Penn Research Team Receives $5 Million Grant to Use Big Data to Improve Health

PHILADELPHIA – A team of physicians and researchers across Penn Medicine and the University of Pennsylvania have been awarded a four-year $5 million grant to develop and test algorithms aimed at predicting adverse health events in real time. The project, Smarter Big Data for a Healthy Pennsylvania: Changing the Paradigm of Healthcare, is funded by the Commonwealth Universal Research Enhancement (CURE) program, and will examine the use of the algorithms in the hospital, at home, and in the community.

The researchers will develop strategies to pull and blend information from diverse sources – from insurance claims and electronic health records, to social media and patient surveys – in an effort to refocus the health care delivery system to a more proactive approach that anticipates clinical events before they happen, allowing for the potential to avoid dangerous and costly health events such as heart attacks and chronic disease flare-ups.

“Our overall goal is to make new data assets smart and actionable, combining mathematical, statistical, and health economics perspectives to change the way health care is delivered,” said Daniel Polsky, PhD, professor of Medicine at the Perelman School of Medicine at the University of Pennsylvania, executive director of the Leonard Davis Institute, and one of the project leads. Polsky pointed to the unusual combination of institutions, expertise, practitioners, data, and hardware brought together in one project as a major key to the success of the initiative.

Penn Medicine will provide the big-data engine through its Penn Signals platform, which has already successfully produced real-time heart failure detection and sepsis predictive algorithms during its piloting phase, and will allow for the scaled processing of millions of patient records. 

“Having the support of the health system leadership, and working closely with Penn’s Chief Data Scientist Michael Draugelis, has been critical to building a big data research agenda,” said co-lead Kevin Volpp, MD, PhD, professor of Medicine and Medical Ethics and Policy at the Perelman School and professor of Health Care Management at Wharton, and director of Penn’s Center for Health Incentives and Behavioral Economics. “This project is a robust collaboration between the data science team at Penn Medicine and our academic team of physicians and social scientists, and has the potential to benefit Pennsylvania’s most vulnerable populations by addressing health disparities and moving our health care system beyond the walls of the hospital and into homes and communities where the most vulnerable face the greatest challenges.

The project will deploy three separate research teams to explore predictive algorithms:

• The in-hospital team is led by Amol Navathe, MD, PhD, assistant professor of Medical Ethics and Health Policy, whose research focuses on the use of advanced health data analytics and technology to improve health care. Navathe will work closely with the Penn Signals team to develop prediction models of patients at risk for in-hospital complications of common surgical care, including gallbladder surgery, colorectal surgery, and total joint replacements.
• The at-home team, led by Mitesh Patel, MD, MBA, an assistant professor of Medicine and Health Care Management, will work to develop models to dynamically predict changes in out-of-hospital risk for 30-day readmission by monitoring medication adherence and physical activity in the home and through the integration of these home data sources with the data from insurance claims and the electronic health record.
• To monitor community health, the third team, led by Raina Merchant, MD, MSHP, an assistant professor of Emergency Medicine and director of Penn’s Social Media & Health Innovation Lab, will integrate social media data with statewide data to build and validate a tool to monitor and predict high-morbidity health conditions (heart disease, cancer, chronic lung disease, stroke, and unintentional injury) at a community level. This team will also build computational models to use Twitter data for monitoring and predicting dynamic public health events in the state (e.g., influenza, food-borne illness, infectious outbreaks, and acute environmental exposures).

The project also includes a training component to improve the pipeline of minorities entering science and public health careers who are equipped to use Big Data. It will provide opportunities for training in Big Data research through LDI’s Summer Undergraduate Minority Research (SUMR) program, now in its 17th year. Undergraduate scholars from Lincoln University, one of Pennsylvania’s historically black universities, will participate in this 12-week summer internship.

In addition to collaborators across Penn Medicine and the University of Pennsylvania, the team will also include researchers at both Temple University and Carnegie Mellon.