Response to Commentary on Taylor J, Baldwin N & Spencer N (2008) Predicting child abuse and neglect: ethical, theoretical and methodological challenges. Journal of Clinical Nursing17, 1193–1200 in Journal of Clinical Nursing18, 621–622


  • Julie Taylor PhD, RN, MSc, BSc(Hons), RNT,

  • Norma Baldwin BA, Dip. Soc. Admin, MPhil,

  • Nick Spencer MPhil, FRCP(UK), FRCPCH, DCH

We welcome the continuing debate about the difficulties inherent in providing effective care and protection for vulnerable children and the reflective response of Harbison to our article. This is particularly timely in view of another huge media focus on preventable child deaths.

We are entirely in agreement with Harbison about the need for rigorous, systematic work in assessing risk and need. Such work needs to be based on the most robust research evidence available and on the best practice models. She acknowledges the pressures to look for certainties where none may exist: we too consider this a crucial and damaging issue. However, it will not help the cause of child protection if we risk replacing one flawed system with another, just because we can think of nothing better.

Our paper is a critique of predictive methods and part of the current attempt across countries to improve effectiveness in assessing risk and need. Hard questions about the state of our knowledge and skills must be faced if progress is to be made. Other papers are available about the state of science – and art – in assessment and its centrality in good child care and protection practice (Daniel 2007, Department of Health 2004, Flaherty et al. 2002, Kolko & Swenson 2002).

We did not, however, invite health visitors simply‘to rely on their own reasoned assessments and judgements … using evidence and practice wisdom.’ We are clear that the activity of assessing risk is part of the role of health and social work professionals. We accept the possibility – even likelihood – of bias in undertaking such complex and sensitive work. We did ask that health visitors document their observations and the evidence for their judgements. We made a clear distinction however between prediction and assessment.

Far from intending to ‘obscure the necessity for making these judgements and, therefore, to obscure the processes by which they make them’, we like Harbison are committed to shedding light on these issues and improving the conceptual base for professional action.

We wrote: ‘Our argument is that greater effectiveness in preventing harm in individual families is most likely to come through improving skills in working with parents and carers to identify what are the specific and changeable difficulties, risks and needs which may impede their ability to offer safe, nurturing, developmental care to their children’ (Taylor et al. 2008).

We emphasised the need to document and assess ‘What is happening for this child at key moments of each day’. Our focus is on this child, in these circumstances, with documented evidence of day to day risks of harm and specific protective factors. We are entirely convinced of the need for rigorous and detailed study of the circumstances of a child, overseen and critically supported by skilled supervisors. We are not confident that a predictive tool can achieve this. We agree that there is no evidence that practitioners’ judgements are any more predictive than such tools. The problem lies in the perception that such tools and their scores are relatively robust and provide clear answers. Practitioners may then over emphasise their value, without dealing with the ethical, theoretical or methodological challenges.

Systematic assessment of the kind which the Department of Health (2004) and Scottish Executive (2007) assessment triangles represent, is crucial. Both of these assessment frameworks give guidance about the need to assess: what the child needs to grow and develop; what is needed from those who care for the child; and from the child’s wider world. The factors on which the triangles focus include those associated with risks of harm in populations. They expect detailed attention – concrete rather than abstract – to family income, housing and household stability, relationships, supports and social isolation, mental and physical health, drug and alcohol abuse. Such assessment must be probing, recognising the worker’s responsibility for this child, at this moment and responsibility to ensure that others who have a role to play are actually discharging their responsibilities.

Much of this is common sense. To say so is not to make light of the need for continuing improvements in knowledge and practice, but to recognise how much we already know about what works in protecting children. In Lord Laming's (2003) report into the death of Victoria Climbié, he wrote ‘I am convinced that the answer lies in doing relatively straightforward things well’. It was perhaps unfortunate that the report then went on to list more than 90 recommendations for change! More constructive would be to concentrate on Lord Laming’s priorities to support organisations in ensuring that workers have the training, skills, supervision and support to make systems and services work for children, meeting their needs, protecting them from harm.

Ultimately, attention needs to be focused on the social and economic circumstances in which harm is likely, as any reliance on individual identification of abuse is a very limited strategy for dealing with child abuse and neglect (Spencer & Baldwin 2005). Primary prevention through the promotion of family friendly policies is a more constructive approach. This will require a shift in focus from family pathology to societal pathology. It is our view that doing straightforward things well in day to day work with children, including probing, rigorous, child focused assessments, alongside support for families struggling to bring up their children against the odds, may be more effective than pursuing the fallacy of actuarial, predictive tools.