- Participants: IDEEA, CHU Lille, CHU Rouen, KITE, Vidal
- Objectives: To identify adverse events and their links with demographic, medical or therapeutic data of the patient. The identification will be carried out through data mining and semantic mining.
- Work Process: The first task in this WPis data mining where large databases will be explored and verified using statistical methods and specific algorithms. The goal of the latter is to enable the members of this WP to develop a semi-automatic data mining technique that will explore regularly the databases of patients in hospital in order to detect whether adverse events continue to occur or not and if new adverse events can be detected.
The second task has to do with semantic mining which is a technique where relevant information is extracted from texts. In our case medical letters, discharge reports, orders, procedure reports, Adverse Drug Event (ADE) reports, drug approval documents and clinical guidelines. The process of mining will be iterative, this implies the refinement of mining techniques, the exploration of alternative techniques and discovery of new knowledge.
- Expected Results: First results of data mining, semantic mining and a combination of both. Multi-terminology server.
Return to The Project
|
