Inférence causale et modélisation avancée : mieux comprendre et évaluer pour améliorer la fiabilité des résultats en santé

 

Geneviève Lefebvre

Université du Québec à Montréal [UQAM]

 

Domaine : santé des populations

Programme Chercheurs-boursiers - Junior 2

Concours 2019-2020

My research interests focus on developing new data analysis tools to answer concrete health questions. One of the applications considered will be to determine whether the medications commonly used against asthma have an impact on birth weight by affecting or not the duration of pregnancy. I suggest addressing this problem through a certain class of statistical models, called mediation models, which are used when the effect (or action) of an exposure could indirectly affect the outcome studied through an intermediate variable.

Being directly applicable to several other health research problems, some methodological themes related to mediation models will be studied and deepened. I will also focus on the application of advanced statistical models that can shed light on other important topics such as overdiagnosis of attention deficit disorder with or without hyperactivity (ADHD) in children from Québec and the long-term sequelae of treatment against pediatric leukemia among survivors.