The solution to deal with the ills that affect emergency services may lie in artificial intelligence. Researchers from Inserm and the University of Bordeaux have developed an algorithm capable of classifying emergency room visits for trauma through the analysis of clinical reports.
Trauma accounts for a third of emergency visits and 9% of mortality in France. The trauma can be related to everyday life accidents or traffic accidents, but also to suicide attempts, for example.
At each visit to the emergency room, caregivers must fill out reports that include a certain amount of information: statement of symptoms, condition of patients, or circumstances of trauma occurrence.
From there, the researchers found that this data remained largely untapped. Above all, these reviews are unstructured texts, with technical terms and abbreviations. The researchers’ project is called TARPON (Automatic Processing of Emergency Room Summaries for a National Observatory), the results of which have been published in the journal Internet medical research journal Artificial intelligence.
97% correct classifications
And that’s where artificial intelligence comes in. In order to extract the relevant information from the reports without having to review them one by one, the researchers developed an automatic processing technique based on artificial neural networks. The analysis of the information is carried out anonymously.
The system was trained from a sample of 500,000 adult emergency reports at the Bordeaux University Hospital. A powerful server, installed in the hospital, makes it possible to implement the algorithm and thus classify the lesions according to their type.
Result: 97% of the reports are classified correctly thanks to this algorithm. This work will allow the establishment of a national trauma surveillance system, but also for epidemiological analysis.
As of this summer, the study of the data will be able to begin on the technological platform of the Health Data Hub. The project will be extended to fifteen emergency services in France.
Source: BFM TV
