RESUMO
INTRODUCTION: In all cases of suspected child abuse, accurate risk assessment is vital to guide further management. This study examined the relationship between risk factors in a risk assessment matrix and child abuse case conference outcomes. METHODS: Records of all children hospitalised at United Christian Hospital in Hong Kong for suspected child abuse from January 2012 to December 2014 were reviewed. Outcomes of the hospital abuse work-up as concluded in the Multi-Disciplinary Case Conference were categorised as 'established', 'high risk', or 'not established'. All cases of 'established' and 'high risk' were included in the positive case conference outcome group and all cases of 'not established' formed the comparison group. On the other hand, using the Risk Assessment Matrix developed by the California State University, Fresno in 1990, each case was allotted a matrix score of low, intermediate, or high risk in each of 15 matrix domains, and an aggregate matrix score was derived. The effect of individual matrix domain on case conference outcome was analysed. Receiver operating characteristic curve analysis was used to examine the relationship between case conference outcome and aggregate matrix score. RESULTS: In this study, 265 children suspected of being abused were included, with 198 in the positive case conference outcome group and 67 in the comparison group. Three matrix domains (severity and frequency of abuse, location of injuries, and strength of family support systems) were significantly associated with case conference outcome. An aggregate cut-off score of 23 yielded a sensitivity of 91.4% and specificity of 38.2% in relation to outcome of abuse categorisation. CONCLUSIONS: Risk assessment should be performed when handling suspected child abuse cases. A high aggregate score should arouse suspicion in all disciplines managing child abuse cases.