Change-point detection method for the prediction of dreaded events during online monitoring of lung transplant patients

Nassim Sahki (1, 2) , Anne Gégout-Petit (1, 2) , Sophie Wantz-Mézières (1, 2)

1 BIGS - Biology, genetics and statistics
2 IECL - Institut Élie Cartan de Lorraine

Context • Survival after lung transplantation is about 80% at 1 year and 50% at 6 years. • The two main complications responsible for deaths in lung transplant patients are infection and/or rejection. Main objective • Test the monitoring of lung transplant patients by connected sensors ; • Propose a methodology for real-time prediction of a serious event (infection and/ or rejection) via the change-point detection in the evolution of the multivariate signals collected by these connected sensors.

Domaines: Statistiques [stat]; Mathématiques [math]; Informatique [cs]

Nassim Sahki, Anne Gégout-Petit, Sophie Wantz-Mézières. Change-point detection method for the prediction of dreaded events during online monitoring of lung transplant patients. Annual PhD students conference IAEM Lorraine, APIL 2019, Dec 2019, Nancy, France. ⟨hal-02392756⟩