Publications by authors named "Thomas Petitjean"
Article Synopsis
- The study aimed to create and validate a natural language processing (NLP) pipeline capable of identifying 18 medical conditions in French clinical notes, including various comorbidities from the Charlson index, while ensuring privacy in a collaborative research environment.
- The detection pipeline employed both rule-based and machine learning techniques, utilizing a large language model and annotated clinical notes from three research studies focused on oncology, cardiology, and rheumatology.
- Results showed high accuracy metrics, including a macro-averaged F1-score of 95.7, indicating that the collaborative effort significantly outperformed other methods, demonstrating the effectiveness of secure teamwork in developing advanced medical AI models.
Background: Older adults desire to stay independent at home for as long as possible. We developed an interactive website to inform older adults and caregivers about ways to achieve this.
Objective: This study aimed to perform an in-depth exploration among potential end users about how to improve the interactive website to better inform older adults and caregivers about ways to stay independent at home.