Lab tests are usually the first item ordered for patients, whether they are in the ED or a specialty clinic. Since most clinical decision-making depends on these results, visibility and timeliness is crucial. Laboratory information systems (LIS) have been around for years and represent a mature market in the hospital world. Today, just about every hospital in the United States has an automated lab system, and many are on a second or third generation. It seems when it comes to LIS, the biggest question for CIOs is best of breed or enterprise.
And right now, according to Jason Hess, director of research, clinical ancillary at Orem, Utah-based KLAS, the pendulum is swinging back and forth. With no real interoperability standard for lab and notoriously hellish interfaces, is there really a compelling reason for best of breed?
It depends on whom you ask.
According to Vince Ciotti, principal at Santa Fe, N.M.-based HIS Professionals, if it's the lab department that is making the decision, best of breed will usually be the first choice. He says that occurs more often in larger hospitals for a simple reason: revenue. “In bigger hospitals, pathology usually dominates,” says Ciotti. “A pathologist will generate $40 to 50 million a year in revenue [from the lab] and he'll do his own thing — and don't get in his way.”
In smaller hospitals, however, Ciotti says an integrated LIS is much more prevalent. “You get down to a 150-bed hospital and a pathologist is just one of the team,” he says. “He'll go with an integrated system because he's a team player.”
Dallas-based Baylor Health Care System, with its 14 hospitals, is a good example. CIO David Muntz says the organization's governance process encourages users to make decisions about which systems they want. Baylor's pathology department typically chose best of breed. Baylor, which currently has a Kansas City, Mo.-based Cerner lab system, recently committed to a new LIS from Clearwater, Fla.-based SCC Soft Computer. “If it was just an IS decision, I would have chosen to upgrade with Cerner, perhaps,” Muntz says. “But the users have to be satisfied with their product.”
Muntz follows a philosophy of focusing on the “customer's customers.” For lab, he says, that means not just pathologists, but the end users of the lab information — nurses and doctors. “If you don't focus on that, you won't be successful,” he says. “Get the end user involved.”
If PHRs gain traction, that end-user may soon be the patient. Like many hospitals, Muntz says Baylor gives paper copies of lab results to patients, but that is one area where laboratory systems have room for improvement. “We don't really address how to interpret what's on the page, and that should be part of the human engineering that goes on in laboratory roll outs.“
With so many hospitals like Baylor using best of breed solutions, interfaces become more important, but rarely less painful to maintain. Because there is no standard format, all codes are proprietary, and testing is required for each one. And, say many, it will only get worse. According to Hess, as closed loop medication administration becomes an industry standard, the strength and reliability of the LIS interfaces will be challenged even further. Hess cites the example of the nurse who has to wait for the results of a patient's sodium level before giving a medication. “That's a scenario where an interfaced lab may not have the same level of interoperability,” he says.
Another potential issue is pushing pathology data to the PACS system for anatomical pathology that includes images. “How much of that data is getting pushed to a solution that can be seen by the rest of the hospital?” Hess asks. Many say they feel that an interfaced LIS cannot have the access and transparency of an integrated one. KLAS, which recently reached out across the country to survey hospitals on LIS, plans to release its LIS perception report in November to help answer that question.
And when it comes to LIS, what many call integration often turns out to be anything but. According to Ciotti, if a vendor has true integration (even at the cost of reduced functionality) the CIO is lucky. Because so many EMR vendors have been on a buying spree, Ciotti says many of those “integrated” systems are in name only. “I call that ‘interfarce,’” he says, “when a vendor buys a lab system and claims, ‘We integrated the brochures, the PowerPoints — everything but the systems.’”
He claims the larger vendors, who have acquired niche LIS, are building interfaces under the covers and calling it integration. Ciotti says that under those circumstances, a CIO really has no control of the interface. “That's what drives a CIO nuts, because the vendor is totally responsible for that.”
At Baylor, Muntz has Atlanta-based Eclipsys for his EMR, and interfaces the LIS using a Quovadx (Irving, Texas) Cloverleaf interface engine. He says the challenge of the LIS interface is not on the technical side. “We can write interfaces very easily,” he says. “The challenge is getting the users who are involved in the validation process to be able to share their time with their regular responsibilities and then to be able to test it.” He says validation and testing is particularly troublesome for the lab because they provide about 75 percent of the data on which clinical decisions are made, and that's a lot of testing.
Peggy Donovan, CIO at two-hospital Olathe Medical Center, Kansas City, has been on both sides of the fence when it comes to choosing a LIS — prior to her CIO position, she came to Olathe in 1989 as lab director, and her first job was to find a LIS. Now on its second generation LIS, in 2002 the hospital decided to go with Cerner's Millennium EMR, and in 2003, it went live on Cerner's PathNet LIS.
Donovan says the multidisciplinary team that included clinicians and executives made the decision to go with a patient-centric, integrated solution. “That quickly narrowed the field,” she says, “because there really weren't that many truly integrated products on the market. It was an organizational decision, and it's proven to be a sound solution.”
Donovan says one of the big wins of her LIS is auto verification, which checks if samples are in the normal ranges — and normal for the particular patient, too. “We built logic within Cerner Millennium so those normal results are released,” she says. “If you pick a LIS that is a robust solution, you can optimize and individualize for your own facility.”
Another win for Olathe is in the ED, Donovan says, where physicians order labs and click on an icon to see results. And since her second hospital is also on Cerner, Donovan says that transfers into the ED can gain valuable time for the patient. “Since we're on the same system, if a patient gets transferred to our ED in the larger hospital, the ambulance pulls up and they're already doing the second screening because the lab results are already there.”
Both CIOs say that by using physician portals, their docs can securely view lab results from anywhere. Yet there's more afoot in the LIS arena than the interface or integration questions. Some new lab technologies offer promise, and yet still more challenges.
At Baylor, Muntz says he is excited about the prospects of genomics, proteomics, molecular flow and similar new disciplines. “They are going to present new data challenges because of the volume of data we have to deal with,” he says, “not to mention all the ethical questions when you start to deal with genetics.” Muntz says trying to keep the expression of a gene, and the sheer number of pieces of data it will require, is far removed from keeping a blood count. “We're taking these exotic things and they're going to become ordinary, routine and practical,” he says. “That's the real goal of automating.”