Artificial intelligence for the evaluation of profiles, trajectories and management of vulnerability in health: COVID-19, frailty of the elderly and cancer
TRADIAB
Trajectoire du risque de complications et réponses thérapeutiques dans le diabète de type 2
Underway
01/10/2021
-
01/10/2024
Main objectives
This project aims to identify, from the data of the AP-HP health data warehouse (EDS), trajectories of development of complications and responses to treatment. The identification of variables associated with these trajectories will allow to identify the subjects most at risk of diabetes evolution towards complicated forms and those who could benefit the most from diabetes treatments in terms of protection against the occurrence of complications
AP-HP
Challenge
Diabetes is a chronic disease that can cause multiple complications (ophthalmologic, renal, cardiovascular, podologic), sometimes severe (blindness, dialysis, amputations). These complications do not affect all diabetics in the same way: some remain free of complications or develop mild complications while others will develop serious complications. Even if the main determinants of risk are known, it is difficult to identify early the patients who are most at risk of developing complications
Methodology/Technology used
Use of statistical learning methods to extract data and build predictive models using a large number of variables
Expected outcomes
- Identification of profiles at risk of complications allowing the implementation of effective prevention strategies at an early stage
- Identification of therapeutic response profiles (variation in the effect of a therapeutic strategy according to certain parameters)
- Medium to long term development of tools for the analysis of care pathways and the definition of indicators of quality of care
- Identification of therapeutic response profiles (variation in the effect of a therapeutic strategy according to certain parameters)
- Medium to long term development of tools for the analysis of care pathways and the definition of indicators of quality of care
Contacts
Gaël Varoquaux () & Louis Potier ()
Members
AP-HP, Inria, Inserm
Other projects
Description
Learning a deep representation of patient records for event prediction and patient segmentation
Names of partners involved
AP-HP, Inria