The great challenge of using electronic medical data in the age of artificial intelligence
November 2023.- Hercules DalianisProfessor in Computer and Systems Sciences, from the University of Stockholm, Sweden, visited Chile in the framework of the Fondecyt project "Creation of linguistic and computational resources for clinical language in Spanish". "Creation of linguistic and computational resources for clinical language in Spanish", a project of the DCC academic and IMD of the academic of the DCC and IMC UC, Jocelyn Dunstan Escudero.
"Privacy is key when working with data in the healthcare area. Practically all the information we collect belongs to patients who are vulnerable, for example, to different types of discrimination," says IMFD researcher Jocelyn Dunstan Escudero. One of the great challenges when working with medical data are the violations that the use of this data can bring to those who provide it: privacy and security is one of the main challenges for those who seek to work with large volumes of this data.
Hercules Dalianis has experience in clinical data anonymization. In the activity, held in the Sala Colorada of the Extension Center of the Catholic University, he tested live his application for clinical text recognition, in addition to explaining the guidelines to establish the anonymization of medical data that his teams have used, which allows the use of the databases giving guarantees of privacy of the data that the People deliver to the medical teams.
The work "Creation of linguistic and computational resources for clinical language in Spanish" seeks, through the use of different data science techniques, to optimize resources to achieve the great goal of reducing waiting lists in public health services. seeks, through the use of different data science techniques, to optimize resources to achieve the great goal of reducing waiting lists in public health services.
"In this project, we created the first annotated corpus in Chile, which is currently available and released to the scientific society. One of the things that is interesting to highlight is that the construction of this corpus was achieved thanks to the Transparency Law, with which the team of researchers managed to obtain eleven million interconsultations, which were requested individually from each of the 29 health services that exist in the 15 regions of the country", said the researcher in the activity.
This project was funded by the State of Chile through Fondecyt, with the support of the Ministry of Health, the Chilean Safety Association, the Center for Mathematical Modeling of the University of Chile (CMM) and the Millennium Institute Foundational Research on Data (IMFD).
Check out the registration of the presentation here: