleven of which were Quantitative papers, and one qualitative paper. The study designs among all the papers reviewed were all different and some had mixed methods, some were longitudinal studies others were correlation studies and yet still there was the single ethnographic qualitative paper. In view of this the author elected to utilize a mixed method framework as proposed by Caldwell et al. to analyse the papers (see Appendix C for the framework) [36].
All the twelve papers extracted in this systematic review were read and quality appraised, to assess the extent to which the papers have attempted to reduce methodological bias. Petticrew and Roberts proposed that assessing quality relates to the internal validity of the study, and whether it is free from methodological bias [45]. Assessing quality in systematic reviews as in the case of this review, relates to the methodological quality. Higgins and Deeks outlines that appraisal of a study is an attempt to judge the degree to which the study is an approximation of the truth [34]. In addition, Bettany-Saltikov suggests that a comprehensive appraisal requires a systematic approach in which the strengths and weaknesses, and validity of the research can be judged [30]. Thus, the twelve extracted papers were appraised against the eighteen questions set out in the Caldwell assessment tool, and the papers marked against the criterion "yes", "no" or "partially" [36]. The results were represented in tabular form and will be discussed further in the results section of this systematic review.
In addition, these authors chose a Likert scale to assign a numerical value to the answers gained in the Cardwell assessment tool and also presented the results of the assessment with a numerical value; adding to the ease of assessment of the overall quality of the papers. While this will not affect the inclusion criteria, as the findings from all the papers will be presented, it is envisaged that this valued scale presentation will enhance the appreciation of the quality of the papers selected. For each individual answer to the eighteen questions in the Caldwell tool, a value will be added 0=No, 1=Partially, and 2=Yes, allowing a maximum score of 36 for each paper. Papers with a score of 36 will then be judged as the best quality papers (refer to Appendix C for these). This process carefully analysed all the extracted research papers included in this systematic review, the authors' attention must now turn to the data extraction process undertaken.
The data extraction phase in this review was the most challenging [30]. Thus, careful consideration was given to extract data relevant to the salient issues being studies, which are aimed at answering the research question. As such, a specified data extraction matrix was created considering the population being studied, the interventions, the comparative interventions, the outcomes and recommendations. The process of the data extraction was therefore aimed primarily at answering this question. Higgins and Deeks proposed that in order to standardise the process and improve the validity of the results, a data extraction form should be created and its effectiveness checked (see Appendix D for data extraction matrix) [34].

Data analysis and synthesis methods
Parahoo suggest that data analysis is a process of information gathering techniques and the organisation of raw data [37]. There are many different tools used in the data analyses and synthesis, depending on the type of data being analysed. For example, quantitative data may include measurements such as correlational measures, nominal measurements, ordinal measurements or even ratio measurements, while qualitative data is more concerned with presentation of description data using themes, sub themes and meta-synthesis [46]. Because the studies reviewed for this paper have included a wide range of research designs, and included mostly quantitative papers but also included a qualitative paper, and more importantly as the studies have all yielded insufficient similarities in the variables being measured. These authors have chosen to present a narrative synthesis of the data collected [47].
The narrative approach, therefore, involved collating the data from the studies included in each review paper to present a summary of the outcome supported by a discussion. This included identification of emerging themes relating to the impact of exposure to domestic violence on adolescent males. This is in-line with Bettany-Saltikov view that the use of thematic presentation of results is commonly used for qualitative papers. However, where studies included in order to answer the research question are insufficiently similar, with a range of study designs and methods, or where mixed methods and both qualitative and quantitative papers are included, it may be best to seek emerging themes to present a narrative synthesis of the findings [30]. The selected papers for this review were all dissimilar in the variables being measured, in the methodologies used as well as outcomes obtained. Despite the extensive heterogeneity among the papers what emerged was that the outcomes were extensively either externalising factors or internalising factors, it was therefore found that presentation of the findings using the externalising and internalising themes would best allow accurate presentation of the findings. Internalising behaviour was therefore taken to represent and included measurements of mental health concerns such as depression or anxiety, identity problems, and coping issues. While externalising behaviours included issues such as violence, conduct disorder, withdrawal, behaviour concerns, bullying, aggression, or drug use.

Results of the quality assessment of included papers
All the included papers were critically analysed using the eighteen questions set out in the Caldwell et al. assessment tool, and the papers marked against the criterion "yes", "no" or "partially". As eluded earlier, these authors have also chosen to utilise a Likert scale to assign a numerical value to the answers gained in the This article is available in: http://healthcare-communications.imedpub.com/archive.php Cardwell assessment tool. For each individual answer to the eighteen questions in the Caldwell tool, a value was added 0= no, 1= partially and 2= yes, allowing a maximum score of 36 for each paper, a score of 36 was deemed the highest quality [36].
Results from the critical appraisal of the included papers using the Likert scale tool revealed that all the papers were of high quality with scores ranging from a minimum score of 30 to a maximum score 36. From the Cardwell et al. tool it was clear that all the papers had clear titles which reflected the content of the studies as well as clear aims, methodologies, study designs with rational for choice, all had clearly defined populations, abstracts and rational for each study [36]. The Qualitative paper included in the study will be discussed additionally at the end of this summary.
From the documented credentials the researchers all appeared to be credible and suitably qualified to undertake the studies done. Credibility check is particularly important in this review as it serves as a common criterion used to evaluate research studies and is one of the most important factors in establishing trustworthiness [37].
While ethical considerations were not explicitly documented for all the papers, each had comprehensive literature reviews, except for paper 1 and paper 2. One would suggest that this may be due to the specificity of the population and variables being studied, and possible lack of previous research in that area, this was acknowledged by (Hindin and Gultiano 2006) in paper 2. However, Beauchamp and Childress argued that it is important to note that ethical issues often arise at different stages in a study and may be dealt with at a particular stage rather than under a specific heading of the study [48].
Eight of the papers adequately described the sample and were reflective of the population; however this was not found to be the case in papers 3, paper 6, paper 8 and paper 11. In paper 3 for example, the sample was a convenience sample of inpatient adolescent on a psychiatric unit, for whom the self-reporting questionnaire design were not reliable and did not represent the population as a whole this limitation was acknowledged by the authors of the paper. While non-probability sampling affords the researcher a ready population, the sample chosen may or may not be representative of the entire population [37]. In addition, paper 6 reported to have lost some of the data during collection as such were unable to report entire findings of the study. The implications and consequences of this loss of data are not known. However, the researchers did acknowledge this attrition concern in their reporting of the findings.
Among the papers selected there were 6 papers found to have reliable data collection these included paper 2, paper 3, paper 4, paper 8, paper 9 and paper 12. Paper 1 and paper 2 was found to have data collected from inpatients records in a residential unit which were found to have discrepancies at times, and prevented generalizability of the findings. The researchers in paper 1 accepted this as a limitation to the study reporting that this data was used due to time constraints, financial constraints and accessibility issues.
It may be argued that a candid discussion of limitations could assist readers in correctly interpreting a particular study and frank and open discussion of limitations may inform future studies, which potentially could result in higher quality if these limitations of earlier studies are addressed [49].
Other concerns with data collection in paper 2 included issues of self-reporting of depression when depression rates were being measured as a variable the researchers acknowledge this could have resulted in underreporting therefore affecting the findings. Other issues related to lost and missing data as well as reduced attrition rates were reported among the papers undertaking longitudinal studies, such as in paper 6, paper 10 and paper 11. Attrition rates can result in bias from skewed data and may result in unpredictable and inaccurate outcomes.
In paper 7, the data was also collected from samples in a residential psychiatric unit, which was not representative of all adolescent males and the variables being measured could not reliably answer the questions being asked in their research. Though there were concerns regarding loss, missing or unreliable data collection among the papers, data analysis among the papers were mainly clearly outlined and explanations provided to support the approaches adopted. However, due to the variation in the study designs and the variables being compared and assessed among the papers, there were a variety of tools utilised in the analysis of data.
For ten of the papers, the methods of analysis appeared valid and reliable. However, Calvete and Orue highlighted that due to the schemas being measured, which included the measurements of feelings of mistrust and victimization as well as perceptions of victimizations, the analysis of the data could be unreliable due to difficulty measuring these schemas [50].
All the papers had clearly and appropriately presented results with comprehensive discussions and conclusions. However, in terms of generalizability of the findings of the papers, only two could be considered as having results which could be said to be a general reflection of adolescent males. There was an over representation among the papers of convenience sample collection from residential and treatment centres which limited generalizability of approximately half of the papers. This type of sampling offers simplicity but represents a weakness due to selection bias [37]. In addition most of the papers were found to have limitations due to population specific characteristics which did not allow for application of findings to general adolescent populations such as in paper 2, paper 7 and paper 9.

Summary of included studies
All the studies were related to adolescent males, whose ages ranged from twelve to twenty years.
Ethnicity and Social class of the participants varied from study to study, which in itself had implications for the findings of some of the studies as in paper 4. This study looked at family violence and children's behaviour and mental health, all the participants in this study were reportedly selected from what was reported to be lower class two parent families. In the context of British social stratification, lower class refers to individuals who are semi-skilled and manual workers. Since the authors of this study did not explain what was meant by lower class, the results of the study cannot therefore be said to be transferable. While various methods were used to recruit the samples and included non-probability and random sampling, cluster sampling, and randomised sampling, among the twelve papers half of the studies' utilised populations from residential treatment centres and units which contained a convenient sample of adolescents receiving treatment and support for behaviour concerns. It could be argued that this convenience sampling has implications for generalizability due to possible selection bias [46]. Sample sizes varied across the twelve papers resulting in some researchers reporting their sample sizes to have been too small to adequately address the phenomenon being studied which has been a limitation to their study as in paper 3. Conversely others reported larger sample sizes of over five thousand participants as in the study paper 6. Interestingly none of the twelve studies were British based studies, almost half were American, the rest Danish and Finnish, one was Spanish, Filipino and one Israeli. Nevertheless the implications of this will be discussed in the strengths and weaknesses section of this paper.

Summary of the papers in terms of Prevalence of domestic violence exposure on adolescent males
While the authors have not produced actual figures in support of their claims in the presentation of the results of the studies in relation to the prevalence of exposure to domestic violence and outcomes, results generally highlighted that there were varying outcomes.
All the studies were concerned with measuring the effects of, or the relationship domestic violence witnessing had on adolescent males through the measurements and assessment of different outcome variables (See the Appendix).
Paper 1 was the only paper to have assessed the significance of witnessing domestic violence and resultant animal cruelty in adolescent males. Six of the papers were concerned with effects on behaviour, conduct disorder, and violence, while three of the papers were concerned with exposure and its effects on mental health, another three were concerned with coping responses, identity development, and exposure and its impact on drug use. The heterogeneity of the papers and varying results emanating from the studies reviewed, were better summarised into two themes, internalising and externalising behaviours: Internalising behaviour was taken to represent and included measurements of mental health concerns such as depression or anxiety, identity problems, and coping issues. While externalising behaviours included issues such as violence, conduct disorder, withdrawal, behaviour concerns, bullying, aggression, or drug use [28].
Seven of the papers reported results which were classified under the externalising behaviour theme. These included paper 1, 3, 4, 5, 7, 9 and 10 all the other papers 2, 6, 8, 11 and 12 were classified as internalising behaviours themes. Below is a table depicting some of the inconsistencies in the findings, see Appendix E for the entire Data extraction Matrix.

Externalising factors
Among the externalising themes, findings from paper 1 were that adolescents who were cruel to animals had significantly higher histories of exposure to domestic violence and were two to three times more likely to have been exposed to domestic violence. There were some limitations in this study in relation to measuring the variables which affected reliability of cause and effect and so intern affected the reliability of the study, results were presented therefore as probable cause.
Among the six papers looking at domestic violence exposure and effects on behaviour the findings were mixed, while paper 3 found exposure to parental violence to have no statistical significance associated with the likelihood of violent or nonviolent conduct disorder in adolescent boys they found conduct disorder to be higher among the males sampled. Limitations to the study resulted in conclusions being drawn that domestic violence effects cannot be commented on in relation to conduct disorder. In addition, paper 4 reported in their longitudinal study that boys who were exposed to domestic violence in childhood, reported less internalisation problems in adolescence, and concluded that early exposure in the absence of continued exposure had little effect.
Meanwhile, paper 7 found that boys who witnessed interparental violence were 2.5 times more likely to be at risk of becoming a victim of bullying and were less likely to become bullies as a result of witnessing inter-parental violence than girls.
In contrast, paper 9 reported that more males committed violent crimes, and reported more drug use as well as frequency of drug use but were more likely to use drugs if their peers did the same. Alternatively, paper 10 found that exposure to domestic violence had less impact on development of aggressive behaviours in males than females. The authors suggest that this may be as a result of the paper focusing on measurements during adolescent stages rather than in childhood using longitudinal studies as in many of the other studies. It is worth noting that aggressive behaviour in this particular study referred to getting angry and fighting.

Internalising behaviours
In the papers measuring witnessing domestic violence and its relationship to adolescent mental health symptoms, both paper 2 and paper 11 found that males reported and exhibited less depressive symptoms and accessed less professional mental health services as a result of exposure to domestic violence than females. Paper 2 reported this outcome even though there were differences in the prevalence level for both genders. In terms of other internalising behaviour such as those in paper 5, it was found that domestic violence exposure was significantly related to delinquency, depression, withdrawn and marginal to anxious behaviours more so in females than in males and more so among exposed adolescents in comparison to those who were not exposed to domestic violence.
In relation to identity development, paper 8 found all adolescents to have encountered identity problems, when exposed to domestic violence and this in turn influenced problematic gender role identity development, and the development of selfconcept, as well as having negative effects on psychological wellbeing, there were no gender differences found in relation to this concept.
Paper 12 is the only qualitative study included in this systematic review, which focussed on coping responses to domestic violence exposure. It found that all the adolescents reported varying degrees of effects which included, moving from place to place, feelings of depression, suicidal tendencies, feeling of inequality, ambivalence, anxiety, sadness, and poverty. In addition drug and involvement in crime was also reported to have resulted from their exposure.
Finally, we have presented the results of the search strategy, the results of the paper selection process, the characteristics of the papers and data extraction from the papers. In addition, the results of the quality assessment of the included papers and a summary of the findings of all the twelve papers included have been presented. The author will now present an overall discussion and analysis of the systematic review undertaken and then conclude the study.

Analysis of the literature in relation to the systematic review question
In terms of the quality and critical appraisal of the included studies, it could be argued that the papers were mainly of high quality according to the findings of the assessment tools applied. Though the Likert scale tool applied could be considered a subjective measure as outlined by Cardwell, it allowed a numerical value to be assigned to each paper allowing easy comparison. These authors however considers the additional application of the Cardwell et al. critical appraisal tool to have been a particular strength to this critical review process and found it to have revealed a number of methodological concerns among some of the included studies [36]. It was obvious that there were data collection concerns and therefore subsequent implications for validity and reliability of findings among some of the studies. For example, in Duncan et al. study which assessed relationship between exposure to domestic violence and animal cruelty, data was collected from inpatient records rather than from interviews which was acknowledged as a gross way of assessing data which may affect reliability and variability. However, the study concluded that adolescent males who displayed cruelty to animals were 2 to 3 times more likely to have been exposed to domestic violence

Synthesis of the findings
Among the twelve papers reviewed, it was found that the findings would be better discussed in themes due to the heterogeneity of the papers as discussed earlier. . This has been supported by Duncan who reports adolescent males who were cruel to animals were two to three times more likely to have been exposed to domestic violence [51].
In addition Cavette and Oruze found that in general, exposure to domestic violence contributed more to the development of aggressive behaviours in females than in males, and witnessing domestic violence had an impact on aggressive behaviour development, both proactively and reactively in both genders.
Moreover, the main findings in the study by Mustanoja, et al. was that males who witnessed inter-parental violence had a 2.5 fold risk of becoming victims of bullying during adolescence compared to those who had not witnessed domestic violence [55]. Furthermore Fagan and Wright reported that more males committed violent crimes and violent offences, they were less likely to be exposed to severe violence between parents and were in fact less consistent in reporting witnessing domestic violence than females. In addition they were found to have more reported drug use with increased frequency than females, and had higher levels of delinquency, delinquent friends and were more likely to be influenced by friends taking drugs to do the same [52]. The influences of peer group are thought to be more important in adolescent years than that of parental influences. In a more general sense Idemudia and Makhubela results showed adolescence encountered identity problems when exposed to inter-parental domestic violence, which in turn influences problematic gender roles, gender identity development and the development of self-concept. Domestic violence exposure was reported to negatively affect psychological wellbeing [59]. Additionally the study by Aymer found adolescent males who were exposed to domestic violence at home, also experienced more poverty, and had increased alcohol and drugs use, and moved from place to place more frequently. This was reportedly due to antisocial behaviours and family exclusion from society. In addition they experienced more feelings of depression, suicidal ideations, ambivalence, anxiety, sadness, and being unequal [60].
This systematic review highlighted that there were a number of disparate ways in which domestic violence impacted adolescent males. This ranged from animal cruelty, drug taking, violence and aggression, through to feelings of depression, suicidal ideations, and feelings of sadness and ambivalence. However while there appears to be no common agreement as to how it impacts on these individuals, what the evidence suggests is that there is consensus that this vulnerable group of individuals are negatively impacted by exposure to domestic violence.

The strengths and weaknesses of the review
The potential strength or weaknesses of a systematic review lies in the transparency of each phase of the synthesis process, allowing the reader to focus on the merits of each decision made in compiling the information (Garg, 2008) However in utilising the framework of Caldwell these authors aimed to archive in-depth analysis and increase precision of the result by using a meaningful, but relatively easily understood method of analysing the data emanating from each of the papers used in the review [36].
These researchers acknowledge the value of collecting data from key people involved in the field of domestic violence but found the time limitations to produce this systematic review restrictive and so acknowledges this as a weakness to the study.
A further weakness to the study was the inclusion of papers written in English language only, which excludes other primary papers written in other languages and so potentially presents bias findings. Alternatively, it may be argued that language may not be of great importance and the selection of papers from a number of different countries gives a representation of different cultural perspectives on domestic violence. Similarly, while it may be argued that since none of the studies included in this research were of British origin, one would argue that the findings could also be applicable in developing training and education for practitioners working with similar issues in Britain, given that westernised adolescent males will be of similar developmental stages [63].
In relation to the methodological flaws of some of the papers included, these authors acknowledge that a significant number of these papers contained results that were not wholly generalizable due to convenience sampling and data collection issues such as self-reporting which can be possibly inaccurate or bias and therefore had implications for the findings of the paper.
In addition the review highlighted considerable heterogeneity among the different studies which has contributed to the lack of consensus in relation to the prevalence of exposure. While this may demonstrate a lack of conclusive findings to some extent, it does not indicate a weakness, as the emerging outcome is based on a robust process of analysis.
Taking all these issues into consideration, these authors propose that the presentation of the findings of this research process will advance the knowledge towards increasing the current overall awareness and understanding in relation to the impact of domestic violence on adolescent males.

Implications of the findings for practice, research and recommendations
The results of this review have several implications for professionals who work with adolescents. Firstly, the results suggest there is a relationship between the exposure to domestic violence and social outcomes for adolescent males, even though there is disagreement of what the impacts are. However, for policy makers, practitioners or other health professionals interested in addressing the prevalence of the effects of domestic violence on adolescent males, the reviewed results suggests several approaches would have to be adopted due to the diverse social effects on adolescent males. As it is well known, that there is a relationship between domestic violence and social deprivation, Radford et al. and Matczak macroeconomic policies therefore must address the root causes of domestic violence in society. From a wider policy perspective it is thought that there needs to be greater education via the media to reach out to all levels of society that domestic violence is unacceptable and could have significant negative consequences on those exposed to domestic violence, and on the identity development in children and adolescence. Similarly, practice educators are an important resource contributing to policy formation and subsequent embedding of policy initiatives. Practice educators should ensure greater awareness of established protocols of information sharing, consent and confidentiality. In addition, they must ensure that all education programs include consideration of local response pathways for adolescents exposed to all levels of risk, including clearly defined processes for responding to disclosure [64,65].
Secondly, the results of this review also suggest potential targets that could maximally prevent the impact on vulnerable adolescents may be early detection and intervention. Professionals such as teachers, social workers, those employed in the police and criminal justice system and health care professionals must all be equipped to identify, intervene and interrupt at any stage where evidence suggest that an adolescent is at risk of suffering from the effect of domestic violence.
Many of the research studies suggest that the current approaches in many ways can be regarded as reactionary, as they proposed interventions such as psychological therapy Hindin and Gultiano, or parenting classes, Fagan and Wright as they seek to treat the problem after it has occurred. As such, future interventions may need to adopt prophylactic approaches aiming to address the issue of domestic violence and its causes from a preventative perspective or at the embryonic stage before it impacts on these vulnerable individuals [52,53].
As we have already established some of the causes and consequences of domestic violence, it is envisaged that future research should now focus on what are the most effective approaches in addressing the issue, taking into consideration that this age group is quite likely to be media savvy, and the media may well be a mediating factor. Secondly some of the papers also indicated that there was a relationship between exposure and impact on adolescents. Therefore future research also needs to identify what length of time, and what frequency these occurrences take before having an impact, as well as which adolescent males may be more likely to be impacted. It is thought that this evidence may assist professionals with timely intervention strategies

Conclusion
The aims of this review was to systematically identify studies related to the impact of exposure to domestic violence on adolescent males and to describe the key features of these studies as it relates to this issue. In particular the exposure to domestic violence was explored across the domain of prevalence and consequences on social wellbeing as it relates to adolescent males. From the findings it is expected that recommendations for practice education can be made.
The effects of domestic violence in relation to women and children has been well documented, however the information relating to adolescent males has not yet been well researched, and is at best sparse. Through this systematic review the authors aimed to add to this body of knowledge. Thus, we recommend that the issue of the impact of domestic violence on adolescent males is not just an issue for professionals, but should involve a wide array of stakeholders as its impact is likely to have far reaching consequences for a wide cross section of society.
Practice educators and practitioners therefore could play a critical role in networking with NGOs and voluntary organisations in the development of social support networks to support victims of the exposure to domestic violence. Similarly, there is a potential for health and social care practitioners to undertake research, involving these victims to explore possible ideological and cultural factors which engender and perpetuate the phenomenon of domestic violence and resultant exposure to adolescents. It is envisaged that such research studies can provide the necessary background for policy-making, planning, and implementation of new policy initiatives in terms of addressing the impact on male adolescents In conclusion, each of the findings of this review study warrants replication due to the varying nature of the outcomes. Future studies with ethnically diverse samples will add to the generalizability of the present findings to adolescents groups in general. Future research might also consider investigating the role of the media and where professionals are best placed to address the consequences of the impact of domestic violence on adolescent males. It would also be interesting that future research may have quite different outcomes in non-western societies that focused more on the importance of the family who may regard adolescence as a transition period into independence from family and coming of age.   Domestic 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 4 5 6 y y y y y y y y y y y y n y y n n n y y y y y y y y y y y y y n y y y y y y y y y y y y y y y y n y y n n n y y y y y y y y y y y y y n y n y y