Grant: Foundation Grant

<html> <body> <p>My research program is focused on statistical methods for inference in epidemiology, with focus on methods for perinatal epidemiology and for pharmacoepidemiology.<br> <br> The study of pregnancy outcomes is limited because the appropriate base cohort of conceptions cannot be constructed from typical observational data. This leads to compromises in defining cohorts and measures of risk, which can introduce bias. Further, processes that occur during pregnancy are either poorly recorded (e.g., maternal weight gain) or unobserved (e.g., fetal growth). Further, observational studies record data on these processes only at a few time points (sometimes only at birth). Finally, most research to date has not appropriately considered the role of fetal growth and maternal weight gain as mediating variables. I intend to apply conceptual and statistical tools for causal inference to the study of maternal exposures and pregnancy outcomes, incorporating recent research on methods dealing with mediation to account for the potential mediating effects of fetal growth, maternal weight gain, and preterm birth.<br> <br> My second research area is pharmacoepidemiology. I am the methods lead for the Canadian Network for Observational Drug Effect Studies (CNODES), funded to conduct rapid-response studies of drug safety in administrative databases. Such research is limited due to confounding, especially by indication. My research involves development and implementation of methods for targeted semiparametric inference to avoid this. Methods for causal inference typically require estimates of the propensity score to account for measured confounding; in my ongoing research I am developing and evaluating approaches to model the propensity score for the most efficient causal inference for exposure-outcome associations. CNODES is an ideal environment for methods development. The study data serve as a platform for methods work, and methodological developments produced by my research team can be quickly integrated into the ongoing drug safety projects.<br> <br> I have published 95 papers in the past five years, and my H-index is 49 per Scopus. My research has been published in top journals and many have been recognized as fundamental contributions. My method for correcting gestational age data in has been implemented in Canadian, Ontario, and French references for birth weight for gestational age. My methods have been used in work that has had public health impact on breastfeeding, and on socioeconomic pathways to preterm birth, and on drug regulation. My research team, coupled with program experts Kramer, Suissa, Tagalakis, Bodnar, Moodie, and Ernst, provide substantive and methodological expertise to ensure the success of this research program.<br> <br> Most of the ongoing research projects in my research team are being led by trainees, and all of them provide training opportunities. My research team involves several students and fellows who work in a collaborative environment, and all trainees have access to relevant journal clubs and working groups.<br> <br> These projects and my collaborative work allow me to be active in development of statistical methods and implementation of methods of direct relevance to epidemiologic and other researchers, and in knowledge translation from methods to substantive research. This work has both long-term impact in terms of improving the methodological quality of health research, and direct impact when my collaborative work is used to drive changes in health care.</p> </body> </html>