Abuse and Protection Issues across the Lifespan: Reviewing the Literature

The literature is reviewed: (a) comparing models of abuse and protection applied at each stage of the lifespan; and (b) exploring abuse and protection issues with respect to individuals over time. A paucity of comparative and lifespan work in the field of abuse and protection is reported. Within the available literature, different types of knowledge are drawn upon and contrasted definitions of ‘abuse’ and ‘protection’ employed. Accordingly, the most significant findings of the review are in the dissonances and the gaps surrounding the substantive findings. Examples are presented to demonstrate this, while new directions for discussion and research are proposed.


Introduction
This paper reports on a review of the literature which: (a) compares and contrasts abuse and protection issues as they apply to individuals at different stages of the lifespan; and (b) examines continuities and changes in abuse and protection issues, with respect to an individual over time. The main impetus for the review was a research project into interagency adult protection practice in Scotland conducted by the authors (Hogg et al., 2009). Whilst this research adapted methodology from an earlier study of child protection (Daniel, 2003), no explicit comparison between the two studies was intended. In practice, however, we and our respondents frequently drew on comparisons with other contexts to help us interrogate processes and concepts. For instance, does the victim's impairment make this case distinctive from other instances where men abuse their female partners? Would/should professionals have acted differently if the victim was a child? We wished to explore the extent to which such comparisons hold.
'Abuse' and 'protection' have been constructed differently, at different historical periods and within different policy and theoretical frameworks, as they are applied to different groups. The concept of child abuse, and of child protection as a state activity, has had currency for some decades. The 'social problem' (Blumer, 1971) of abuse within adult intimate relationships rose to prominence more recently, and has been labelled in various ways: for instance, 'wife battering', 'partner abuse', and 'domestic violence'. This neatly reflects conflicted and/or evolving understandings of the centrality of gender, type of (potential) harm inflicted, marital status and occurrence in the home to the definition of this concept. More recently still, 'elder abuse' in family settings began to gain recognition as a social problem. Whilst child, domestic and elder abuse continued to be largely separately researched, the arguments for comparing findings, primarily from a 'family violence' perspective, gathered pace in North America in the 1980s (e.g. Gelles, 1987) and in the UK a decade later (Kingston and Penhale, 1995). In non-family residential settings, concerns about abuse of residents have existed since the 1960s, giving impetus to the development of care in the community. Instances have continued to emerge and galvanise debates about adult and child protection since that time (e.g. Levy and Kahan, 1991;Buckinghamshire County Council, 1998). Such debates have linked abuse to various extents with the nature of institutions themselves. Meanwhile, mistreatment of people with particular support needs in community settings have informed discussions about the heightened 'vulnerability' of certain individuals, groups and/or situations (e.g. Social Work Services Inspectorate, 2004). Within these contexts, certain types of abuse have been more strongly associated with certain groups of individuals: for instance, adults with learning disabilities and sexual abuse (Brown and Turk, 1992;ARC/NAPSAC, 1996;Craft, 1996).
Current UK policies proceed on the basis that adults who are considered particularly vulnerable, for instance due to frailty or cognitive impairment, who are of any age and at risk of abuse in any setting, constitute a meaningfully distinctive category. This is not to imply that individuals do not move in and out of this category, but it is to imply that a common protective framework is appropriate. There is a separate policy response to domestic abuse, defined in this paper and in a range of UK contexts as the emotional, financial, physical or sexual abuse of an adult (usually a woman) by her current or former partner (usually a man). The policy response to child abuse is different again -though with some inconsistencies in laying down where 'childhood' finishes and 'adulthood' begins. We wished to explore the extent to which these assumed commonalities and distinctions between groups in social policy are grounded in theory and research.
Our literature search identified a body of work that we refer to here as the lifespan/comparative literature. However, this is not a comprehensive, nor a unified body of work: indeed, this is the primary finding of the literature review. In this paper, we demonstrate how we have structured the literature we found and give some examples of what it has to say. Our intention in doing so, in view of our primary finding, is to stimulate debate about where the gaps might lie.

Method
We searched a number of databases (CINAHL, IBSS, Medline, ASSIA, British Nursing Index and Archive, National Library for Health, BNB) to identify papers concerned with 'abuse', 'protection' or their common synonyms in relation to the 'life course' or 'lifespan', and/or to two or more groups usually treated as distinct, e.g. adults and children. Internet search engines were also used. There were no exclusion criteria with respect to date or place of publication. However, inclusion was limited to publications in the English language.
In identifying literature of relevance, we were guided by the following questions: 1 How have models of abuse applied at each stage of the lifespan been compared? What justifications have been given for similarities and differences in models of abuse? 2 What is known about continuities and changes in an individual's risk of abuse across the lifespan? 3 How have models of protection applied at each stage of the lifespan been compared?
What justifications have been given for similarities and differences in models of protection? For the purposes of this review, we define a 'model of abuse' as a way of defining and/or explaining abuse. We define a 'model of protection' as a strategy underpinned by beliefs about appropriate and/or effective ways to prevent and/or respond to abuse. Crucially, any such search strategy operates as a series of filters ( Figure 1). The review did not set out to draw new comparisons between distinct bodies of work, or to construct a lifespan model. The results do not reflect one concrete, pre-existing phenomenon. Rather, inclusion in the review is dependent on authors having identified or constructed a phenomenon, having labelled it 'abuse', and having considered it across life stages or affected groups usually treated as distinct. Considering the subtleties of this route to inclusion is important to meaningful interpretations of the findings. So too is attention to the different types of knowledge represented. Empirical research, theoretical and opinion papers are all included: their contents are evidence-based to varying degrees and contested to varying degrees. When approaching the findings, we found it helpful to bear in mind the following questions: 1 What is the political, theoretical and evidence base of the comparisons and continuities which have been explored? 2 Have existing comparisons and/or continuities been sufficiently explored? 3 Are there other comparisons and/or continuities it would be useful to explore?
These questions can be applied separately to the literature organised under each of the headings as detailed below.

Results and discussion
We have organised the findings of the review thematically. Each theme addresses a factor which might affect or characterise a situation potentially defined as abusive, and which has been discussed in the lifespan/comparative literature. The themes fall into four broad divisions: cultural/societal factors, environmental factors, factors relating to others in close contact with the individual at risk and factors relating to the individual themselves. These divisions draw on ecological models of abuse, which locate 'vulnerability' in the interactions between individuals nested within a set of broader contexts (Bronfenbrenner, 1979;Belsky, 1980;Sobsey, 1994) (Figure 2). Whilst functional in terms of ordering our findings, we accept that this model is a partial representation of many situations; for instance, the individual at risk might also abuse others. Attention is drawn to Schiamberg and Gans (2000) for an innovative incorporation into the ecological model of abuse of two individuals' interlocking ecologies over time.
The lifespan/comparative literature points to important parallels and contrasts not only by age but also by disability and gender. These dimensions have been explored to varying degrees in the literature addressing each of the factors set out in Figure 2. Below we discuss the literature organised under one factor heading from each of the ecological levels, selected to exemplify features of the lifespan/comparative literature as outlined above. There is a particular attention to the treatment of age, disability and gender in our discussion. We also begin to consider, under each heading, whether there are other explorations to be made. Table 1 shows the contrasts and continuities which have been explored in the lifespan/comparative literature, which specifically relate to discriminatory social attitudes. We first review the contrasts and then the continuities. We then address the implications for future discussion and research.

Example 1: Cultural/societal factors -discriminator y social attitudes
Within a series of commentaries comparing contemporary understandings of elder abuse with those of child abuse, at a time when the former was gaining recognition as a manifestation of family violence, an underpinning societal devaluation was proposed to be one way in which it was distinct. For instance, Penhale (1993: 104) refers to western stereotypes of children as 'vulnerable' but older people as 'burdensome' and suggests this takes its toll on professional awareness and responses, mediated through statutory responsibilities and powers. Harm to older people within families has long been conceived of as a private rather than a community concern by society at large (Schene and Ward, 1988;Kohn, 2003). Despite the awareness-raising of organisations such as Action on Elder Abuse, and the increasing procedural if not statutory footing of adult protection in the various jurisdictions across the UK, this is arguably still the case to some extent today. Penhale and Kingston (1995) suggest that society's own creation of the necessary conditions for elder abuse to flourish (i.e. ageism) accounts for the absence of any 'moral panic' analogous to that to which child abuse gives rise.
Similarly, societal ageism has been argued to shape understandings of domestic abuse against older, as opposed to younger women. The propensity to view such circumstances through an 'elder abuse' as opposed to a 'domestic abuse' lens, and the frequent comparison of 'elder abuse' with 'child abuse' as opposed to with 'domestic abuse', is itself argued to evidence views which are patronising to older people and dangerously gender blind (Vinton, 1991;Penhale, 2003;Hightower et al., 2006). Feminists highlight oppressive results where male domestic tyranny in the context of older couples is not approached by professionals and others as such, but rather as an excusable response to 'carer' stress, in turn linked to female victims' own characteristics (e.g. Brandl, 2000; Straka and Montminy, 2006). Moreover, whilst younger women's voices were powerful in constructing 'domestic abuse' and the policy response, older women's voices have been marginalised both in this and in professionally dominated 'elder abuse' discourses (Whittaker, 1995;Harbison, 1999). Amongst the findings of qualitative studies attempting to address the imbalance is evidence of the way ageism can compound the experience of domestic abuse. Isolation by their partners, for instance, requires less effort than it might have done in former years, because of the myriad social factors already excluding older people from life outside the home (Seaver, 1996). Whilst Walker's (2000) three-phase cyclical model of escalating tension, acute violence and 'loving contrition' has received considerable support in work with younger women, older women experiencing longstanding abuse report that the contrition phase appears to wane with age (Seaver, 1996). Furthermore, older women may have been socialised into expectations of their submission to men and life-long commitment to marriage more rigidly than younger counterparts, impacting on perceptions of their options if abused (Aronson, 1995;Zink et al., 2003).
Comparisons of social attitudes as risk or protective factors in the context of residential care have taken the level of research activity and the numbers of inquiries into abuse in different sectors as broad indicators of the value placed on care recipients. For instance, Manthorpe and Stanley (1999) argue that the high number of inquiries in children's, as opposed to adults' institutions, probably reflects a distinction in public reaction when abuse has been disclosed, rather than higher rates of prevalence. Stevenson (1999) makes a similar point in relation to her comparison of levels of research activity in children's homes as opposed to homes for older people specifically; Manthorpe and Stanley (1999) link these trends to the stronger policy implications, where children have been harmed in situations into which the state has intervened to place them. By contrast, older people's residential care is increasingly depicted in the language of 'consumer choice', however false this might have proved in practice. Manthorpe and Stanley (1999) make further distinctions between adult care groups in terms of inferred public concern about abuse; they note in particular that abuse of adults in mental health services is less likely than abuse of any other institutionalised group to become the subject of formal inquiry.
Taking a broader overview, Craft (1996) has usefully highlighted the continuities between adult groups in residential settings, particularly older adults and those with learning disabilities, in terms of their societal devaluation. This might influence the attitudes of residential staff, and has certainly influenced the historical strength of public demand for safeguards. The same sense of continuity characterises texts on abuse and disability, in respect of societal discrimination as well as other factors. For instance, disabled people are frequently disadvantaged within mainstream services including women's refuges and child protection systems (Westcott, 1994). Persistent negative stereotypes may lower the perceived worth both of disabled children and of disabled adults in the minds of perpetrators (Sobsey, 1994). Similarly, patriarchy and gender-based oppression are presented by some as unifying features, more analytically meaningful than the division of abuse into types according to the victim's age (e.g. Whittaker, 1995). Two papers were also identified documenting elevated risk of abuse in bisexual and homosexual people across the life course in comparison with heterosexual people, hypothesised to be linked in part to perpetrators' discriminatory attitudes (Tjaden et al., 1999;Balsam et al., 2005).
At a basic level, it would be possible to identify other continuities and dichotomies to explore under this heading. For example, how does discrimination interact with models and experiences of abuse over time in the context of minority ethnicity? Are there differences, as opposed to commonalities, in the ways discrimination and abuse interact for disabled children and disabled adults? Perhaps, too, there is scope for more coherent theorising. Certainly, it would be naïve to hope for complete consensus about the types of discrimination that might be thought to underpin abuse, to what extent and in what ways. Three thoughtful chapters on psychological, sociological and feminist perspectives on family violence in Gelles and Loseke's (1993) edited book, for instance, provide some evidence of this, with much of the comment also transferable to extra-familial abuse (Gelles, 1993;O'Leary, 1993;Yllo, 1993). These theorists welcome debate however, as do others like them, and accept the potential for complementarity to significant degrees. It seems that more nuanced understandings of the intersections of multiple discriminatory structures could enhance our various models of abuse. For instance, the discourses of abuse and gender, on the one hand, and abuse and disability, on the other, seem ripe for more extensive dialogue. Dialogue is pre-existing in the specific case of disabled women (e.g. McCarthy, 1999;Copel, 2006); however, the envisaged learning has broader applicability than this. For instance, the ways each cluster of perspectives considers, represents and responds to abused individuals whose opportunities and/or capacities to make decisions seem restricted could be compared and potentially mutually refined, to the benefit of affected individuals of all abilities and either gender. Table 2 shows the contrasts and continuities which have been explored in the lifespan/comparative literature specifically related to isolation. These interrelate, and we discuss them together below. We then address the implications for future research and theorising.

Example 2: Environmental factors: isolation
Evidence from North American family violence research, which has used large samples and statistical methods, associates isolation from the broader community with elevated risk of all forms of family violence (e.g. Gelles, 1997). Isolation is also commonly reported as an outcome of abuse (Finkelhor and Pillemer, 1988;Penhale, 1993). These associations are drawn upon, alongside professional experience and extrapolation, in British academics' comparative discussions of familial elder and child abuse. Here, older people are considered at relatively greater risk of isolation as opposed to children, who are regularly seen by a range of professionals in the course of daily life (Penhale, 1993). Increasing isolation in later life might also distinguish older women's experiences of domestic abuse from those of younger women, as noted above. Additionally, qualitative researchers examining older women's experiences, such as Zink et al. (2003), report that fear of isolation and loss of social status can differentially influence older women against leaving abusive relationships.
Discussions of the risk of isolation from the outside world in residential care have followed a similar pattern to discussions of societal devaluation. Namely, isolation is considered a common risk factor (Craft, 1996;Stevenson, 1999), whilst finer distinctions between groups have also been proposed. For instance, Stevenson (1999) speculates on common barriers to children's and older people's relatives, respectively, monitoring standards within residential settings. Older people's relatives, for instance, might feel unsure about the lower boundaries of acceptable care quality, or the impact on their relative of a possible complaint. A more frequent tension in children's homes is with parents perceived by staff to be 'part of the problem': valuable inputs might therefore be ignored (Stevenson, 1999: 23). Manthorpe and Stanley (1999) speculate on a link between family involvement and the perceived 'deviance' associated with residence in certain institutions. These conceptions might colour staff attitudes to relatives and their views, as well as relatives' attitudes to their family member. A corollary would be that adults with learning disabilities and older people in care more often have family advocates than do children and adults with mental health problems. There is some evidence of this (Manthorpe and Stanley, 1999). Beyond their fostering of family links where possible, engagement with the community, and with professionals external to the home, are argued to be key protective factors against abuse across residential settings (Stevenson, 1999).
This subdivision of the lifespan/comparative literature exemplifies the different types of knowledge represented. The continuities and comparisons which have been made draw on informed professional opinion, views of some of those affected by abuse as expressed in qualitative studies, and a body of quantitative evidence whose tenets are contested (e.g. Whittaker, 1995). More research and inter-disciplinary discussion would appear to be required to draw maximal learning from such tentative comparisons. The similarities and interconnections across age and service user groups that have already emerged, however, as well as the linkages between ecological levels demonstrated in this paper, would seem to lend support to more joined-up theorising.

Example 3: Factors relating to others in close contact with the individual: legal and cultural underpinning of involvement
Example 3 draws on a particular contrast made explicit in the literature comparing elder and child abuse in family settings, but illuminating a large range of more implicit Table 3 Life course continuities and changes in the literature which relate to individuals' histories of abuse Continuities and changes (across the life course) • Revictimisation research • Elder abuse survivors' retrospective accounts of earlier abuse • Cycle of violence and disability links between discourses. Namely, different expectations about the involvement of the perpetrator with the victim have been highlighted in cases of alleged abuse by caregivers of older people as opposed to children. It has been proposed that definitions of abuse might be more clear-cut, because no cultural precedent exists in recent history for the physical chastisement of older people by their carers. Conversely, caregiving relationships between relatives and most older people are not underpinned by legal duties, as is the case for children (Korbin et al., 1989;Phillipson and Biggs, 1995); nor does contemporary Western culture lend equal weight to the duty of older people's relatives, as opposed to parents, to provide care and support. Similarly, there are fewer legal and cultural imperatives to guide decisions as to who should participate, once elder protection procedures are underway. Some adult co-residents are not carers; few older people's carers have rights and responsibilities to become or remain involved in the way that parents do. Furthermore, the possibility of separating some adults, in order to prevent abuse, is less ideologically fraught than separating children from their parents (Stevenson, 1996).
Whilst not explicitly considered elsewhere in the lifespan/comparative literature, consideration of cultural expectations about the relationship between particular victims and perpetrators, including preparedness to challenge these, is vital to any attempt to protect and/or empower. Oppressive constructions of disabled people's dependency on others, of the relations between ethnic groups, or of women's status in relation to men, for instance, have continuing potential to shape individual relationships, not least abusive ones. Again, we begin to see in this the connections between groups and/or life-stages more usually treated as distinct, as well as the interconnection of the levels of the ecological model.

Example 4: Factors relating to the individual -histor y of abuse
Example 4 concerns an individual's history of having been subject to abuse. The literature identified relates to continuities and changes across individual life courses, rather than contrasts between groups. Table 3 shows the bodies of work exploring these continuities and changes. We introduce the respective discourses and then discuss the possibilities for bringing them together.
The study of 'revictimisation' has generated a considerable, if disjointed, literature, particularly in North America (Messman-Moore and Long, 2003). Revictimisation is the repeat victimisation of an individual who has been abused before, with research focusing on an observed association between prior victimisation and increased risk for subsequent victimisation. In practice, revictimisation has most often been operationalised as sexual victimisation of a woman who was sexually abused as a child. A seminal narrative review (Messman and Long, 1996) and subsequent meta-analytic review (Roodman and Clum, 2001) found positive evidence for the existence of the phenomenon at least as thus defined, despite methodological limitations and wide variations between studies. A number of factors have been proposed to impact on rates of revictimisation. These have been reviewed from a number of perspectives and grouped in various ways into theoretical models (Messman and Long, 1996;Grauerholz, 2000;Breitenbecher, 2001;Roodman and Clum, 2001;Arata, 2002;Messman-Moore and Long, 2003;Classen et al., 2005;Marx et al., 2005). Most of the factors discussed, and certainly those empirically investigated to any significant degree, focus on the psychological effects of child sexual abuse on women's later functioning: for instance the hypothesised effects on attachment style, substance use, sexual behaviour and recognition of risk (Grauerholz, 2000). The evidence with respect to any one factor is very often very mixed (Breitenbecher, 2001); Messman-Moore and Long (2003) therefore propose a shift towards examination of the broader function of these factors, that is the psychological vulnerability and/or increased contact with potential perpetrators which have different causal pathways in different individuals. Grauerholz's (2000) model of revictimisation is arguably most helpful in terms of future research (Arata, 2002;Messman-Moore and Long, 2003). She has organised existing knowledge and new hypotheses into four ecological levels: the former falling largely into the inner levels and the latter into the outer. For instance, at the cultural level she proposes a tendency to blame the victims of sexual abuse may result in lower self-esteem in abuse survivors than others, whilst simultaneously disinhibiting potential new perpetrators who may devalue the survivor accordingly if aware of past abuse.
Study samples in revictimisation research have predominantly comprised young, white, middle-class North American women, with no recorded disabilities or mental health problems (Messman-Moore and Long, 2003). This limits the scope for findings to be generalised. The few quantitative studies including broader age-ranges of participants have found higher rates of revictimisation than those with younger samples (Roodman and Clum, 2001). Isolated studies have found support for the existence of revictimisation amongst adults of both sexes (Desai et al., 2002), homosexual people (Lie et al., 1991;Heidt et al., 2005), women with mental health problems (Darves-Bornoz et al., 1995) and women from minority ethnic groups (Urquiza and Goodlin-Jones, 1994). However, wider-scale, more inclusive research is clearly overdue. Moreover, some commentators have moved to examine the linked effects on future vulnerability of a broader range of traumas than 'abuse' alone (Classen et al., 2005;Marx et al., 2005).
Qualitative research from outside of the US has examined the earlier abusive experiences of people who also came to be abused as older people. Pritchard (2001a) interviewed female survivors of elder abuse, and found that past experience of domestic abuse and/or child abuse was common amongst this group. In a second study, she interviewed male elder abuse survivors, finding some similarities in types and life-time patterns of victimisation (Pritchard, 2001b). Similarly, Walsh et al.'s (2007) interviews with older Canadian people uncovered lifespan patterns of victimisation amongst some of the respondents; they also commonly voiced support for the theory of revictimisation as it applied to their generation.
Interestingly, a link between prior victimisation and increased risk of future victimisation has also been drawn by researchers into abuse and disability, though largely independently of other revictimisation research, and never isolated to the same degree from the range of other factors conceptualised as increasing risk. Nonetheless, Sobsey's (1994) cycle of violence and disability, informed and supported by his wide-ranging review of empirical research, illustrates a number of mechanisms by which initial violence creates heightened risks of future violence in this specific context. For instance, violence in the home might lead to removal of disabled children to institutional settings, which have historically been also far from safe. Exposure to persistent violence might lower self-esteem and self-protective skills. Most fundamentally, violence can cause disability, and disability increases the risk of abuse, through a number of mechanisms operating at every level of the ecological framework.
With respect to these three bodies of work, and the potential links between them, we wish to pose some tentative questions. On the broadest of levels, what can each teach the other? In particular, the theme of competing emphases on individual versus social contributors to risk of abuse recurs here, and an exchange of ideas about this might be fruitful. It has been argued that the revictimisation discourse has focused too greatly on individualistic factors in attempting to explain re-abuse (Grauerholz, 2000). By way of contrast, the UK 'adults at risk'/'vulnerable adults' discourse might be argued to focus too greatly on inherent and/or socially constructed 'vulnerability' underpinning sometimes multiple abuses, at the expense of also examining the effects of individual experiences over time. The UK child protection literature is well advanced in its examination of mechanisms by which adverse circumstances and negative experiences need not lead to more negative experiences in the future. This discourse of resilience is expanding to encompass a lifespan perspective, which might also be usefully connected to the three bodies of work reviewed above.

Conclusion
This paper has presented a review of the literature exploring contrasts and continuities between models of abuse and protection, between groups of people and across the life course. It has used four examples, one from each level of an ecological model of abuse and protection, to demonstrate that useful comparisons have been drawn, but that there are many more to explore, or to explore more fully. The subtlety of these interconnections appears not to be matched by current models of protection, which are fragmented across the lifespan according to divisions imposed by convention, social policy and/or academic convenience. The review found no evidence that similarities and differences sometimes assumed in professional, academic and policy discourses have been informed by comparative or lifespan research.