This study developed and validated suicidality surveillance indicators using NLP on electronic health records (EHRs) from 15 Parisian hospitals. Analyzing 2,911,920 records, an interrupted time-series analysis revealed a post-pandemic increase in suicide attempts, particularly among girls aged 8–17. The study demonstrates the feasibility of jointly analyzing multi-hospital EHR data for timely population-level mental health monitoring, highlighting the need for increased attention to violence against women, especially at young ages.

Romain Bey, Ariel Cohen, Vincent Trebossen, Basile Dura, Pierre-Alexis Geoffroy, Charline Jean, Benjamin Landman, Thomas Petit-Jean, Gilles Chillet, Kanoko Sallah, Xavier Tannier, Aurelie Bourmaud, Richard Delorme