Data-Driven Medicine

Five promising trends toward improving patient care

Published: Monday, May 22, 2017 - 12:02

People have touted the potential of big data and computation in medicine for what feels like decades, promising more effective and personalized treatments, new research discoveries, and smarter clinical predictions. But only recently have these technologies made it to the clinic, where they can actually improve patient care.

At University of Chicago Medicine, several collaborations between physicians, researchers, and computational experts have produced such pioneering applications, from the pathology lab to the critical-care wards.

A recent feature in Medicine on the Midway, the UChicago Medicine alumni magazine, offers a tour of these local innovations, spotlighting several Computation Institute (CI) researchers and partnerships along the way. Science Life recently ran the feature as a five-part series, and we’ve broken out the CI contributions to each installment. Read about how agent-based modeling, predictive analytics, high-performance computing, and other computational and data-science efforts are already helping cure disease and improve patient health.

Data-driven medicine: The introductory section speaks to CI faculty and fellow Samuel Volchenboum about the current state of medical data science, and profiles SIMPL, a collaboration between Volchenboum’s Center for Research Informatics and the UChicago Medicine pathology core that tracks testing and recommends treatments for cancer cases.

Scaling up cancer research: Last year, vice president Joe Biden visited the UChicago campus for the public opening of the Genomic Data Commons, a National Cancer Institute effort to combine cancer data from multiple sources and make it easier for researchers to extract discovery from that information. This section talks to GDC head and CI faculty and senior fellow Bob Grossman about the project, and visits Argonne National Laboratory to talk to CI Senior Fellow Rick Stevens about new initiatives to use those data for virtual studies on the supercomputers at Argonne.

Special agents to help speed research: Agent-based modeling (ABM) is used by movies and video games to create massive battle scenes and imaginary worlds that feel realistically complex. Now, ABM experts, including the CI’s Gary An, Jonathan Ozik, and Chick Macal, are developing new models to understand disease and find promising drug targets, as well as constructing models of the hospital itself to improve operations and simulate difficult situations.

A statistical crystal ball: Electronic medical records have made it much easier and faster to collect patient data in real time. However, quickly analyzing those data and turning it into patient interventions remains largely unexplored territory. With the UChicago Clinical Research Data Warehouse, physicians and data scientists (including the CI’s Anoop Mayampurath) can now develop algorithms for predicting cardiac arrest and conduct huge studies of sepsis, cancer, and other diseases using real patient data.

Data collection—home edition: Patients spend only a small fraction of their lives at the doctor’s office or in the hospital, and data from their activities outside of clinical settings also holds great value in assessing health and finding the right therapies. This section looks at how UChicago researchers are using Fitbit, smartphone apps, and voice recorders to collect data at home, fueling customized treatment plans and research studies.

About The Author

Rob Mitchum

Rob Mitchum is the communication manager at Computation Institute in Chicago.