College of American Pathologists
Printable Version

  Feature Story


cap today

Everyone into the pool—a ne approach to data

May 2004
Ed Finkel

Thirty-five years ago, each major bank in the United States had its own credit card—and they were all losing money.

So the banks decided to collaborate and compete at the same time by creating the Visa card, for which they would share the costs of development and operations while separately marketing their own affiliated cards. Now they share a $3 trillion business as part of a for-profit collaborative.

That account roughly parallels the benefits laboratories and other partners could gain from the national clinical electronic network for medical information under development by the Patient Safety Institute, or PSI, says Johnny Walker, chief executive officer of the Plano, Tex.-based nonprofit. The independent, secure PSI network will give providers, labs, and patients access to a consolidated patient view of clinical information from multiple sources.

"The data will not reside in a central repository. The data will reside at its originating source," says Walker, who expects it will take "five to six years" for the network to cover the nation. "PSI maintains the giant master patient index, which knows where all the data is and all the different record numbers at various facilities, and it provides the network linking capability."

The network provides access to recent information such as laboratory tests, medications, diagnoses, allergies, and immunizations, Walker says. "Data that exists on a patient is gathered up from all the places where there is electronic clinical data of high value to providers," he says. "With patient permission, received with the swipe of a card in the hospital or physician’s office, the physician gets a comprehensive patient-centric view of the patient’s clinical record in four to five seconds."

Hospitals and other facilities typically will integrate the network into their existing computer software, Walker says. While PSI can and does provide an operational view of data where required, its "main focus is not providing a software product for people to use; it’s providing a secure third-party clinical network that everybody can access like the Internet," he says. "PSI provides patient data from outside the organization and does it by enhancing existing software applications and systems rather than requiring that they be replaced in order to gain connectivity."

The Patient Safety Institute, a national organization that brings together patient, physician, and hospital leaders, recently finished a demonstration of the electronic network at the three Swedish Medical Center facilities in Seattle, Walker says. Begun in December 2002, "It was a rousing success. Patients universally opted in, [and] the hospital didn’t want to give it up, partly because of its value but also partly because physicians said they were going to walk out if PSI was taken out."

The network’s next expansion will be in Delaware, where physicians, patient representatives, payers, the state government, and the state’s eight hospitals are "finalizing the plan for a statewide rollout" that will occur later this year, Walker says. "We are in the planning stages to do something with a number of other states that are in the process of securing their financing. Financing is the hurdle. Delaware has some of the financing, and they feel comfortable they will shortly secure the rest. They have gotten broad support from foundations, employers, payers, hospitals, physician groups, and the state government as well as the federal government."

Physicians have been supportive for a number of reasons, Walker says, ranging from "ease and speed of clinical information access" to the ability to view such information in their home offices, which is attractive to "physicians striving for that life balance with work and home." Most important, he says, "The consumers have universally opted in. The consumer opt-in option is one of the PSI principles. We don’t want them to be co-opted and then have to find a way to extract themselves from some system if they don’t want to participate."

Walker sees several reasons laboratories would want to participate:

  • Competitive advantage. "They either want to get one up on the other one [competitor], or they don’t want to get left behind," he says.
  • Connection to providers. "It’s an incredibly inexpensive connection to their large and growing base of providers," he says. "Better to be at the front end than the back end."
  • Chance for lower costs. "It’s a very large potential decrease in their network cost structure," he says.
  • Patient-centric viewing. "It’s an opportunity for patient-centric viewing of lab data—companies currently don’t have that," he says. "Most are limited to physician-centric views, and this limits direct-to-consumer offerings."

Access to a single view of results from different organizations means dealing with multiple lab standards. "This doesn’t create a problem for physicians reviewing results if the relevant lab range is provided with each result," Walker says. "But multiple standards do negate the computer’s power to provide decision support, graphs, alerts, and the like."

Thus, the need for Logical Observation Identifier Names and Codes, or LOINC, an extensive standard coding system for identifying lab test results. "We need to get on a standard," he says. "As a result, we are seeing a real push now from people who have access to this data saying, ’Now I have a reason to get with the hospital and get them on the program to convert to the industry LOINC standard. PSI provides the value proposition to convert from their proprietary or old standard to the evolving industry standard. We’re seeing large LOINC interest being birthed." For those lab companies that want to provide or leverage the patient-centric view, he adds, "they have the problem that we solve, which is lots of places and organizations where the same patient is identified under different numbers, and therefore not recognized as the same person."

Labs that deliver this unified, patient-centric data first will be the preferred vendors, Walker predicts. "There’s absolutely a first-mover leverage opportunity for those with enhanced capabilities, leveraging the ability of this universal cross-patient-centric view for additional client value and additional market share, while certainly reducing the cost of delivering it," he says. "As lab data proliferate, firms that make distribution the least expensive are going to be the ones favored by the payers."

PSI has announced national partners such as Rx Hub, Laboratory Corporation of America, and Quest Diagnostics as well as additional hospital labs and regional providers, Walker says. On the political side, the network has received endorsements from the Western Governors Association and the Democratic Leadership Council and support from the Department of Health and Human Services. Some of the federal funding has come through the Department of Homeland Security for bioterrorism preparedness.

"There’s not a lot of money floating around to build in-state clinical networks, but it is growing," he says. "We’re finding that states are very adroit at saying, ’I’ll build a network that delivers both bioterrorism network surveillance and clinical data because PSI can do both.’ The more creative states are finding multiple ways to skin a cat, including sharing between states."

Once it becomes more of a national network, Walker says, the states will charge a transaction fee based on usage similar to fees collected by Visa and other credit card companies. PSI’s research has shown the total cost to build the network will be about $2 billion, while the value to payers—private insurers, Medicare, Medicaid, self-insured employers, capitated providers, and the uninsured—will be about $40 billion per year.

"It’s not a hard economic value proposition. It’s a timing issue: How do I get the $2 billion so I can collect the $40 billion?" Walker says. "And everybody wants somebody else to do it. The challenge is to create a toll-road concept." He adds, "We’re trying to allocate the costs equitably among all the projected users. The states get a say in how they want to allocate it within their state. PSI will work with them to provide national guidance. . . . The stakeholders in each state will ultimately make the decision."

Walker says in addition to spreading the network across the country, PSI has at least two enhancements in mind for the system in the coming months and years: to give patients the ability to access their own data—which PSI hopes is an early part of the Delaware system once it’s up and running—and to give physicians, labs, and others the opportunity to convert historical data to LOINC.

"The PSI network will be able to allow the patient to choose his own portal of choice—that could be a lab company, for that matter, or a hospital, or whatever—and then see the data where they want to see it," he says. That gives the patient the ability to "append additional data that would become part of the record that would be submitted to the doc as a patient note on items that were omitted or possibly incorrect."

This ties into PSI’s patient-centered philosophy, Walker says. "If patients are not able to see their record or make decisions on it, why should anyone give permission to access their record?" he says. "The patient now has the opportunity—the lab has the opportunity in working with the patient—to present a much more comprehensive view."

PSI does not have immediate plans of its own to convert the historical data to LOINC, but Walker figures another entity could. "In the near term, there is an opportunity for somebody to figure out how to convert historical data to LOINC, just like they’re doing with the ongoing current data—and that will be a very attractive [product] and something people will pay for," he says. Without that piece, "it’s still not immediately attractive. With current lab data converted to LOINC, the real value of standardization for decision support, graphing, and integration only comes into play when the historical data is also converted to LOINC—or time elapses till the historical data is no longer relevant to the user."

Ed Finkel is a writer in Evanston, Ill.