Abstract
Background
Child maltreatment includes physical, psychological, sexual abuse and acts of neglect. Among the resulting non-accidental injuries, burns are responsible for an important morbi-mortality. The main objective was to build a detection algorithm of non-accidental paediatric burns (NAB), using ICD-10 codes in the hospital resumes from the French Hospital Discharge Database (HDD).
Methods
Children aged 0 to 16 years old hospitalised at the University Hospital of Tours from 2012 to 2017 with a coded burn were included. “Probable” or “possible” HDD cases of NAB were defined based on specific ICD-10 codes during the inclusion stay or the previous year. A chart review was performed on all the HDD cases and HDD non cases matched on sex and age with a 1:2 ratio. Performance parameters were estimated for three clinical definitions of suspected child maltreatment: excluding neglect, including neglect with restriction then broad definition. For clinical cases, report to the judicial authority (RJA) or worrying information (WI) was searched.
Results
Among the 253 included children, 83 “probable” cases and 153 non cases were analysed. Sensitivity varied from 48% (95%CI [36-60]) to 90% [55-100] when excluding neglect, specificity from 70% [63;77] to 68% [61;74]. The positive and negative likelihood ratios varied respectively from 1,6 [1,2;2,3] to 2,8 [2,1;3,7] and from 0,7 [0,6;0,9] to 0,1 [0,0;0,9]. The proportion of clinical cases with no RJA/WI without reason varied from 0 (when excluding neglect) to > 85% (with broadest definition); all corresponded to a possible isolated neglect.
Conclusions
The performances of the algorithm varied tremendously according to the clinical definition level of child maltreatment. Neglect is obviously difficult to clinically detect. Training for healthcare professionals and qualitative studies on obstacles to RJA/WI should be added to this work.
Key messages
The performances of an algorithm to detect non-accidental pediatric burns (maltreatment) using the French hospital discharge database dropped when including neglect, difficult to diagnose clinically. Training for healthcare professionals and qualitative studies on obstacles to the judicial authority (RJA) or worrying information (WI) should be added to this diagnostic study.
Challenges of a French Hospital Discharge algorithm to detect non-accidental paediatric burns
