« Previous | Contents | Next »
Listen
CHAPTER FOUR: RISK ASSESSMENT IN CHILD PROTECTION
INTRODUCTION
Child protection work has become much more reactive and less optimistic since the 1970s when the death of Maria Colwell in 1973 resulted in a moral panic and preoccupation with culpability, blame and retribution (Parton, 1996). Child abuse became like a disease which can be predicted, diagnosed and treated, without any acknowledgement of the social context. There was no question as to how risk was constructed by professionals, only that it was present. In child protection work in particular, the main aim of risk assessment now is to gather forensic evidence, prioritise cases and predict risk. There is thus a recent split in child welfare work internationally, between child protection and family support:
- Child protection - a neo-liberal approach to private issues as public issues and the adoption of a focus on risk. Kemshall (2002: 113) describes neo-liberalism as 'privatisation, de-regulation and marketisation of the state sector'. This type of approach is mainly adopted in the UK, North America and Australia.
- Family support - proactive intervention with children and families through health and educational services primarily, giving primacy to the family and a less specialist role for child welfare professionals. It respects family diversity and family values but it fails to explore power inequalities and patriarchal models of control (especially in relation to child sexual abuse by adults). This type of approach is mainly adopted in continental Europe.
Trotter et al. (2001: 9) give a useful illustration of the difference between the two approaches:
"The difference between the European and British systems is well illustrated by the common characterisation of the first child protection visit in which the British child protection worker comments: 'I am here to investigate a report of suspected abuse against your child'. The European child protection worker comments on the other hand: 'I am here to see if I can help you with any problems you might have with your child'."
Much of the literature on child protection focuses on the forensic nature of child protection procedures, namely risk assessment and management, the centrality of evidence gathering and surveillance of perpetrators, and this chapter explores this forensic approach in relation to legislation, guidance and tools. It compares and contrasts different models of practice and explores the rights of families in child protection and the tensions between the various professionals involved.
LEGISLATION, GUIDANCE AND TOOLS IN CHILD PROTECTION
The Children Act (1989) and the Children (Scotland) Act 1995 both aimed to balance the protection of children with the rights of parents and to limit compulsory intervention to a minimum. Nevertheless, further compulsory intervention orders were developed in Scotland, namely, the Child Protection Order which allows for the detention or retention of a child in a place of safety for up to 8 days; the assessment order, which allows for an assessment of a child for up to 7 days; and an exclusion order, which allows for the removal of an alleged abuser from a household (instead of the child being removed) for up to six months. McGhee and Francis (2003) note that even with the climate of managerialism and accountability (albeit greater in the criminal justice and community care fields than in child protection currently), these orders have been little used in child protection cases in the UK to date. However, these authors infer that there has been a greater involvement of legal services within local authorities following the implementation of the Children's Acts north and south of the border, which may have stifled professional discretion to intervene where deemed appropriate. However, Francis et al. (2006) also suggests that the minimum intervention principle now operating in most areas of social work is having a significant influence on the raising of child protection thresholds. For example, social workers are more willing to suggest that removing a child may result in greater harm than leaving the family intact and they try to find alternative methods of containing or resolving problems within the home environment.
Whilst intervention may not have become more intrusive following new legislation, risk assessments per se have been encouraged in recent years. An initial evaluation of the Department of Health's Assessment Framework was published three years following its instigation in 2000 ( DOH, 2003). The original aims of the framework were to make practice across departments more consistent, for assessments to be more child-friendly and family-oriented and for assessments to inform the future work with the child/family. However, it has been noted that there are over 750,000 permutations of significant risk factors within this framework, thus making predictability less accurate (Edinburgh City Council, 2004). The DOH evaluation found that parental involvement in the process of assessment had increased substantially but not so much for children and young people themselves. Social workers also felt that the increased participation of family members enhanced the quality of the assessment but many expressed concerns about their ability to undertake a good assessment in terms of information gathering and inter-agency working. Although their workload was seen to have increased because of the assessment process, they were nevertheless likely to spend more time working directly with children and families, and collaborating with other agencies, as a result of this process.
In a more recent critique of the DOH Framework, Millar and Corby (2006) suggest that the structure of the framework allowed for greater accountability to, and communication with, service users, which parents and other family members appreciated. These authors also argue that the prescriptive nature of the framework, whilst criticised in earlier evaluations, can indeed be helpful in establishing dialogue and credibility between worker and the family. Nevertheless, Booth et al. (2006) maintain that the timeframes given for the DOH assessments are too tight and restrain the building of such dialogue and credibility. However, they stress the need for adequate resourcing to address the needs of vulnerable people.
Of 29 local authorities in Scotland that Francis et al. (2006) included in their evaluation of risk assessment in relation to child protection orders, only five used the DOH's Framework for Assessing Children in Need. Although respondents were aware of several alternative models of risk assessment, these were seldom used in any systematic way because they were deemed not comprehensive enough to be used in all situations (pre-birth risk; addiction of parent; Children's Hearings reviews, etc), resulting in many social workers resorting to pragmatic professional judgement, willingly or otherwise. Equally, less than one third of their sampled authorities issued specific guidance on assessing 'significant harm', not least perhaps because it was rarely defined. However, there was some ambivalence about whether it should be rigidly defined, which may preclude professional judgement and instinct (at least amongst the more experienced social workers). Most agreed that a broad definition would be helpful without being overly prescriptive or stifling of professional discretion.
So as to enable greater conformity and consistency across Scotland, the Scottish Executive (2005c) proposed a programme of change - Getting it right for every child, incorporating the development work undertaken on an 'Integrated Assessment, Planning and Recording Framework' ( IAF). Getting it right for every child concentrates on planning and action to address children's needs. Integrated assessment should be undertaken only where the child's circumstances require it and should include the views of all stakeholders (including children and families). Assessments apply in a holistic fashion to all children and 'travel' with them between authorities and between agencies. 'Getting it right for every child' - which was subject to consultation in 2005 - outlines a programme which would give authority and responsibility to agencies and professionals to work with children, families, local communities and partner agencies; to take action which is appropriate, proportionate and timely and which focuses on improving outcomes for children and young people.
The development work for the Integrated Assessment Planning and Recording Framework was piloted in two areas of Scotland (Gibson et al., 2006) but was found to result in poor assessment record keeping and action planning, and a failure by some staff to follow up on potential risks to children. This was deemed to result from a lack of integration and analysis of varying agency files. Although the researchers had concerns about inter-agency working at the assessment stage, parents were more positive about both the assessment processes and the resultant additional services put in place to support them (Gibson et al., 2006).
Legislation is also currently being developed (the Children's Services (Scotland) Bill 3) which requires inter-agency cooperation and information sharing and streamlines the Children's Hearings system to focus only on cases which require compulsory measures of care/control in cases where significant need is identified (including patterns of behaviour which give cause for concern). The bill itself would place duties on agencies to work collaboratively to support children and to make clear plans for children with complex needs; to ensure that the views of children and their families are taken into account; and to refer to the Children's Hearings system only where compulsory measures were considered necessary to help the child.
In terms of risk assessment tools in child protection, actuarial models of risk assessment are criticised by many writers, but Trotter et al (2001) note that mainland European child protection systems tend to be less hierarchical and forensic in practice, thus offering more autonomy for practitioners to use professional judgement in risk assessment and management in comparison to the UK, North American and Australian child protection systems. In New Zealand, the Risk Estimation System ( RES) developed in the 1990s, is a nationally-adopted, computer-based, consensual model of identifying areas of severity of abuse, vulnerability of children and likelihood of further abuse. However, Stanley (2005: 80) argues that '[r]isk is slippery and can escape the disciplines of positivistic risk assessment technologies'. He goes on to suggest that 'the calls for more generalised and common frameworks do little to establish deeper analytical understandings of how definitions and assessments of risk are formulated' (ibid: 86). Munro (2004: 881) also notes that numbers in actuarial calculations 'have an air of authority and objectivity that can mislead people into crediting them with more accuracy than they deserve'.
Warner suggests that in the UK, there is limited knowledge of known risk factors in relation to child protection work, no theoretical basis for assessments and a failure to revise initial assessments. In Australia, Warner suggests that risk assessment tools are used more pragmatically, to cover the backs of professional organisations rather than to further the interests of client safety. The USA has a more formalised and widespread use of risk assessment tools than other developed countries although the evidence to suggest that these inform professional judgement is limited (Ryan et al., 2005). In some US states, only 50 per cent of substantiated cases of child abuse result in a service being offered, even though child welfare is 'big business' in the USA and often privatised and unregulated (de Panfilis & Zuravin, 2001). At least 42 US states use at least one type (actuarial, clinical or combined) of risk assessment tool in child welfare decision making (Loman & Siegel, 2004). One such actuarial tool is the Family Risk Assessment ( FRA), which encourages accurate and consistent predictions of family risk of future child abuse and neglect in order to enable consistent decision-making about further work with families. Although it proved to have higher than average inter-rater reliability, its evaluation (Loman & Siegel, 2004) suggested that it could only accurately classify risk in two out of three cases and should therefore not be used in isolation. These authors also noted that misclassification arose more in families with low to moderate risk factors who were less likely to receive a service (especially in urban areas) as a result of such a classification. Other problems with this tool were that it could not readily differentiate recent past from distant past incidents of abuse, mitigating (protective) factors which may offset the risk factors, and incidents of mental health problems.
VARIATIONS IN PRACTICE IN CHILD PROTECTION
Francis et al. (2006) suggest that inter-agency cooperation varies between different agencies in different geographical areas. Equally, differing agencies have different values, cultures and interpretations and language about risk; levels of trust and confidence in other agencies may thus be low. These authors particularly singled out the divide between social work and education in Scotland in relation to child protection training and procedures, where they described these variations in practice as 'deep-seated differences in professional values, culture, language and attitudes' (ibid: 9). They find, amongst others, that thresholds of risk vary not only across agencies but also within agencies, notably with older, more experienced workers operating at a higher threshold of risk than their recently trained counterparts. Likewise, Gold et al. (2001) found that the age and experience of the workers resulted in differing assessment outcomes. A further study (Brown and White, 2006) assessed the literature on integrated working in children's services and similarly found cultural differences between professionals. They also suggested that integrated working demonstrated limited evidence of improving outcomes for children and families.
In their comparative study of risk assessment practice in Canada and Israel, Gold et al. (2001) found that whereas Canada prioritises the rights of children over the centrality of the family, Israel still prioritises the rights of the family to care for their own children. As a result of these differing philosophies, assessments and interventions reflected the professional, political and social mood in both countries. In terms of children's versus family rights, in the UK the paramount interest is the child's rights, not least because it is possible that involving the family in a risk assessment may create further harm or anxiety for the child (Marshall, 2006; pers. comm.). Hayes and Houston (in press) argue that in decision making in child protection within a family group conference arena, there may be power struggles between the child and the family and between the family and the professionals. However, these authors argue that such an arena tends to be seen favourably by both professionals and families and challenges the often risk-averse nature of child protection work. A further Canadian study of family group conferencing found that bringing the various parties together served to 'interrupt' thinking on how to reduce family abuse and engendered a strategy for the future and a commitment to the programme of work (Pennell and Burford, 2006: 126). This review did not, however, uncover any specific literature relating to risk assessment or management in family group conferencing and the primacy given to the rights of children versus their families. It is ironic, perhaps, given that New Zealand is the birthplace of family group conferencing, that Stanley suggests New Zealand is moving away from a family-oriented approach to one of child protection, albeit currently combining 'family-centred decision-making with a forensic eye on child protection' (2005: 38). The move towards a child protection approach in New Zealand came with the introduction in 1992 of the Paramountcy Principle, which privileges child protection in situations where the rights of parents are in conflict with the rights of the child. Social workers are mandated, through the Children, Young Persons and Their Families Act 1989 to remove children at 'serious risk of harm'.
Horwath (2005) notes that the role of the social work team in determining the focus of an assessment, the definition of neglect used and the subjective factors within individual social workers may all impact on assessments. She also states that while the Irish do not have a child protection register, they do have a child protection notification system and she found three different approaches to risk assessment in her sample of social workers in Eire:
- assessments which merely confirm that abuse took place;
- assessments which confirm whether the child has or will suffer harm;
- assessments which confirm not only potential harm but also the impact of that harm on the child's future well-being.
In a vignette exercise undertaken by Horwarth's sample of social workers, 50 per cent of respondents focused on the present crisis (abuse), 22 per cent considered the context and the pragmatics of the situation; and 18 per cent considered the child's wishes and longer-term well-being (Horwath, 2005). Horwath found that over half of the actual cases in her sample of child neglect incidents were closed because the neglect was unconfirmed or carers/agencies did not respond to requests for information. However, this study took place before a framework for assessment was developed in Eire. Howarth also notes that child neglect is more prevalent than other forms of child abuse in the western world but research is limited, mainly to the USA and UK.
Taylor and Meux (1997) also suggest that the identification of risk is influenced by the occupation to which professional respondents belong. Britner and Mossler (2002) argue that good child protection practice should encompass the rights of the child, the integrity and rights of the family and the duties and powers of the state. Different professionals represent those three interest groups and yet little research to date has explored the differences between professionals in the factors which influence their decision making (Stanley, 2005). Britner and Mossler (2002) found different professionals assessed risk not according to the characteristics or circumstances of the client but of the organisation within which they worked. Their study found that the different professionals involved in the same cases placed different emphases on factors influencing the risk assessment process and outcome. Such findings are not conducive to a common language of risk, as will be explored further in Chapter 5.
DISCUSSION OF RISK ASSESSMENT IN CHILD PROTECTION
The focus in the child protection field is on service responses to children at risk rather than the child's own developmental needs. Equally, because definitions of risk are unclear, different professionals have different perceptions of risk level and severity, and given the call for inter-agency cooperation, this is a worrying fact, as Little et al. (2004: 106) point out:
"When one professional talks to another about a child at risk, there is likely to be some misunderstanding and in the worst-case scenario they will be talking completely at cross-purposes."
In an international review of child protection literature over the preceding five years or so, Trotter et al. (2001) identified a key theme of a focus on procedures and forensic evidence in child protection, especially in the USA, UK and Australia. One consequence of this was the criticism of child protection services as ineffective in providing 'genuine assistance to families' (ibid: 7). Although there is much literature on the pros and cons of different approaches to risk assessment, there is a dearth of empirical material on the effectiveness of risk assessment tools versus professional judgement. Parton (1997) suggests that child welfare practice has been 'colonised by juridical priorities' of risk and harm (quoted in Trotter et al. 2001: 8). Another prevalent theme in this literature review was that child protection systems are net widening (e.g., drawing more children into the system because of lower thresholds of risk), when many children and families need more welfare than protection interventions.
' Messages from Research' ( DOH, 1995) argues for a family support rather than forensic approach and for better relationship building between worker and client. However, Bell (1999) argues that there is a conflict of interest between partnerships of worker and client and the dual task of assessing and managing risk within families. Equally, Trotter et al. (2001) suggest that partnership and risk assessment are not obviously compatible concepts as they produce conflicts of interest and rights. Equally, partnership rarely applies to worker and client but more to inter-agency working.
European models are deemed better in terms of giving workers greater autonomy and responsibility, without recourse to legal measures. Much of the literature acknowledges the better outcomes in reducing child abuse if welfare needs are paramount. Time spent on risk assessment could therefore be more gainfully employed on problem solving with families. Britner and Mossler (2002) suggest that many studies have found an inverted correlation between the need to remove children from their families and the level and intensity of service provision to that family following removal. It is as though the removal of the child means the removal of the problem.
SUMMARY
This chapter has examined the literature on risk assessment in child protection, focusing almost exclusively on the literature on child abuse and neglect. As with the other two fields, risk assessment in child protection is seen by practitioners and commentators as focusing too much on the process and not enough on the outcome of assessing risk. This results in 'bifurcation' (prioritising high risk groups at the expense of other needs/groups), being risk averse (to avoid litigation), intervening to make defensible decisions and reducing the professional autonomy of workers. Legislation in this field, more so than in community care, takes into account the rights of families as well as the rights of the child and often this can cause tensions in decision making. As in community care, there is also the issue of differing agencies having differing priorities and it seems from the literature that social workers tend to defer to the decisions of the judiciary and the medical profession in relation to risk assessment in child protection. There is also coverage in the literature of differing approaches to children's versus families' rights between the UK, North America and Australia compared with mainland Europe for example, where the latter see intervention to bolster family cohesion as the priority rather than the removal of the child to a place of safety.
Risk assessment tools in child protection are criticised for being overly actuarial and time-consuming to complete, and little is known about the related risk factors. Culpability in child protection is also focused on the family (rather than external factors such as poverty in terms of child neglect, for example) and/or on the practitioner, and there is little evidence of corporate responsibility for child protection issues. This often results in practitioners working defensively and applying objective and often compulsory measures to families rather than building trusting relationships with families.
In the following chapter, these broader issues of organisational structure and culture, which are prevalent in all three social work themes, are explored in greater depth.
« Previous | Contents | Next »