February 2012
Feature Story

Anne Paxton

Dr. Kandice Kottke-Marchant at the January opening of the new laboratory building at the Cleveland Clinic, where she and others are projecting a fourfold increase in their reference testing business by 2018.

Sidelining expansion plans, torpedoing new equipment purchases, and handing down budget ultimatums. When lean times loom or spending outstrips revenue, and hospitals seek ways to stem red ink, these are often the strategies of first resort.

The task of getting reference testing costs under control, however, may call for a different game plan. Reference testing is a $6.2 billion market, growing about eight percent per year, and many hospital laboratories cite the taming of reference lab testing costs as one of their leading sources of frustration. At the 20-hospital University of Pittsburgh Medical Center Health System, for example, the send-out genetic testing bill at just the core academic hospitals has risen dramatically in the past five or six years and is well into seven figures, says Jeffrey Kant, MD, PhD, director of the Division of Molecular Diagnostics in the Department of Pathology.

“These new expensive tests are increasingly driving it,” he says. “The diagnostics companies market directly to physicians to ‘develop the market’ and get an early return on their investment. It’s a lot of money, and depending on insurance coverage, the hospital may frequently be stuck with the bill.”

But some hospital laboratories, both large and small, say they are reaping savings from kinder, gentler approaches to the problem of controlling reference testing costs. Among their strategies: education, conversation, and, in the make-versus-buy choice between doing tests in-house or sending them out, sometimes finding a middle way.

Variations on those strategies have saved UPMC hundreds of thousands, likely millions of dollars in reference testing costs over the years, Dr. Kant estimates. When he arrived in the mid-1990s, molecular diagnostics was a single division but not highly integrated operationally. “We were organized a little like Yugoslavia in the sense that each area of emphasis—for example, microbiology, genetics, etc.—was freestanding and had its own personnel and in some cases its own location.”

Over time a number of these units have been consolidated into one clinical laboratory under a single system of oversight, operationally integrated into one 5,000-square-foot space. “My hospital vice president asked me to take an active role in overseeing test utilization, including our send-out testing, and recently we’ve begun to take a comprehensive review of our genetic testing experience, primarily our hit rate, in terms of how often there is a positive result. We then break it down by disease, by ordering service, even by the clinician who is ordering the test,” says Dr. Kant.

“The majority of the hospital’s genetic reference testing orders, about 75 percent, come from our children’s hospital,” he says. Some areas are higher users than others. “So we have conversations with clinicians and genetic counselors periodically, and we’ve gotten them over the years to adopt a tiered or sequential approach to testing. Interactions are bidirectional and occasionally we’ll have cases where the clinician says, ‘I really need all three of these tests at once,’ so we just do it.”

“But we also get requests for large gene panels where within that panel a single disorder or gene encompasses a plurality of cases. So we’ll call up and say, ‘Did you realize 80 percent of the positives in this disorder come from one test—which, by the way, we can get for you for six percent of the cost of the panel. How about we start there?’ And invariably they say, ‘Okay. That makes sense.’”

His laboratory makes an effort not to control but to consult on this testing. And that approach is having an impact. “When we say, ‘Why don’t you do the $500 test instead of the $10,000+ panel,’ each one of those conversations is $10,000 to the bottom line, and over the years that mounts up.” Other areas of the hospital that Dr. Kant doesn’t directly oversee are taking the same approach to solid tumor testing for adults, he adds.

Educating physicians about the panels is one part of the task; another part of utilization management is selecting the best value where there are several available options, Dr. Kant says. “So we’ll research the disease and say, ‘What’s really important in testing for this disorder are the following elements,’ and we’ll sometimes compare the different offerings of several reference labs.” In a case last spring, the laboratory priced two or three different labs, finding that the charge for a single specific test ranged from $1,150 to more than $4,000. “But essentially you were getting the same thing at all of those labs.”

As part of an academic medical center, the UPMC lab has a broad spectrum of responsibilities, but it is able to draw on pathology house staff, resident and fellow physicians, to do most of the legwork of collecting and comparing price information, which Dr. Kant says provides valuable experience and role models. “Rarely do you get a disorder where there is only a single lab offering a test, so after we have done our research for the first request, when we get subsequent orders for analyte X, we have a database that contains our preferred vendors, and all tests for analyte X will go to that vendor. UPMC has entered into a joint laboratory venture with Quest Diagnostics, and that’s typically where we look first.”

Calculations of return on investment are a key part of decisions on whether to bring the highest-volume genetic sequence assays in-house, Dr. Kant says. “When we do our capital budget requests and propose a new piece of equipment to bring a test in-house, hospital administration may have targets like six to 12 months to recoup the investment in an equipment purchase. That would obviously be a higher priority than an instrument that required five years to pay off. In fact, we are negotiating right now for a sequencer with our administration.”

In the case of some inherited genetic diseases, he says, ”you’re often confirming a fairly convincing clinical diagnosis. Is that a good expenditure of money? I could see where people might debate that, and sometimes it’s hard to know.”

With cancer, “there’s a preoccupation with biomarkers that have an impact on the prognosis of the patient, and to the extent that these markers allow you to target a particular patient with therapy, or save money from an expensive therapy, they are extremely useful and I would say cost-effective. If a new test allows you to prolong the life of a patient, that’s a win. Or if you don’t have to spend a huge amount of money on a therapy that doesn’t work, that’s a win.” However, many other tests that have been done for years are not being discontinued, “so any new testing tends to be aggregated on top of existing testing, and that tends to drive up the cost of care for patients.”

Improving the laboratory’s information system has been key to refining implementation of UPMC’s molecular reference lab strategies, Dr. Kant notes. “We switched our genetics and microbiology services a little over a year ago, with one objective being to enhance the ability to track test results, including reference testing, with more granularity. Our text-based LIS was used for anatomic pathology and didn’t allow one to easily segment things, or easily count positives and negatives to look at our experience.”

Now the laboratory is able to pull information more efficiently from the LIS, he says. “If you send out lab tests and you get one to two percent of patients positive, that’s the kind of test I can take to clinicians and say, ‘You’re getting two positive patients for every 100 you’re testing. Why do you think the yield is so low? If it isn’t a screening test, their clinical criteria might be too broad for ordering the test.”

Dr. Kant is also pursuing this strategy actively with hematologic oncology molecular tests. “We have certain tests where we’ve been looking at things like the blood count of patients, and we don’t see a strong indication. We’ll call and ask why they are ordering the test, and they’ll say, ‘We know it’s not very likely, but we’re just trying to dot all the I’s.’” Whether that type of approach is a good way to practice medicine is an important question, in Dr. Kant’s view. Clinicians may be able to say, for certain individual patients, that indications are appropriate. “But in taking a close look at each case there is an iterative process that lets ordering criteria be tightened up over time.”

North Shore-Long Island Jewish Health System in New York, the second largest secular health system in the country, performs more than 160,000 surgical pathology accessions and millions of clinical laboratory tests, but it is still seeking ways to pare the costs of reference lab testing. “We have a strategy to guide reference laboratory testing, and are now working with our clinical colleagues to implement it,” says Jordan Laser, MD, director of molecular pathology.

All reference testing is funneled to the core laboratory’s dedicated Reference Center, which distributes the tests to different outside labs, Dr. Laser says. “These are tests that we don’t perform in-house due to either intellectual property or volume reasons,” Dr. Laser says, noting that because of patent and licensing issues, “There are some tests I’d love to bring in, but I can’t.”

He outlines a three-part strategy, which boils down to determining the appropriate setting for the test. “First, I think it’s very important that the decisionmaking process is patient-centric, that we’re getting all the data our hospitals and practitioners need to manage the patient appropriately. The second point—although ultimately the goal is to decrease costs—is to be cost-blind, ironically. You should ask yourself: Is the result of this test going to have an impact on inpatient management, or will delaying the test to the outpatient setting put the patient to any kind of harm? And if the answer is yes, regardless of cost, that test is going to be sent out.”

But the laboratory’s control over test ordering can be viewed as a threat to, even an attack on, the ordering physician’s autonomy, and a third component involves being aware of that. “As physicians, we’ve been trained to do whatever we think is appropriate to treat a patient,” he points out. “So it’s critically important to stress that this decisionmaking process is not being taken away, but overseen.” Without physicians onboard with the process, he adds, “it’s going to be a brutal, long, and possibly unsuccessful initiative.”

Impending changes in the national health care system and the way hospitals will be paid make reference lab strategies a necessity to curb costs, Dr. Laser believes. “With the explosion of tests, it’s difficult to keep track of what tests are appropriate to order.” For example, he notes, there’s a qualitative BCR/ABL, a quantitative BCR/ABL, BCR/ABL for p190, BCR/ABL for p210, and BCR/ABL tyrosine kinase inhibitor mutation analysis. “For someone not that savvy, these tests all sound similar and they might consider them equivalent. The fact of the matter is they aren’t. Each test gives a different type of result, addresses a different disease entity, or predicts a need to change therapy. It’s very easy to order the incorrect test and not get the results you need for patient care. So we need to have some system overlooking these send-out tests.”

For their part, the reference laboratories are doing their best to keep business brisk. For example, as a leading provider of reference testing, the Cleveland Clinic is banking on a significant expansion in its reference testing over the next seven years. “I think the trend is that hospitals are less likely to be doing some of the cutting-edge molecular testing and will be sending those out,” says Kandice Kottke-Marchant, MD, PhD, the clinic’s medical director of laboratories. However, she expects this may change for at least some tests.

“The new sequencing assays, DNAs, RNAs, and chromosomal microarrays, are typically very expensive in terms of platforms and equipment, maybe in the $500,000 range, and you have to have significant in-house volume to be able to have those platforms. But as manufacturers are making more standalone molecular platforms, the ‘labs in a box’ geared to low-volume labs, smaller hospitals may start bringing in more molecular testing. With the standalone platform, they don’t need to have the infrastructure with the clean room, the dirty room, and the specially trained technologists. Those have been prohibitive for many more moderate-sized hospitals.”

What Dr. Marchant has been seeing elsewhere is a pattern of many small independent hospitals joining together in consortia, with one of the hospitals becoming a core laboratory for the health care system. Such arrangements could possibly affect the dynamics of referrals from small community hospitals to commercial laboratories, she says.

Her laboratory does the testing for the main Cleveland Clinic campus as well as eight regional hospitals, a total annual volume of about 20 million billable tests. “By our projections we’re currently looking at a 40 percent to 60 percent increase per year in our reference lab testing, resulting in a fourfold increase by 2018.” That forecast was a factor in the decision to construct a new laboratory building on the Cleveland Clinic campus, effectively doubling the current lab’s 138,000 square feet. The new building will house all of the laboratory’s microbiology, molecular diagnostics, specialty chemistry, and immunopathology. Is new construction a pattern among large reference laboratories across the board? It might be, Dr. Marchant says. “There are quite a few that are expanding around the country.”

But the Cleveland Clinic also has its own test utilization concerns. The chair of molecular pathology at the clinic, Gary Procop, MD, leads the test utilization committee, which includes clinicians and has been proactive in a couple of areas, Dr. Marchant says. “One is cost management. We focus on Medicare’s diagnosis-related groups, and we’ve gone DRG by DRG to decrease the overall medical care costs of inpatients—not just laboratory work, but pharmacy, radiology, medical equipment, and so on.” For example, with cardiac patients, “we’ve made an 18 percent decrease in the number of blood gases we perform, and that’s brought us considerable savings in terms of lab costs.”

In a related measure, the hospital IT department has programmed a check on about 90 percent of lab tests to make sure they aren’t ordered more than once a day. “Sometimes it’s difficult with hospital IT systems. You may have to look in three different places to see if a test was already ordered, and the doctors just may not know. It’s not like FedEx tracking, where you know where every package is all the time. But with our new IT feature, when clinicians are ordering a test they will see that Doctor X ordered it a few hours ago, and the hospital system is not going to let them place a second order,” Dr. Marchant says.

Another initiative the Cleveland Clinic started in 2011 is the targeting of high-cost testing. “For any test over $1,000, we now have a group of deemed user experts, and if somebody is not on that list, they need to have the approval of a deemed user. Or we may have genetic counselors who can talk to them to ask: Is this the right test? Do you really need it now?” This approach has also been fairly successful, and she has seen increasing numbers of labs using it.

To reduce its own expenditures on send-out tests, the Cleveland Clinic has the goal of developing new tests in-house. “If the test is high-cost and has a high volume, we’ve made a push over the last three or four years to accelerate test development internally. And that’s been successful in the last couple of years. We were spending over $1 million a year sending out chromosomal microarray tests, so we purchased the equipment and hired the FTEs to do it here, and we expect to recoup the cost within a little over a year.”

Standalone hospitals can be at a disadvantage in pursuing such a strategy because their only volume is internal. That’s one reason even small hospitals may want to increase the volume of a test by doing local outreach to physician offices, Dr. Marchant says—in the same way that the Cleveland Clinic is looking to grow its reference lab business.

Splitting billing for a laboratory test’s technical component from billing for the professional component is a trend Dr. Marchant is seeing in a few areas of laboratory medicine. “One is in immunohistochemistry, also flow cytometry and some limited molecular testing, where you can send the specimen to a core lab, and then you’d only get charged for the technical component while your pathologist does the interpretation. The reverse also happens, where a small hospital might do a group of coagulation tests but may not have the expert in-house to interpret them, so they go to an outside lab that offers interpretation as a service. The specimen doesn’t move, but the reference lab provides the expert.”

The former arrangement is at the heart of a partnership between Comanche County Memorial Hospital and the reference laboratory Clarient, says Carol Dittmann, MD, medical director of the clinical laboratory at the hospital, located in Lawton, Okla. Given the limited technology available in midsize to small hospital laboratories such as hers, Dr. Dittmann says, “Partnering with a reference laboratory has allowed me, as a pathologist, to capture business I could not have gotten alone.”

For example, “My lab does not have a flow cytometer or molecular laboratory, but by sending samples to Clarient for the technical component, I can provide interpretations virtually, over the Internet, for flow cytometry or FISH tests such as the Urovysion test for bladder cancer. I may have the original urine cytology and then I can correlate what I’ve seen on the original sample with the Urovysion FISH. I can do the same thing with a bone marrow or lymph node where I have the histology here. This allows the local pathologist, who has control of the surgical pathology case and knows the histology and clinical history, to integrate that knowledge into interpretation of the tests that are performed off-site.”

Her laboratory is also able to provide interpretations for immunohistochemistry stains after the reference laboratory performs the technical component, Dr. Dittmann says. “Our lab at CCMH has an IHC laboratory and we are stocked with the antibodies commonly used, but for those antibodies that are needed only occasionally, when the volume of the test does not justify the cost of bringing that stain in-house, we send the block to Clarient. They perform the stain, and I can still interpret those stains, integrate them into my surgical pathology report, and bill for the professional interpretation.”

Those stains can be viewed online through virtual microscopy, or the glass slides will be sent overnight back to the pathologist for interpretation from the actual slide. Either way, the stain is made readily available to the pathologist without the hospital or pathology group absorbing the cost of purchasing and storing a long list of antibodies. And having the ability to interpret the slides by virtual microscopy shortens the turnaround time.

“The great thing for the local pathologist, when the test is performed in a reference lab, is if I see unusual results and I’m having difficulty interpreting them, with a single click on my computer I’m able to send them back to the reference lab and get a second opinion from a pathologist there. So it’s sort of a built-in safety net for more complex cases.”

The CCMH laboratory does not have the volume of breast cancers to justify purchasing and maintaining an image analysis system for interpreting ER, PR, and HER2/neu IHC tests. But through Clarient’s PathSite Scope IA system, she is able to interpret a patient’s ER/PR and HER2/neu IHC with image analysis, create a professional report including photomicrographs from Clarient’s template, and bill for the image analysis interpretation.

“These technical-only processes allow me to capture more of the market share of professional interpretation in my area.” She calls it a win-win for the pathologist—“a way to increase revenue and provide better patient care at minimal or no cost to the pathologist or hospital.”

The clinical lab at CCMH hospital and the associated Cancer Center of Southwestern Oklahoma use the same reference laboratory for flow cytometry and molecular testing for cancer patients. “Linking the lab results from inpatient testing and the cancer center’s outpatient testing creates a continuity of care and the ability to compare test results between inpatient testing and outpatient testing.” If different laboratories were used for send-out testing, Dr. Dittmann notes, the results might not be comparable. “For example, if you are following a quantitative test for a molecular marker of malignancy, such as BCR/ABL, the only way to compare results is to know that the test was performed by the same lab with the same methodology.”

But pathologists at the hospital closely monitor the send-out testing. “When it comes to referred-out tests, we work together with our oncologists here to be sure we are doing everything we can to minimize the financial impact on the hospital while maintaining the appropriate level of care.” At the laboratory level, one of the technologists notifies the lab manager if an expensive esoteric test, costing more than about $750, is ordered on a hospital inpatient.

“We’ve had some requests for genetic testing for unusual diseases costing as much as $8,000. Where we know it’s going to be absorbed into the per diem or DRG, I’ll have a discussion with the individual physician on whether the test is needed for current therapy or can be done on an outpatient basis.” In general, Dr. Dittmann says, there is a delay between the time cancer patients have a surgical procedure and when they’re going to start chemotherapy, so it’s usually unnecessary to order the test right away.

“In every case where physicians have ordered a very expensive test and I’ve talked with them, they’ve said it was something that could be done after discharging the patient. It was being ordered while the patient was in the hospital as a matter of convenience to the patient, when in truth, it was not going to affect current treatment and could be ordered later.” No physicians have demanded that the test be done on the inpatient, she adds.

Her laboratory routinely evaluates, roughly every quarter, the volume of tests it is sending out, to consider whether it has the methodology, instrumentation, and staffing to bring the test in-house. “Some tests we can’t bring in. But recently we brought in the 25-hydroxy-vitamin D test because the technology became available on one of the instruments the lab already owned. And we’re currently evaluating bringing in free testosterone levels.” They brought in microalbumins a few months ago.

One factor that affects her laboratory’s choice of reference labs is the ability to do third-party billing. “We made it a policy here between pathologists that we would not send esoteric testing, consultations, or testing we can’t provide here, at least in the anatomic pathology setting, to a lab that would not do third-party billing. When we use Clarient for our cancer testing or consultations and those tests can be third-party-billed, that saves the hospital money. It’s an absolute lifesaver for us, and not all reference labs will do it.”

Information technology problems can be an obstacle. “We have Specialty Labs as our clinical laboratory reference lab, and they can do third-party billing, which would make a big difference to the hospital, but they’re unable to do it for us because our LIS is unable to provide to Specialty the necessary information for third-party billing.”

Dr. Dittmann foresees an ongoing battle throughout the country to rein in reference testing costs. Even though the laboratory negotiates pricing schedules with its reference labs, “The cost of and the number of tests continues to go up. Especially when it comes to these cancer tests, there seem to be new tests available almost monthly, and the cancer guidelines for treatment require specific tests to be performed. So there are more tests we are required to offer to patients. But as a mid-sized hospital, we have no choice but to send those tests out.”

These factors underscore the need for pathologists to assume responsibility to help contain the costs of reference testing for the hospital, Dr. Dittmann says. Her advice: “Pathologists should see consultations with clinicians about reference testing as one of their main roles as medical directors of the laboratory.”