The epidemiological data currently available on child abuse in France are very inadequate. They do not make it possible to know the number of child victims, nor the precise context and impact of this violence. Very often, the figures put forward are underestimated. Improving the knowledge of these situations would make it possible to develop targeted primary and secondary prevention programs. In this study, is proposed an innovative method of exploiting existing national databases, the PMSI and SNIIRAM, with the aim of 1) approaching an order of magnitude of the frequency of physical abuse in children and 2) evaluating its contexts and consequences. The PMSI is a high-quality database that has been used for many years for medical research purposes. It provides a large amount of epidemiological information and can be used to create cohorts large enough to detect rare events. The SNIIRAM/EGB data allows the recovery of all data on reimbursed consumption, such as drug treatments, biological examinations carried out or consultations (with general practitioners or specialists).

Inclusion Criteria:

• All children aged 0 to 5 years hospitalized in France between 2007 and 2014.

• Based on ICD-10 codes generated within the PMSI, 3 groups of children were defined:

  1. those with at least one code dedicated to physical abuse,
  2. those with at least one code or "combination of codes" suspected of physical abuse, and
  3. all other hospitalized children, control group.

Exclusion Criteria: NA

children hospitalized in France between 2007 and 2014