摘要：Public concern over, and recent developments in, the field of child protection are well known (Munro 2012). Within these developments, there has been a strengthening of the role of social work with an increased focus on, and recognition of, professional knowledge, skills and ‘expert’ decision making (Munro 2011; Gilbert et al. 2011). Focus on inter-professional and multi-agency practice has developed alongside (Frost and Lloyd 2006; Frost and Robinson 2007; Ruch 2009), and continues to have a clear focus in the recently issued Working Together to Safeguard Children (H.M. Government 2013). This paper enquires into a relatively under-explored area of multi-agency child protection practice, specifically, that of the police (that is,. non-specialists in child protection) making an urgent, first response to a child protection call, often out of hours and without immediate recourse to the expertise and knowledge of child protection practitioners. In these situations, the police are called upon to make key decisions: for example, whether to immediately protect and remove children using police protection powers (Section 46(1) Children Act 1989), to refer on to local authority social services for a s47 investigation or s17 services, or to take no further action. There is exploration of the issues raised by a request from the police to develop an assessment framework as an aid to practice in these situations. The police had in mind an equivalent instrument to a domestic abuse framework already adopted by them. The paper reviews debates, particularly about predictive efficacy, in the construction of assessment and decision-making tools. The nature and distinction between consensus based and actuarial risk assessment instruments are examined, as are challenges for general multi-agency working, alongside the specific challenges for front line police officers. It is proposed that a consensus based assessment framework to support decision making, drawing on empirically tested, actuarially informed risk assessment evidence, which is collaboratively tested with a multi-agency group, is indicated.
