Seminar Series: Janie Coulombe
Recent advances in causal inference under irregular observation times for the outcome
Janie Coulombe, PhD
Assistant Professor
Department of Mathematics and Statistics
Université de Montréal
Short Biography:
Janie Coulombe is an Assistant Professor in the Department of Mathematics and Statistics at Université de Montréal. She completed a PhD in Biostatistics from McGill University (2021) and did a postdoctoral internship at McGill University under the supervision of Dr. Erica E. M. Moodie. Her research focuses on the development of causal estimators with good properties that address the special features of electronic health records data, such as irregular observation times and missing data. Her research is influenced by her experience at the Lady Davis Research Institute at the Jewish General Hospital (Montreal) where she worked with electronic health records and administrative longitudinal data while she was a statistical analyst, before starting her PhD.
Abstract:
Electronic health records (EHR) data contain rich information about patients’ health condition, comorbidities, clinical outcomes, and drug prescriptions. They are often used to draw causal inferences about treatment effectiveness. However, these data are not experimental and present with special features that may affect the causal inference when they are not addressed. One of these features is the irregular observation of the longitudinal processes used in the inference. We focus on an irregularly observed longitudinal outcome on which we aim to assess a causal treatment effect. The work on irregularly observed processes in causal inference is relatively recent. In previous work, with co-authors I demonstrated that the irregular observation of the outcome can bias causal effects. In this presentation, I will review why irregular observation times can be problematic, discuss some examples, and present some recent work in this area of research.
Keywords:
Causal inference, Observational data, Electronic health records, Covariate-dependent monitoring times, Irregular observation, Confounding.
Date: Friday, October 6th
Time: 1:30 pm - 2:30 pm
Location: PHFM 3015 (Western Centre for Public Health and Family Medicine)
