CMU, Pitt, and UPMC create Pittsburgh Health Data Alliance
Healthcare as we know it is about to change. The incorporation of computer technologies into medical care is fueling a revolution in the development of modern medical tools. Carnegie Mellon, the University of Pittsburgh, and the University of Pittsburgh Medical Center (UPMC) recently announced a partnership to combine their expertise and resources in order lead this transformation in healthcare.
This alliance will foster unique collaborations between doctors and computer scientists. Working together, they hope to develop novel ways to utilize the plethora of medical data to both combat illnesses and personalize medicine. The partnership will include two new research and development centers funded by UPMC with $10 million to $20 million per year over the next six years. The Center for Commercial Applications of Healthcare Data (CCA) will be based at the University of Pittsburgh and lead by Michael Becich M.D., Ph.D from the Department of Biomedical Informatics. Carnegie Mellon will be home to the Center for Machine Learning and Health led by founding director Eric Xing Ph.D, a professor of the Machine Learning Department. Researchers from all three institutions will collaborate in the two new centers forming an ecosystem of knowledge and innovation.
Xing envisions the center developing technologies that will enhance a doctor’s ability to provide a high quality of healthcare. “Currently, doctors act as the bridge between the patient and the body of knowledge. The Center wants to expand this bridge to become more data driven and community dependent.” Doctors will help gather the necessary information by asking the right questions and can use the new tools to aid in their decision making.
Innovators like Xing hope to ride the wave of wearable technologies as a means to gather rich and highly dimensional data. He hopes that one day we will have a device that instantly links patients to an ecosystem of doctors; a drastic change from the appointment-based system used today.
Carnegie Mellon’s focus will largely be in the information technology sector, developing ways to utilize data. Xing explained that currently, “Medical data has a very short life cycle; it is just stored and retrieved when necessary. The shift towards personalized medicine aims to use the data all the time. Your data will be used to enhance your experience as well provide clues about other people’s conditions. A large patient network can be used to make inferences doctors already do, but even faster and for larger groups of people. The future will be to use machine learning and artificial intelligence to make decisions. This center aims to lay the groundwork to develop these new technologies.”
As professor of machine learning, Xing has seen firsthand the advancements in the past decade that now enable computers to make inferences based on available data. Previously, computer science has primarily been used to develop operating systems and store data. “The maturity of the discipline has led to success in automation such as automatic speech translation and self-driving cars. Healthcare is ripe to be one of the next frontiers of innovation because it is high stakes and high risk,” Xing said.
Compared with other disciplines, making the correct decisions is extremely important because it directly affects people’s lives. A decade ago, the technology would not have been robust enough to be trusted to make these critical decisions. Xing’s past research involved using machine learning in a variety of fields related to medicine. One project focused on human genomic data by identifying variations in DNA that place individuals at a high risk for a given disease. So far this approach has been done in a weak statistical fashion unable to accurately identify the interactions between variations. In reality it is the combination of these variations that is iresponsible for some complex diseases. “Using machine learning we can build models using big genome data from patient populations and moving forward the goal is to create a personal model for each individual based on their genomic artifacts,” Xing explained. By using personalized genomic data, doctors and researchers can enhance and drastically change healthcare.
Another relevant project in Xing’s repertoire began as a computer vision project where the goal was to develop an algorithm to automatically detect an interesting event in a long video such as a surveillance tape. After this technology was developed, a doctor reached out to Xing and explained that in the ICU there is a critical need to monitor patients and also hold caregivers accountable. “For example, the technology could help ensure that the caregiver washes their hands thoroughly and frequently and can be used to protect against malpractice,” Xing said.
This center will allow innovators to provide doctors with their prototypes and doctors, in turn, will provide feedback to improve the technology. Both scientists and doctors will work to design the new technologies. Xing praised that the center enables interactions and collaborations with doctors that makes the partnership very unique. He noted that “the goal of the center is not to simply publish a paper or make headlines, but to make practical products that can be adopted by doctors and patients for a modernized healthcare.”