Passer au contenu

/ Département de mathématiques et de statistique

Je donne

Rechercher

 

Coulombe, Janie

Vcard
Assistant Professor

Faculty of Arts and Science - Department of Mathematics and Statistics

André-Aisenstadt Office 4243

514 343-7977

Courriels

You must click to get the email

Research area

I am interested in developing new causal estimators (average treatment effet and optimal dynamic treatment regimes) that can be used jointly with observational data. These estimators are built to consider some challenges in analyzing observational data, such as irregular visits, missing data and confounding of the treatment-outcome relationship.

My research is funded by an NSERC discovery grant and I am a Chercheuse-boursière Junior 1 du FRQS. 

Currently, my research focuses on the development of causal estimators for different causal parameters that are more flexible and efficient. With Prof. Shu Yang (North Carolina State University), we developed a new multiply robust estimator (https://arxiv.org/abs/2304.08987) for data with irregular observation times and confounding. That estimator is flexible and more efficient than others from its semiparametric class. 

I am also interested in the many different ways of addressing irregular visit times and in comparing imputation and approaches based on inverse weighted estimating equations (work under review).

For more details, please see my page  https://janiecoulombestat.github.io or my Google Scholar page at https://scholar.google.com/citations?user=UmVoZQwAAAAJ&hl=fr&oi=ao .

 ---

Papers on irregular visits:

Preprint:

Coulombe, J., Yang, S. Quadruply robust estimation of marginal structural models in observational studies subject to covariate-driven observations. (travail en révision)

Peer-reviewed papers:

Coulombe, J., E. E. Moodie, S. M. Shortreed, et C. Renoux (2023). "Estimating individualized treatment rules in longitudinal studies with covariate-driven observation times." Statistical Methods in Medical Research, 32(5): 868-884. 

Coulombe, J., E. E. Moodie, R. W. Platt and C. Renoux (2022). "Estimation of the Marginal Effect of Antidepressants on Body Mass Index under Confounding and Endogenous Covariate-Driven Monitoring Times." Annals of Applied Statistics 16(3): 1868--1890.

Coulombe, J., E. E. Moodie and R. W. Platt (2021). "Weighted regression analysis to correct for informative monitoring times and confounders in longitudinal studies." Biometrics 77(1): 162-174.

Coulombe, J., E. E. Moodie and R. W. Platt (2021). "Estimating the marginal effect of a continuous exposure on an ordinal outcome using data subject to covariate'-'driven treatment and visit processes." Statistics in Medicine 40(26): 5746-5764.

Related to optimal dynamic treatment regimes:

Moodie, E. E. M., Bian, Z., Coulombe, J., Lian, Y., Yang, A. Y., et Shortreed, S. M. (2023). "Variable selection in high dimensions for discrete-outcome individualized treatment rules: Reducing severity of depression symptoms." Biostatistics, kxad022 (sous presse).

Coulombe, J., E. E. Moodie, S. M. Shortreed, et C. Renoux (2023). "Estimating individualized treatment rules in longitudinal studies with covariate-driven observation times." Statistical Methods in Medical Research, 32(5): 868-884. 

Moodie, E. E. M., Coulombe, J., Danieli, C., Renoux, C., et Shortreed, S. M. (2022) "Privacy-preserving estimation of an optimal individualized treatment rule: A case study in maximizing time to severe depression-related outcomes." Lifetime data analysis, 28(3): 512-542. 

Coulombe, J., Moodie, E. E. M., Shortreed, S. M., et Renoux, C. (2021) "Can the risk of severe depression-related outcomes be reduced by tailoring the antidepressant therapy to patient characteristics?" American Journal of Epidemiology, 190(7): 1210-1219.

Coulombe, J., Moodie, E. E. M., Shortreed, S. M., et Renoux, C. (2021) "Coulombe et al. respond to 'Baby steps to a learning mental health-care system". American Journal of Epidemiology, 190(7): 1223-1224.

 

Coulombe, J., Moodie, E. E. M., Shortreed, S. M., et Renoux, C. Coulombe et al. respond to 'Baby steps to a learning mental health-care system'. American Journal of Epidemiology, 190(7): 1223-1224.

Student supervision Expand all Collapse all

Estimation de l’effet causal d’une exposition cumulative sur une réponse continue dans les études enclines à la confusion et aux visites irrégulières Theses and supervised dissertations / 2025-07
Dicaire-Cartier, Mathilde
Abstract
The purpose of this thesis is to estimate the causal effect of a cumulative exposure (i.e., that cumulates in time) on a continuous response measured repeatedly in non-experimental longitudinal studies. More precisely, under certain causal assumptions, the proposed methodology adjusts for confounding and the bias due to irregular observation times. Methods have recently been proposed to address these challenges, but they mostly focused on acute treatment effects, which occur rapidly and are short-termed. To take into account the bias induced by confounders, the Inverse Density of Treatment weighting method (IDT) is introduced. This weight can be used for continuous exposures that follow a normal distribution. Using a proportional rate model for the visits (i.e., for irregular observation times), a second weight (IIV) is introduced and can be used to address the bias due to that irregularity. The Inverse Intensity of Visits weighting models the rate of visits. By combining both weights and by using the method of weighted least squares or the corresponding estimating equations, the Inverse Density Exposure and Monitoring estimator (IDEM) is derived. It can consistently estimate the causal effect of cumulative exposures on continuous responses for studies where patients are observed irregularly and where there is confounding, i.e., a balance of patient characteristics between exposure groups is not insured. Using the properties of two-step estimators, we present a sketch proof of the asymptotic distribution of IDEM that is valid under certain specified conditions. A simulation study is performed to evaluate the performance of the IDEM estimator. Under four scenarios where the exposition and visit models vary, the causal estimates obtained with IDEM are compared with those obtained with the ordinary least squares estimator (OLS) and two other estimators that we call IDT and IIV. The estimates of the four estimators are again compared by applying them in the causal analysis of the Phenobarb dataset available in the MEMSS package of the CRAN repository in R. This longitudinal dataset contains a continuous treatment variable, irregularly observed responses, and potential confounders.

Link to the Papyrus's document

Évaluation de la modélisation et des prévisions de la vitesse du vent menant à l'estimation de la production d'énergie annuelle d'une turbine éolienne Theses and supervised dissertations / 2015-04
Coulombe, Janie
Abstract
Following an internship with the company Hatch, we have access to datasets that are composed of wind speed time series measured at different sites accross the world and over several years. The wind speed engineers from Hatch are using these datasets jointly with Environment Canada databases in order to ascertain the wind energy potential of these sites and to know whether it is worth installing wind turbines there. For a few years, some companies are also offering mesoscale simulations of wind speed based on different environmental characteristics from the site we want to evaluate. We would like to know if it is worth paying for those mesoscale datasets and if they can be used to provide better estimations of the wind energy potential. Among other things, these data could be used to provide a better estimation of the long term mean wind speed. Since we already possess measured datasets, we will also use them to test, with statistical methods, the methodology currently used and the different steps leading to an estimation of the wind energy production. First of all, we will see what are the different methods that could be used to extrapolate wind speed to a wind turbine’s height and we will evaluate those methods with the mean squared extrapolation error. Also, we will study wind distribution modelling by the Weibull distribution and consider its variability over time. Finally, cross-validation and block bootstrap will be used to see whether we should use mesoscale data instead of wind data from Environment Canada or whether it would even be beneficial to use both kind of data to predict wind speed. In summary, the whole methodology used by wind speed engineers to estimate the energy production will be tested from a statistical point of view and we will attempt to propose changes in this methodology that could improve the estimation of the wind speed annual energy production.

Link to the Papyrus's document

Our researchers by theme