Healthcare Collaborative, IBM Partner On Big Data Platform

7 Big Data Solutions Try To Reshape Healthcare

Big data analytics is set to make another big mark on healthcare. IBM, one of the leaders in the space, has partnered with the Premier Healthcare Alliance and four healthcare systems to develop the kinds of analytic tools that will be required for population health management.

Dubbing itself the Data Alliance Collaborative (DAC), the group in addition to IBM includes Carolinas Healthcare System of Charlotte, N.C.; Catholic Health Partners of Cincinnati; Fairview Health Services of Minneapolis; and Texas Health Resources of Arlington, Texas. All four of the systems are members of Premier's ACO Collaborative, which includes 79 healthcare organizations in 40 states. DAC is a subset of the ACO group.

Premier chose IBM as its partner in the venture partly because it already uses a lot of IBM hardware and software in aggregating member data for benchmarking and other purposes, said Keith J. Figlioli, senior VP of healthcare informatics for Premier, in an interview with InformationWeek Healthcare. In addition, he noted, IBM and Premier codeveloped "an integrated payer and provider data model that can become extensible."

[ Effective use of big data in healthcare starts with teaching it in medical school. Read Big Data Use In Healthcare Needs Governance, Education. ]

The IBM data model, or data infrastructure, is the platform that DAC members are using to develop their analytic applications, said Figlioli. The platform is capable of integrating clinical, financial and claims data across care settings, regardless of what kind of system generated the information, he said.

This infrastructure is a kind of advanced data warehouse, based on the IBM Netezza data warehousing appliance. Some DAC members have either never had a data warehouse before or have repositories that are narrowly focused on particular kinds of information, Figlioli said.

Carolinas Healthcare System has a legacy data warehouse that cannot be used to run reports expeditiously, Allen Naidoo, the organization's VP for advanced analytics, told InformationWeek Healthcare. In contrast, he said, the IBM data model is very flexible, allowing him to "get near real time data" and to run reports "on the fly" with Netezza, "which makes processing go at lightning speed."

This sounds like real competition for Health Catalyst, a data warehousing company that recently has gained a number of adherents and investors among large healthcare organizations.

Carolinas Healthcare System, which has formed an accountable care organization, recently began receiving claims data from health plans and employers. With the help of Netezza, Naidoo said, this will provide the system with a "360-degree view of the patient. For example, I can tell from EHR data whether the physician has written a script. From the claims data, I can tell you whether the patient has filled that script. So, if the patient is not compliant with a medication, we'll know that from the value brought by these two data sets."

DAC's members hope to use a similar approach to improve transitions of care. For example, one of the first co-development projects is a tool designed to quickly notify providers about groups of patients who have not filled prescriptions within 24 hours of discharge. This tool might take advantage of Surescripts' community medication histories or other data sources, Figlioli said.

DAC members also are co-developing an "all-cause predictive readmissions model" that analyzes both EHR and administrative data to identify the patients who are most likely to be readmitted before they're discharged. The DAC model will identify risk factors leading to readmissions, tying patients to appropriate evidence-based checklists based on their condition.

IBM cannot commercialize anything developed by the collaborative, said Figlioli. However, he noted, a vendor might sell applications that it develops for the platform itself. For example, he noted, "One leading system integrator is creating assets for this group that will later be commercialized if somebody consumes them."

That integrator is Perficient, a St. Louis-based firm that has built a benchmarking application for PremierConnect, a "business intelligence, data warehousing and social networking solution" unrelated to DAC.

What DAC members come up with won't be proprietary, Figlioli said. "They'll be completely transferable assets. The idea is that there's a co-development agreement with all the data alliance members, and the data alliance will get bigger over time, and the intent is to include all of the trading partners." Within six months, he added, he expects at least four more organizations to join DAC.

From the perspective of Carolinas Healthcare System, Naidoo said, the most important advantage of the collaborative is that it enables the partners to learn from one another as they develop the population health management tools they need. Moreover, he added, IBM has begun to provide DAC members with insights on how companies in other industries, such as banking and financial services, use analytics.