In today’s data-driven world, risk prediction plays a pivotal role in advancing healthcare. Accurate risk prediction models are essential for informed decision-making and can significantly impact both individual and societal outcomes. Accurate risk assessment is the cornerstone of a learning health system. However, developing robust and fair models presents challenges due to the complexities of real-world data, including suboptimal data quality, data incompleteness, sampling bias, and uneven data representation across population subgroups. Over the years, our research group has made dedicated efforts to develop novel statistical methods to address these challenges, with recent work particularly focused on risk modeling using electronic health record (EHR) data.
This workshop series, consisting of six sessions over the academic year, is designed to introduce these new methodologies to the Penn Medicine research community, with a focus on advancing predictive healthcare and precision medicine with EHR data. It will serve as a collaborative forum for discussion and foster opportunities for synergy in new research initiatives. Each session will take place over 1.5 hours during lunchtime (12:00-1:30 pm) and will be divided into four parts:
Introduction (15 minutes): An overview of the session topic, featuring a range of real study examples to provide context and relevance.
Statistical methods (30 minutes): A high-level discussion focusing on the core ideas of the statistical methods, with technical details minimized. Relevant papers will be distributed before the meeting to help attendees familiarize themselves with the content.
Case Studies (15 minutes): Real-world case studies from ongoing research at Penn and beyond will be presented to illustrate practical applications and challenges. Investigators who have utilized the methods will lead these case study presentations and discussions.
Discussion (30 minutes): An interactive segment where participants can share their experiences, ask questions, and brainstorm potential projects. Attendees are encouraged to bring challenges from their own research and explore ideas for relevant new research during this time. A biostatistics colleague will help facilitate the discussion.
This workshop series is offered by the Statistical Center for Translational Research in Medicine (SC-TRM), directed by Prof. Jinbo Chen in the Department of Biostatistics, Epidemiology & Informatics (DBEI) at Penn. It is supported by the Center for Epidemiology and Biostatistics (CCEB), and the Institute for Translational Medicine and Therapeutics (ITMAT). We thank CCEB for its generous support on lunch arrangements and all administrative support, which is essential in ensuring that this workshop is both productive and enjoyable for all participants
