About the presentation: Estimating the effects of discontinuation, deprescribing, or otherwise stopping drug treatments using observational data has been and continues to be extremely challenging. Two newer study designs (cloning, censoring, and weighting and a prevalent new user approach) have the potential to avoid immortal time and other self-inflicted biases common in studies of discontinuation. That said, what exactly each design estimates differ in subtle but important ways. This presentation will go over these distinctions and discuss important strengths and weaknesses of these designs for drug discontinuation studies.
About the presenter: Michael Webster-Clark is a pharmacoepidemiologist who specializes in epidemiologic methods research, with a focus on novel study designs and extending inferences to external target populations.
About the presentation: Synthetic health data are generated from existing electronic health datasets, so that they maintain the properties of the original data. Producing and sharing synthetic data may shorten the time to conduct studies and reduce legislative and ethical barriers to data sharing. Our goal is to provide an overview of current uses of synthetic health data in the Canadian context, and to discuss potential innovative uses of synthetic electronic health data for CNODES.
About the presenters:
Lisa Lix is a Distinguished Professor in the College of Community and Global Health at the University of Manitoba and Director of Data Science in the George & Fay Yee Centre for Healthcare Innovation. She holds a Tier 1 Canada Research Chair in Methods for Electronic Health Data Quality.
Hassan Maleki Golandouz is a PhD candidate in the College of Community and Global Health at the University of Manitoba. His research focuses on developing fairness-aware and equitable predictive models using electronic health data, as well as investigating the role of synthetic data in their development and validation. He also has a background in Computer Engineering and Artificial Intelligence.