UPMC on Monday announced a $100 million initiative aimed at mining and analyzing data in order to provide more precise care. It's a five-year plan that touches every piece of the enterprise: research, bedside care, data and accounting.

Lisa Khorey, vice president of enterprise systems and data management at UPMC, explains the various facets of the project.

[See also: UPMC launches $100M personalized care initiative]

1. Genomics: “One of the five elements of our enterprise analytics program centers on personalized medicine or specifically genomic medicine. Our first use case, though, is much more specific than just genomics. It centers on breast cancer. So, if we bring together what we know of our patients who have consented and have had a genetic scan, and we have their electronic medical record, what can we infer or detect about the treatment of those patients. What is the least toxic treatment that can be given, and what is the variation between those patients that might be an indicator for a different treatment path based on individuals."
2. Outcomes: “In addition to that, though, as part of our program, we are not only looking at what’s the genomic interplay with data from the electronic medical record, but also what are the other aspects that make up – what is personalized medicine to an integrated delivery/financial system – to UPMC. Some of that comes to clinician performance. How are patient populations faring. What are patient outcomes? How do we compare ourselves within populations, across providers, across organizations, to a national baseline to a big focus on patient outcomes and understanding our care delivery.”
3. Data: “Another leg on the stool looks at data mining and finding answers to questions that can influence and change, perhaps, the way we look at medicine, although not necessarily at he point of care. So more hypothesis kinds of testing, like what clinical indicators set a person on a path, what is out of control versus well managed, when is Life actually threatened. What clinical markers can we derive? What patterns exist? What can we lift out of all the data that’s known?”
4. Accounting: “The other two use cases, or legs on the stool, really center around horizontal accounting – so just looking at the cost of care across the total dollar. We seek to understand not just how much things cost, but what has marginal value and where are inefficiencies and every patient having everything that they need to have the best possible outcome. A lot of that centers around the best care management that we can possibly surround our patients with.”

“Those facets of personalized medicine are really the centerpiece that we are building here based upon all our integrated medical health records implementations that have gone on in the last 10 years and the 3.2 petabytes of data that live inside UPMC today, which is vast. So, our intention is to organize and harmonize that into an analytics space with Oracle and with other partners, [not just] to really be able to answer some of these questions and to measure some outcomes, etc., but really to drive toward something much more extraordinary, which is: How can we actually change care delivery? What could we change or derive or know about medicine that just isn't known today?"