Detección temprana e identificación de riesgos de salud con PLN y argumentación1>
(2023 - 2026)
This project intends to develop annotated corpus, basic and advanced NLP tools and models specifically suited to the health domain in order to employ the generated resources to cope with two use cases. The project focuses in uses cases aiming to discover new knowledge on health risks and supporting the early diagnoses in mental and cardiovascular health, mainly, by the secondary use of Electronic Health Records (EHRs) and free text. For the first use case, the team will conceive a NLP system to detect high risk suicidal behaviors in children and adolescents by automatically extracting information from structured (questionnaires) and unstructured texts (free text) to identify evidences that show suicidal risk situations according to the established psychiatric ideation-to-action framework. For the second use case we will develop an NLP system to automatically discover potential risk factors associated with cardiovascular complications (for example, ischemic stroke, cardiomyopathy) after a first episode. To that end we will use patients' EHRs, progress notes and ECGs and confront these new knowledge to the existing scores and guidelines used in hospitals to ameliorate them.
Organization: PROYECTOS DE GENERACIÓN DE CONOCIMIENTO2022
Main researcher: Arantza Casillas, Aitziber Atutxa
Jose Mari Arriola, Aitziber Atutxa, Arantza Casillas, Josu Goikoetxea, Koldo Gojenola, Nuria Lebeña, Maite Oronoz, Alicia Pérez