The volume of personal data kept by health care providers might give you pause, but all that information—ranging from data from smartphones, Fitbits and medical records, to an individual’s genetic and genomic predictors—is helping to shape major advances in digital health in Pittsburgh.
“The Next Big (Data) Thing,” an invitation-only event on April 18, gives the Pittsburgh Health Data Alliance a chance to share how its member universities, health system and tech industry are creating practical applications from data to improve lives and change the practice of medicine.
“Health care is changing dramatically and that will be reflected in the way that it affects patients or consumers,” says Joe Marks, executive director of the Center for Machine Learning and Health at Carnegie Mellon University. “There are some major trends that are creating research challenges, from volume to value-based payment, charging for outcomes rather than services, the move toward telemedicine, precision medicine—all of these are societal trends and they’re going to affect how you are treated.”
The alliance, or PHDA, was formed two years ago by the University of Pittsburgh, Carnegie Mellon and UPMC Enterprises. It aims to marry Pitt’s expertise in medical research, CMU’s world-class computer science and machine learning, and UPMC’s clinical setting and patient-focused analytics. Emerging projects “will give you a sense of how health care will be different in coming decades,” Marks says.
Pittsburgh has the components to become as dominant in digital health as coastal research universities, says Rob Hartman, program manager at UPMC Enterprises, the venture arm of University of Pittsburgh Medical Center, which funded the first six health tech projects.
“Our goal is to create technology that can be commercialized in digital health, and health care more broadly, that can elevate the Pittsburgh region as a leader in health care and health care technology,” Hartman says.
The daylong conference at the Westin Convention Center Hotel includes breakout sessions showcasing wide-ranging advances in precision medicine, consumer-driven health, health IT devices, care management, biometrics, and pre- and post-hospital risk management.
“What we’re trying to do is really accelerate what we call translational health data, going from basic scientific medical research to helping the researchers think about framing their research in the context of unmet market and clinical needs,” says Donald Taylor, co-director of Pitt’s Center for Commercial Applications of Healthcare Data.
Project PUMP, for example, stands for “pressure ulcer monitoring platform”—a set of sensors that go under a hospital bed’s wheels to measure a patient’s movements and ultimately reduce bedsores. Since treatment of pressure ulcers involves rotating the patient regularly, this could be a low-cost solution to a clinical problem that some analysts suggest costs hospitals $50,000 per patient to treat on average.
“This is an example of not throwing too much technology at a problem but just enough technology,” Taylor says.
In two years, the Pitt center has cultivated 82 projects, drawing in investigators from 25 departments across nine schools on campus. “It really involves transdisciplinary teams to work together,” he says.
As personalized diagnoses and tailored treatments become more popular, many people are aware of precision medicine. One of the PHDA’s newest projects—Phylogenetic Models for Predicting Cancer Progression—utilizes models, algorithms and software tools to pinpoint the origin and evolution of tumor cells and enable researchers to predict how a patient’s tumor will progress.
“Most people have some sense that genetic sequencing and analysis is important in cancer research, but there’s a big need in interpreting those results,” Hartman says.
An early-stage biometrics project seeks to understand intestinal activity by analyzing gut sounds. A non-invasive microphone would collect sounds from the intestinal tract to help physicians diagnose and treat gastric illnesses, Hartman says. “It’s an interesting concept that we’d like to explore.”
And what of all that personal data you put out there?
Patients enrolled in one study gave researchers permission to glean data about their social lives and activity levels from smartphones, Fitbits and medical records to predict their risk for readmission to a hospital. Research shows a correlation between activity and the risk for readmission during the month following a surgery, says Hartman.
“There are cognitive and behavioral reasons for why this may be true, as well as physical or activity-based reasons,” he says. “If you’re depressed post-op, that might show up. It’s also true that social support can help. If you’re having some sort of complication and you’re talking with your mother or your friends, those social interactions might help you get the support you need earlier.”