Introduction - Shaken baby syndrome (SBS) or traumatic brain injury caused by shaking (TBI-SBS)—a form of child abuse—is particularly serious in terms of mortality and morbidity. Recent data on these injuries are not available in France. Hospital data from the Program for the Medicalization of Information Systems in Medicine, Surgery, and Obstetrics (PMSI-MCO) could serve as a relevant data source for studying SST, provided, however, that children who are victims of shaking can be easily identified from this source. Method - Hospital data from the PMSI-MCO were analyzed for the 2015–2017 period to count and describe children aged 1 to 11 months residing in France (excluding Mayotte). A group of probable cases—hospitalized children presenting with lesions highly characteristic of SIDS—and a group of possible cases—children presenting with lesions possibly associated with SIDS—were distinguished in the analyses. Results—Over the 2015–2017 period, 1,215 children were selected, comprising 512 probable cases and 703 possible cases. The incidence rate using the most restrictive definition—that is, probable cases—is 22.1 [20.2–24.0] per 100,000 live births. It is 52.4 [49.5–55.4] per 100,000 live births when using a broader definition (probable and possible cases). The clinical signs and injuries found in probable and possible cases differ significantly. Discussion-conclusion - This exploratory analysis shows that accurately determining the number of SHC cases hospitalized using PMSI-MCO data is complex. The implementation of a specific code in the PMSI-MCO for shaken baby syndrome, as envisaged in the 2017–2019 Interministerial Plan for Mobilization and the Fight Against Violence Against Children, would allow for more precise identification of hospitalized SBI cases.