Aggressive TreatmentAggressive Treatment

The health-care industry, which has stubbornly held on to its paper-burdened processes and disparate software-application silos, is trying to figure out how to become an integrated world of digitized information that leads to better patient care and lower costs. The University of Pittsburgh Medical Center started down a new path to get there last week when it struck a far-reaching, $402 million, eight-year deal with IBM.

The most unusual part of the deal has UPMC and IBM each committing to invest $25 million--and up to a total of $200 million during the life of the contract--toward joint development of software and systems to tackle major problems facing health care, including IT tools for coping with epidemics and bioterror attacks, researching and treating cancer, and using electronic health records. If it works as planned, the medical center and IBM will create IT systems and innovations that help deliver health care and insurance better and then sell the jointly developed products and share the profits. "This is a landmark relationship," says Dan Drawbaugh, CIO at UPMC, which operates 19 hospitals in the region, along with a physician group, health-insurance company, and other health-related businesses.

IBM has a lot at stake as well, since the deal represents one of the biggest tests of a nascent business strategy, hinging on what it calls "business-transformation services," exactly these kind of deep, industry-specific deals that involve a wide range of products, consulting, and research. This is IBM's biggest deal in health care.

The more conventional part of the deal--and the largest, valued at $352 million--is moving UPMC to IBM infrastructure, increasing capacity while consolidating its IT operations. IBM will help the medical center go from 40 storage systems to two, while increasing capacity 25% a year to keep up with the ballooning quantity of digital data UPMC will produce. By the eighth year, UPMC expects storage capacity of 825 terabytes, up from 180 today. It expects to go from 162 Unix servers to 61 new ones and from 624 Windows/Intel-based servers to 244 new systems. IBM will bring in 24,000 new PCs and refresh all hardware every three years. The medical center will keep its staff and run its own IT operations but turn to IBM consultants to improve its IT management and software development. The companies predict cost savings of 15% to 20% in IT operations.

The larger impact from the deal will come if, through co-development work, UPMC and IBM can move the health-care industry forward in its use of IT. However, exactly where that investment will land isn't spelled out. Wireless and radio-frequency ID applications in hospitals are a possibility, Drawbaugh says, along with bioterror tools, cancer research, and electronic records.

UPMC has some experience funding companies based on technologies developed there, notably Stentor Inc., which makes digital medical-imaging systems. The most likely first commercial products are tools for dealing with bioterrorism and for helping regional health-care providers electronically share patient data, predicts Dan Pelino, IBM's VP of global corporate marketing.

With this deal, UPMC is pursuing a different approach to electronic health records from most other companies taking a lead in IT health care, many of which are aligning themselves with regional efforts. Such alliances often bring together competing hospitals, medical groups, and insurance plans to work on tests related to sharing clinical information that resides on a variety of systems and platforms. A notable example is the Massachusetts eHealth Collaborative, a group of 34 health-care providers, health plans, and insurers that's launching pilot programs to get three communities using and sharing digitized patient information. Blue Cross Blue Shield of Massachusetts provided $50 million to start that effort.

But UPMC's size and scope makes it a regional medical community of its own in western Pennsylvania, providing care and insurance to patients across 29 counties, from doctors' offices to cancer centers to in-home care and nursing homes. That size means any success the medical center has can be viewed as a "center of evidence for development of these technologies to prove they can work on a national, and even a global, level," Drawbaugh says. UPMC will use electronic health-record software from Cerner Corp., which already is deployed at some of its facilities; UPMC is expanding a co-development effort with Cerner.

Another potential area of collaboration is in combining the medical center's clinical knowledge in areas such as cancer research with IBM technologies such as supercomputing and analytics. In the past year, IBM has forged other alliances in this kind of research to develop personalized health care with some of the country's most-prestigious health-care providers, including Cleveland Clinic and Mayo Clinic.

Mayo Clinic researchers are working with IBM researchers and consultants to map out ways to use patient data to improve diagnosis and treatment plans for some of the most-resistant forms of cancer and other diseases. They've cross-referenced 4.4 million patient records to look for significant patterns. The work between Cleveland Clinic and IBM includes data-mining projects aimed at improving treatment plans for patients with potentially deadly cardiac and other conditions.

IBM last week also disclosed that it's establishing a test lab linking its facilities in San Jose, Calif.; Rochester, Minn.; and Haifa, Israel, to simulate moving standardized medical records. The company also acquired Healthlink Inc., a provider of health IT services and consulting.

But even if UPMC and IBM produce viable products, they may face a reluctant and cash-strapped market. Average IT budgets are only 2.5% to 3% of revenue for most health-care providers, says Gartner analyst John David Lovelock; those spending 4% to 4.5% are considered aggressive.

Ron Strachan, CIO of HealthEast Care System, a St. Paul, Minn.-area hospital and clinic group, knows the budget pressures well. He sees why a research group like UPMC might need extensive data sharing to further its mission. But for most patients, who stay within one care environment unless they move away, the critical shared data could center on what an emergency doctor might need, such as knowing a patient is diabetic or has a heart condition or a drug allergy. Off-the-shelf software will deliver that capability, he says.

Still, Strachan is more concerned about an issue that UPMC's all-under-one-roof approach doesn't address: "In large part, health-care organizations, plans, and insurers aren't willing to share the data that would make these efforts successful. It's a matter of trust."

Clearly there are major technology problems begging for breakthrough innovations. Several technical hurdles stand in the way of the Bush administration's goal of having digital records for most Americans by 2014, said Dr. David Brailer, the administration's point man on electronic medical records and health-care IT coordinator, at an IBM event last week. Consumers need health information that's portable and private, factors that can be contradictory. "There are no right answers. We don't want to see a monolithic response," Brailer said. "We want to see a set of responses."

UPMC's deal with IBM represents one large health-care system's response. Will this pact revolutionize the delivery of medicine in western Pennsylvania? Will it be a model others in the industry will buy? Even as the government pushes the industry to develop vendor-neutral tech standards for a national health infrastructure, there's a pressing need for real-world examples to test the theories and speed adoption. Says Drawbaugh, "Standards are being developed, but at the same time, the implementation of standards needs to occur."

--With Charles Babcock