Causal inference, selection bias and missing data, reproductive and perinatal epidemiology, cancer epidemiology
Dr. Smith is an epidemiologist who focuses on developing and applying methods for causal inference in public health research. Dr. Smith approaches her applied research using a target trial framework, through which she has helped to clarify questions about exposures during pregnancy and to improve understanding of the effects of COVID-19 on birth outcomes. Her work in prostate and breast cancer addresses questions about complex treatment strategies over time. Her previous work on sensitivity analysis extended the E-value framework to quantify possible effects of selection bias, alone and jointly with other biases. She has also contributed to the literature on mediation, including sensitivity analysis for unmeasured mediator-outcome confounding. Currently she is researching methods to assess sensitivity to missing data under various assumptions.