An algorithm developed by the Dijon University Hospital to detect battered children
It has existed since 2019, but a recent study based on 170 cases validates its effectiveness. An article was published this week in the Epidemiological Bulletin of Public Health France. The algorithm designed by the Dijon University Hospital to detect cases of child abuse was applied to the files of beaten or potentially beaten little patients. Result: the younger they are, or even babies, the more reliable the instrument.
Unmasking medical nomadism
The algorithm was effective in over 90% of cases when dealing with infants. “This instrument makes it possible to identify the various hospitalizations of children throughout France. welcomes Catherine Quantin, head of the biostatistics department at the Dijon University Hospital and public health professor behind the project. If a child is taken care of a few weeks apart in different establishments for fractures, in the event of medical nomadism, the algorithm detects it. Or the doctors do not have these elements which can put them on alert.
The algorithm also takes into account the location of the lesion, which makes it possible to know if the statements of the parents are compatible with the injuries of their child. Or diagnosing abuse as soon as possible is essential, because this violence can be fatal or leave lifelong scars. One in ten shaken babies will die, while three-quarters of those who survive will have serious consequences, particularly neurological ones.
Mapping violence against children
In addition to the particular cases that this algorithm points out, it also makes it possible to give a national map of violence against children. Besides, in 2020, the Dijon University Hospital had measured a 50% increase in acts of mistreatment of minors during the first confinement compared to the previous three years. Data that corroborated the finding of the 119 hotline for children in danger.