Making the Case for Greater Certainty in Child Protection

Paediatricians, as with all healthcare professionals, have a duty to protect children from harm. They are called on to make the case for inflicted injury, often during a busy service week. If an inflicted injury to a child is not identified and acted on appropriately, the child or a sibling may present with further, serious or fatal injury, and, conversely, a decision in favour of inflicted injury that is not adjudicated as such may have serious human and societal consequences, including needless separation of the child from the family. This commentary and the related article by Arthurs et al are bookended by two such situations: the cases of two children removed from their respective parents on the basis of suspicious injuries, only to be returned to their parents following reappraisal of the radiological evidence by courtappointed experts, and the tragic cases of two children abused and murdered by their parents. The Viewpoint article was written in response to press coverage surrounding the cases of the two children, removed from, and then returned to their parents. The authors, eminent in the fields of child protection, paediatric radiology and the law, ask, ‘are we getting it right’ in safeguarding children? In broad statistical terms, we are getting it right, for example, the rates of fatal child maltreatment in England and Wales have fallen steadily over the past 30 years and the rates for children on Child Protection Plans in England have remained relatively constant since 2013. Yet statistics belie the human and social costs in individual situations that go might go wrong. This commentary sets out how we can be better at getting it right. Arthurs et al discuss ‘diagnostic uncertainty’, a term often used in relation to cases of suspected child maltreatment. There will always be a degree of uncertainty in relation to causation of injuries that might have occurred within the home, when there are no independent witnesses to, or recording of the event, and when any perpetrator admission is deemed unreliable. (There is a question as to how best to communicate such uncertainty to the multidisciplinary team and the courts). But, in getting it right, we endeavour to achieve a degree of certainty that approaches zero error rate. This is beyond the burden of proof required in the family courts, where terms such a ‘balance of probabilities’ (UK jurisdictions) or ‘reasonable medical certainty’ (many US jurisdictions) apply and are generally defined as ‘more likely than not’. In achieving this, a shift in clinical decisionmaking has occurred along a continuum from an overreliance on intuitive skills—important as they are, but subjective and valueladen—through heuristic ‘rule of thumb’ thinking, with its attendant bias, to analytical thinking, in order to arrive at a conditional judgement about the probability of inflicted injury. This ongoing shift is being driven by the emergence of a strong evidence base, and with this, the development of analytical tools, such as applications of Bayes’ theorem. A discussion on the evidence base for Bayes theorem in discriminating inflicted from accidental injuries is beyond the scope of this paper. Their application in an infant with a solitary bruise will suffice. While decisions to initiate child protection investigation are multifactorial, the multidisciplinary team can place the onus on the paediatrician to make the case without equivocation. Through a series of landmark longitudinal studies, the Cardiff group have added to their earlier work on Bayesian analysis by demonstrating a strong relationship between the presence, number and location of ‘everyday bruising’ in children at different stages of motor development, and the pattern of bruising in children with inherited bleeding disorders. 5 This work contributed to their validation of Pierce’s TEN4 (torso, ear, neck) Bruise Clinical Decision Rule; that the number and TEN location in children of any age, and a single bruise in a TEN location in a premobile infant, is ‘supportive of abuse’, when inherited bleeding disorders have been excluded. These and other computerassisted models using probabilistic approaches will no doubt become even more discriminating for inflicted injury when adequately powered comparison studies become available. But there will still room for analytical thinking and selfquestioning by the paediatrician—‘how comfortable am I with this decision?’—discussion with peers and child protection experts at the time, and, at a later phase in the process, peer review. The radiologist plays a fundamental role in identifying the presence of additional head and skeletal trauma in young children presenting with possible inflicted injuries, through skeletal surveys and CT head scanning, with relatively high positive yields obtained. Issues relating to radiological opinion was at the heart of the journalists’ criticism of the two cases outlined by Arthurs et al. They point out that image interpretation can be subjective. For example, the ability to distinguish between a classic metaphyseal lesion fracture and a normal anatomical variant such as a metaphyseal spur can be difficult, with disagreement even among experienced paediatric radiologists. Wellconducted, multicentre prospective trials to potentially allow more objective radiological assessment may go some way to alleviate such uncertainties but it is probable that there will always be a degree of subjectivity as in the interpretation of any image, be it radiological or otherwise. To reduce the likelihood of error in radiological interpretation, ‘double reporting’ of skeletal surveys by two radiologists ‘with at least 6 months specialist paediatric training, including experience of suspected physical abuse in children’, is recommended by the Royal College of Radiologists (RCR). However, this is not always feasible due to a national shortage of radiologists. The RCR 2020 Census found that 33% of consultant posts remain unfilled. More pertinent to child protection, there is also a shortage of paediatric radiologists in both tertiary centres and smaller hospitals. This has led to a smaller pool of those with sufficient expertise to provide double reporting, with relatively inexperienced radiologists being involved. These issues have been extensively commented on by the British Society of Paediatric Radiology National Working Group on Imaging of Suspected Physical Abuse. The development of appropriately funded regional radiological networks such that second or expert opinions can be incorporated into formal Child Protection, Birmingham Women and Children’s Hospital NHS Foundation Trust, Birmingham, UK Radiology, Birmingham Women and Children’s Hospital NHS Foundation Trust, Birmingham, UK