Approximately 20,000 fetuses are diagnosed with a congenital anomaly each year in France. Of these, 30% receive diagnostic and prognostic evaluation or medical-surgical treatment in the Ile de France region, half of them at the Necker-Enfants malades Hospital. Data on the long-term outcome of children with complex congenital malformations are quite limited and are based on small series, which often do not take into account the fetal period.
The development and possibility of using databases in perinatology is hampered by the lack of human resources to complete the data and maintain these databases over the long term. Indeed, the relevant clinical information in neonatology is contained in the patient files which are mainly made up of poorly structured textual documents such as hospitalization or consultation reports. The data must therefore be collected manually and the collection of data can only be entrusted to operators with medical knowledge or even in neonatology, given the specificity of the pathologies.
The MACADAMIA project brings together the Neonatal Medicine and Intensive Care Unit, Obstetrics and Fetal Medicine Department of the Necker-Enfants Malades Hospital, TEKLIA, the Laboratoire de Traitement et Communication de l’Information at Telecom ParisTech and the Information Sciences for Personalized Medicine team of the University of Paris. It was awarded and funded by the Fondation Université de Paris under the call "Sauvez la via" in 2020.
The MACADAMIA project will develop a semi-automatic data extraction tool which, through collaboration between Artificial Intelligence and the medical staff, will allow a significant saving of time in the constitution of databases for clinical research, while ensuring a quality data collection. A collaboration between the Necker-Enfants Malades Hospital and TEKLIA has been initiated to take advantage of the Arkindex document processing platform through a research partnership.