I am fortunate to lead a team of brilliant data scientists and engineers who are moving the needle on critical care research and patient care. As the Director of Data Science at the Weil Institute, I help members leverage data science to enhance their research. As a faculty in Michigan Medicine, I focus my research on analyzing data from various sources including electronic health records, bedside monitors, radiology images and wearable devices, utilize various analytical techniques such as machine learning, signal and image processing, causal inference and continuous and discrete optimization to create clinical decisions support systems for various medical conditions ranging from in-hospital deterioration to congestive heart failure.

More recently, my team has focused on the broader impact of risk prediction models on clinical outcomes, delivery of care and healthcare organization. Our team has led the deployment of multiple predictive models into Michigan Medicine’s EHR and have been studying the impact of these models on delivery of interventions. We have been conducting usability studies that are focused on understanding the optimal approaches for integration of predictive models into provider workflows. Our ultimate goal is to conduct clinical trials on these models to assess their impact on patient outcomes. Finally, my team has created a tool for post-market surveillance of predictive models that are deployed into EHR . This is an often-overlooked aspect of model deployment; the impact of the model outputs on the clinical practice can interfere with outcomes of interest in a way that makes evaluation of the model performance after deployment challenging.