The University of Pittsburgh Medical Center said this week that it plans to invest $100 million over five years in informatics infrastructure, training, and support aimed at making personalized medicine a reality.
Over the next two years, the center will install hardware and software to create a data warehouse that will combine clinical, financial, administrative, and genomic data from more than 200 sources of information across UPMC, the UPMC Health Plan, and outside entities, including labs and pharmacies.
To that end, UPMC has tapped four technology partners — Oracle, IBM, Informatica, and dbMotion — to provide the tools needed to build its warehouse. These include analytic and predictive modeling applications for clinical and financial decision-making, which are expected to help improve patient outcomes, enhance research capabilities, and reduce costs, among other gains.
The new infrastructure will provide the tools needed to adopt a “comprehensive analytics approach” that will “enable us to treat each patient in a personalized way to produce the best possible results,” Steven Shapiro, chief medical and scientific officer at UPMC, said in a statement.
It also provides an opportunity to apply healthcare analytics best practices learned as part of a collaboration with the Carnegie Mellon Software Engineering Institute, Lisa Khorey, UPMC’s vice president of enterprise systems and data management, added.
“We’re now ready to turn what we’ve learned into action to support our goal of developing new models of affordable, effective, patient-focused health care,” she said in a statement.
The new warehouse will also solve an existing problem at the institution —integrating data stored in silos spread across the UPMC health system, Khorey told BioInform.
For example, “we have lots of great data warehouses and electronic health records [across UPMC] that are up and running and are crucial to our operations today,” she said. So “hardware and appliances that we purchased from IBM and Oracle are intended to help those existing warehouses function in a more efficient way.“
This infrastructure will also provide the computational heft that the center needs to process, analyze, and derive insights from patient data that could lead to improvements in care and reduce the time required to move research findings from the bench to bedside, she added.
She also pointed out that each hardware company offers specific capabilities needed for the warehouse, which informed UPMC’s decision to work with multiple IT partners.
The first beneficiaries of the system will be members of UPMC’s breast cancer community, according to Khorey, who will use the system to address questions around quality measures, patient outcomes, and the cost of care as well as to integrate genomic and electronic medical record data.
Cast of Characters
The exact terms of UPMC’s agreements with its tech partners aren’t being disclosed, though Oracle and IBM have provided details about which products in their portfolios the institution will use.
According to Oracle, the center will install applications such as Oracle Enterprise Healthcare Analytics, Oracle Health Sciences Network, Oracle Fusion Analytics, and components of its Fusion Middleware that will support cost-based accounting and ensure regulatory compliance, as well as protect sensitive information.
UPMC will also use Oracle’s Exadata Database Machine as the hardware foundation for an analytical environment that will support activities at its new Center for Innovative Science — a $300 million structure that is expected to be completed in 2014 — which will focus on personalized medicine and gaining a better understanding of cancer and aging to improve patient outcomes and reduce over-diagnosis and unnecessary treatments.
Meanwhile, UPMC has tapped IBM to develop a predictive analytics strategy based on IBM’s Business Analytics services and software.
According to IBM, UPMC will use the IBM Healthcare Data Models and IBM Business Analytics and Optimization software. It will also use the IBM Initiate software for all patient identification, resolving patient identity issues, and improving its search capabilities.
For their parts, Informatica is providing an integration platform that will help primarily with moving data, while dbMotion will provide an interoperability framework that allows the aggregation of clinical data and applies semantic harmonization to the data so that “we can aggregate the longitudinal patient record across organizations and/or normalize the dataset so that we can then feed it into the analytics environment in a meaningful way,” Khorey told BioInform.
In addition to hardware and software, UPMC’s planned investment will also cover internal salary costs, training, and other operational support for its system. Overall, the institute expects to spend roughly 60 percent of its investment on capital and about 40 percent on operating costs.
Meanwhile, the center will be keeping an eye on tech trends in the healthcare and life sciences space — for example new developments in genome sequencing and software markets — that will inform infrastructure spending during the last three years of its initiative, Khorey said.
“The rate of change in technology right now and in this space in particular is so great [that] we didn’t feel we could make informed choices today about what we might need or want in those later years, so we very purposefully left that open,” she said.