Qi Long, PhD
Professor, Department of Biostatistics, Epidemiology and Informatics
Associate Director, Penn Institute for Biomedical Informatics
Director, Center for Cancer Data Science
Director, Biostatistics and Bioinformatics Core, Abramson Cancer Center
University of Pennsylvania
Abstract: Enormous amounts of data have been generated as part of health care delivery and by digital devices such as electronic health records (EHRs) data and mobile health (mHealth) data. Such rich, yet complex data offer remarkable opportunities for advancing an intelligent, learning health system, and, at the same time, present significant analytical challenges. In this talk, I will share our research group’s recent journey on developing innovative statistical and machine learning methods for EHRs data and mHealth data to advance intelligent health. I will first present a project on developing distributed learning methods for predicting medical events using EHRs data from multiple healthcare systems without sharing individual-level data. Our approach can use both structured and unstructured EHRs data collected over irregular time points and has the potential to be incorporated into an EHR system as a clinical decision support system. The second project seeks to develop a clinical decision support system for accurate diagnosis of kidney obstruction which is based on a robust Bayesian prediction model trained using renogram imaging data and relevant clinical variables in EHRs in the absence of a gold standard. The third project is aimed to develop an mHealth tool for improving pain management of sickle cell disease patients by developing a hybrid statistical and mechanistic model for optimizing individualized treatment recommendations. Our recent experiences have demonstrated that building on the core principles of statistical thinking a trans-disciplinary health data science approach offers great promises to accelerate innovation and fully harness the power of ever-growing health data to tackle complex real-world problems in medicine.
The meeting agenda includes a presentation, programmatic announcements, and updates from Cancer Center program leaders.
This group meets once a month from noon – 1 p.m.