College of American Pathologists
Printable Version

Breaking bottlenecks at the front end

June 2000
Cover Story

Anne Paxton

On the surface, Kaiser Permanente in Berkeley, Calif., would appear to be a fine candidate for a total laboratory automation system. Its regional reference laboratory, part of The Permanente Medical Group Inc., performs 70 percent of the tests ordered on 3 million health plan members in northern California-some 12 million tests a year. In a 1997 continuous quality improvement project, the laboratory identified its manual sorting of 12,000 tubes a day as a problem area that urgently needed a cure.

"There was a bottleneck and the tubes were taking a shift and a half to get sorted. We were interested in growing the business. But our turnaround time was getting longer-and you can’t grow when you’re giving poor service," says Gene F. Pawlick, MD, director of the laboratory.

Instead of turning to a comprehensive automation system, however, TPMG chose a modular, task-targeted system from Roche Diagnostics-one that automated the preanalytical processes of sorting, decapping, and aliquoting-and a postanalytical archiver. "We considered TLA," Dr. Pawlick says. "But we decided not to go in that direction."

Columbia-Presbyterian Medical Center in New York City drew a similar conclusion following its consolidation three years ago. It chose last year to install a Labotix system, including an automated centrifuge, an uncapper, an automated aliquoter, and a specimen sorter.

With more than 2,000 accessions a day, why not a total laboratory automation, or TLA, system? "We didn’t think it was economically feasible," says Daniel Fink, MD, director of the core laboratory, located at New York Presbyterian Hospital, Columbia Presbyterian Center. When the hospital looked at preanalytical automation only, "We felt we could get 80 percent of the benefit [of automation] for 25 percent of the cost," he points out.

Gary V. Catarella, MT(ASCP), MBA, vice president of laboratory services at Catholic Health System in Buffalo, NY, says even though the central laboratory there runs 1.4 million of the 2.7 million laboratory procedures the system bills each year, it did not have the volume for TLA.

When it started setting up the central laboratory for five hospitals last November, Catholic Health System installed a Beckman Coulter front-end system. "We started considering front-end automation when we were doing a business plan for our centralization and looked at the volume of specimens we’d be receiving," Catarella says. "We knew we needed to reduce some steps, standardize our testing, and streamline our processing, so that is primarily why we were interested."

Move toward modularity. The choices these laboratories made reflect a broader trend, says Robin A. Felder, PhD, director of the Medical Automation Research Center and professor of pathology at the University of Virginia, Charlottesville. It’s a shift from full-scale automation solutions to more contained or modular approaches. Stand-alone preanalytical devices that may be coupled to TLA systems but that can also operate independently, he believes, are the next wave in laboratory automation and will soon be "sold like popcorn." His prediction: "This is where automation will finally get its foothold and start proliferating."

The pressures to automate are legion. Top-level mandates to reduce laboratory costs have become routine and continue to prod many hospital laboratories to explore automation approaches. But the severe shortage of skilled technologists and technicians is now an equally compelling factor. A few short years ago a personnel glut led training programs to shut their doors throughout the country. Thanks to the prolonged economic boom, however, human resources departments are scrambling to fill openings in the laboratory.

A recent survey by the Clinical Laboratory Management Association found that about 24 percent of managerial and technologist positions are unfilled, and nearly 40 percent of laboratories say there is an insufficient applicant pool. "Obviously, there’s a huge labor shortage in the laboratory business. There just aren’t enough people," says Ronald Berman, marketing manager for clinical chemistry and automation at Beckman Coulter.

Coinciding with the shortage is heightened concern about medical errors and a possible shift in philosophy about how to prevent them. The Institute of Medicine report, "To Err Is Human," released last year, unsettled the public with its estimate that 44,000 to 98,000 deaths-at least the equivalent of a packed jumbo jet crashing every three days-are caused by errors made in the hospital.

The report also shook up the medical establishment with a clear message and a new mandate: Human error is inevitable; stop looking for individuals to blame; focus on improving procedures and processes. "Patient safety," the IOM stressed, "does not reside in a person, service, or department, but emerges from the interactions of components of a system."

The White House and Congress have launched initiatives to act on the IOM’s recommendations. But the report has already infused quality assurance efforts with a new urgency, at the same time that manufacturers are making it easier to automate. A growing number of them are "unbundling" total laboratory automation systems and offering attractively priced equipment that automates the most labor-intensive functions of the laboratory-the preanalytical area.

"It’s a very manual area, and there are no economies of scale," notes Dr. Fink. Without automation, "if you double the workload you have to double the people."

Second to the economic advantage, however, is a combination of quality and safety, says Peter Urwin, former president of Labotix and now senior vice president and general manager of Lab-InterLink. "We know of laboratories where people suffered serious injuries from broken tubes, mistakes were made using the wrong specimens on tests, tubes have been misplaced. Automation helps reduce that."

At least six companies are marketing or preparing to release products designed to reduce the hands-on work needed at the front end. Ontario-based Labotix recently merged with the U.S. firm Lab-InterLink; together they have some two dozen installations of preanalytical equipment. Roche Diagnostics and Beckman Coulter have completed beta testing and have several installed systems, and Olympus, too, offers a preanalytical product. Abbott/Tecan’s Genesis FE 500 system is in its final clinical trials, and the company plans a release shortly.

As a result, laboratories that lack the volume to justify total laboratory automation today are finding it feasible to make the move gradually-or even planning to automate only functions at the front end. "It’s obvious that this technology is perfect for small, medium, and even large labs," Dr. Felder says. "Laboratories with 300 or 500 samples a day can do them pretty readily with human labor. But preanalytical systems are reasonable for laboratories conducting tests on 1,000 samples a day, and with more than 1,000 samples, laboratories will want to consider multiple copies of these processors."

Blurred terminology. Although the term TLA describes a number of approaches, generally it indicates a comprehensive system that couples conveyor belts for transporting specimens with discrete workcells to provide centrifuging, sorting, routing, and actual placement of the tubes in and out of the analytical systems.

But the distinctions in current use are not clear-cut. Urwin sees a blurring of the line between partial and total automation. "It’s definitely true," he says, "there are more people interested in front-end perspectives, and many people look at that as a way to enter the market. But as we start discussing it, they invariably want to put some instruments on line too. Is that front-end automation, or TLA?" Many of Lab-InterLink’s customers who have not technically plunged into TLA have nevertheless purchased pre- and postanalytical equipment.

Lab-InterLink, he notes, sells a mixture of systems. "We offer modules to build anything from preanalytical to full-blown automation. For example, a laboratory in North Hollywood is running in excess of 1,500 tubes an hour, yet just using front-end automation. That’s a pretty big lab, and we have others running less than 1,500 tubes a day that have close to total automation."

Urwin agrees that smaller hospitals are showing more interest in automating their laboratories. "This business, like all business, started with the early pioneers. They tended to be the larger labs that had more money, were quicker to experiment, and found automation easy to justify. The market now includes a second tier of laboratories who were waiting to see how the early adopters fared. Now people are looking at it and saying, ’You mean it’s real? It actually works?’"

"We’re moving into a phase where we’re marketing to more conservative laboratorians who want a stable, proven product, and don’t want to take any risks," says Labotix marketing manager David Clarke.

Noting that TLA might have been oversold or too expensive and garnered a bad name perhaps because of some early installations where TLA did not go so well, Clarke says that, in its most recent brochure, Labotix doesn’t refer to TLA at all. "We actually de-emphasized the terminology. We say we have complete systems, which include preanalytical, analytical, and postanalytical automation."

A wider market. Laboratories in the United States have never embraced TLA the way labs in Japan did starting in the 1980s. In Japan, some 170 automation systems have been installed, while only about eight percent of North American hospital and commercial laboratories are even considered suitable for TLA. Among the key obstacles, Dr. Felder points out, are the extent of consolidation of the laboratory, its physical layout, the available market for laboratory services, and the affordability of automation.

There are still hopes that total laboratory automation will become widely adopted. Lab-InterLink and Labotix Automation have installed several TLA systems in North America and Europe. The merged company’s president and chief executive officer, Jack Holthaus, predicts the TLA market will "grow significantly as hospital management realizes that automation lets them provide safe, accurate testing while controlling costs."

But, Dr. Felder says, "Given the limited market for TLA, automation vendors had little choice but to develop modular workstations that provide the scalability to accommodate the large variety of productivity needs."

"Now labs are deciding they can essentially reduce TLA to its most valuable elements," he continues. Rather than preparing to pay millions of dollars for a TLA system, laboratories can gain key advantages from automation in the $250,000 to $750,000 price range.

"I was interested in something that would allow us to be very flexible and agile and respond rapidly to any change that might occur either internally or externally," Dr. Pawlick says. Roche Diagnostics (at that time Boehringer Mannheim) approached him after hearing TPMG was experiencing a bottleneck in receiving and suggested that several laboratories in Germany were solving that problem with preanalytical automation.

So Dr. Pawlick traveled to Germany to see for himself. "I saw probably six labs with these instruments and it was amazing the throughput they had. They were light years ahead of what we were doing in the States at the time. They had huge labs with a lower volume of people. When I saw them in action I said, ’This is exactly what we are looking for.’"

That was how TPMG decided to select the Roche system, becoming the first installation of Roche’s system in the United States and its largest installation in the world to date. Dr. Pawlick calls the task-targeted automation system a smashing success. "We no longer get 12,000 specimens a day. It’s up to 16,000, basically without any new FTEs in that area. So we’ve had a 33 percent increase in work volume without adding any staff and with no compromise in turnaround time, which is unheard of."

A simple addition at the post-analytical end also made it possible to reduce staff, he notes. "As we did workflow studies, it became very apparent that one of the great applications on tap was the ability to archive specimens after they leave the test area." So Kaiser’s laboratory bought four instruments dedicated to archiving.

Known as "Archie," the instruments record the row and position number of the tubes electronically. "So if a cholesterol was done three days ago, we can find which rack the specimen is in and pull it out. Ten years ago that was all done manually and it was a real problem." In this way, automating archiving has also eased the process of doing add-on tests, Dr. Pawlick says.

Building incrementally. Tom Wilson, director of laboratory integration marketing for Roche Diagnostics, explains the company’s approach as a continuum of automation solutions, starting at the task-targeted level and proceeding to TLA. Modular preanalytics, launched in 1999, comes in different configurations depending on the functionality customers require. Two other products, the PSD I, a primary sorter and archiver, and the VS II, an aliquot system that can also sort and archive, are now installed in nearly 30 sites.

"The market’s desire for TLA is significantly less than several years ago. In fact, in terms of requirements, laboratories would choose task-targeted automation 16 to 1 over TLA," Wilson says. "But within that set, roughly half of those people would want to have the ability to build incrementally toward a TLA solution in the future. What it comes down to is that people have a vision-but they want to manage risks and budgets and ensure their business plans are met over a period of time."

Laboratories handling 500 samples or more per day make up Roche’s major market, he says, but a hospital with 300 samples a day might be interested if it is on an aggressive outreach campaign. Even a 200-per-day customer with a high staff turnover at the front end of the laboratory may be interested. "These are generally less skilled, lower paid individuals, and if you have difficulty recruiting or have a high turnover," Roche’s automation solutions can help, he says.

"We’ve talked to our customers for the last few years, and they all need to improve productivity, reduce costs, and run more tests," says Jeff McHugh, Beckman Coulter’s vice president of diagnostic commercial operations for the Americas. "The biggest opportunity to get those kinds of improvements is on the preanalytical side, without question."

Beckman Coulter, which introduced its Power Processor last year as a stand-alone front-end automation system, shipped about 25 of them last year and expects to ship another 40 to 50 this year. Although a few sites have their Power Processor connected to an LX 20 to automate distribution of the tubes to the analyzers, most of them employ sneakernet-where a staff person picks up the sorted specimens and walks them over to the different analyzers. Only about 40 percent of them are potential candidates for the LX 20, the company estimates.

More than half of the company’s current customers now say they are interested in some sort of front-end automation, Berman points out. TLA systems are usually customized but typically would run close to $3 million, whereas the Power Processor can be acquired for around $450,000. "Because of the price changes, [the preanalytical automation] is really getting down to the mid-volume accounts, the 200-bed hospitals," he notes. After two years of sales, Beckman has found many customers now want to add accessories or options. "They want to take a very logical, progressive approach," Berman says. "They really seem to want scalability."

Beckman Coulter has installed at least as many systems in foreign countries, largely in Europe and the Far East, as in the United States. "Those customers are very similar," Berman says. "But they’re not as willing to go to a core lab design to consolidate workstations." Because there tend to be more turf issues among physicians who control, say, an immunochemistry department, or immunoassay, these countries have more of the traditional "silo" type of laboratory. "The core lab is more of a U.S. concept, but it’s slowly being adapted worldwide," he adds.

One of Beckman Coulter’s first customers to obtain striking results was St. Mary’s Hospital Center in Montreal, the only CAP-accredited hospital laboratory in the province of Quebec. In a recent assessment of front-end automation’s impact on its turnaround time, St. Mary’s reported that stat chemistry TAT improved by 28 percent, and telephone calls from unhappy physicians dropped from 28 per day to fewer than five. Immunoassay TAT improved by 47 percent.

Productivity, or reportable results per staff person, improved by 17 percent, and even samples not processed on the automated system improved: routine hematology by 36 percent and stat hematology by 29 percent. Although the TAT for coagulation did not improve significantly, according to Ralph Dadoun, St. Mary’s vice president for corporate and support services, the hospital correctly suspected that by automating two thirds of the volume of blood specimens, it would improve the entire testing process.

Cautionary voices. That is not to say there is universal acclamation for front-end automation. When he was associate vice president and laboratory director at Aultman Hospital in Canton, Ohio, Scott Whalen, PhD, did research on front-end automation some four to five years ago and found the options were limited. Aultman chose to automate its high-volume chemistry, hematology, and immunoassays, but not the front end. "When we looked at those very early systems, the cost was quite high, and there were a lot of technical barriers. There really weren’t any cost-effective solutions," Dr. Whalen says.

"When you look at automation in the area of doing tests, it’s really a software-driven technology. Transport from one instrument to the next is important, but more important is how the information is handled," he explains. Front-end automation really brings in more of the hardware-and hardware-associated problems along with it. "For example, if you’re an outreach lab and physicians may use your lab but for insurance reasons also use others, there may not be standardization of tubes across all referring physicians."

That was one problem that made the cost of front-end automation prohibitive, in his view, and he’s not sure how markedly the situation has changed. At Aultman, "There probably was something that would have worked, if our goal was a totally automated laboratory, but the people doing that function were paid less and were extremely efficient." Despite having 1.5 million accessions per year, "I didn’t have that many of them even though there was a high volume."

In his current post as chief operating officer of Mt. Carmel West, part of a three-hospital system in Columbus, Ohio, that has about the same test volume as Aultman, Dr. Whalen is conducting the same research. He says it is too early to conclude that front-end automation is appropriate. Laboratories should evaluate vendors’ claims of increased efficiency, he advises. In many cases, he believes, since front-end labor tends to be less expensive, it may be just as efficient to keep the process operating manually, unless the employment market is so tight that the laboratory can’t find reliable people to do the work.

On the other hand, he adds, a front-end system would probably reduce the number of mistakes. "As a general principle, the number of times you can decrease handling of the tube, you’ll decrease errors," he says, pointing to safety features of the newer systems that significantly reduce human intervention in capping, uncapping, and aliquoting.

Alverno Clinical Laboratories in Hammond, Ind., was not actively in the market for front-end automation in 1998 because it wanted to finish consolidating before budgeting the capital expenditure for automation. But when it had the opportunity to be a beta tester for Roche Diagnostics’ PSD I, "We actually installed our first piece of automated equipment almost three full years ahead of the plan," says James C. Sparks, PhD, president and CEO of Alverno, a not-for-profit laboratory formed by the Sisters of St. Francis Health Services to provide core laboratory services for its hospitals in northwest Indiana and southern Cook County, Illinois.

After the PSD I and Roche modular analytics system were installed in 1999, Alverno installed a modular preanalytics system early this year. "There was a huge bump in our ability to be productive," Dr. Sparks says. "What we’ve done is essentially tripled our capacity to handle chemistries, drugs of abuse, therapeutics, and, eventually, immunochemistry without adding staff."

While the laboratory has not been automated long enough to assess quality differences, "I can tell you there is a qualitative difference in the technologists’ ability to get the work done. They’re happier people, and our processors, rather than being stuck on rote jobs-look, spin, rack, and send-actually are able to use technology to do those things, and they’re upbeat too."

The core lab connection. In many, if not most, laboratories that opt for front-end automation, the process is inseparable from the creation of a core laboratory. "You have to have a pretty significant test volume," Dr. Sparks points out, "so core labs are the major market. We joke that we need to ’feed the monster.’ The expense isn’t outrageous, but it’s not insignificant, and to get a return you need lots of volume."

Loyola University Medical Center in Maywood, Ill., decided in the winter of 1998 that it would work with Beckman Coulter to incorporate its front-end automation equipment, the ACCELNet system. "To do that, we needed to create a core lab," says Stephen E. Kahn, PhD, section chief and director of chemistry, toxicology, and near-patient testing, and associate professor of pathology and biochemistry at Loyola.

It took six months of planning to combine the automated areas of chemistry, hematology, immunoassay, coagulation, and fluid sera analysis with a specimen receiving-processing area into one large laboratory section. In late June 1998, the core laboratory operation went live. It now processes more than 3,000 tests per day.

"Because we were Beckman Coulter chemistry users, and because we had over the previous years made a number of reductions in staff positions through attrition and eliminated overtime, we really didn’t have the fat to trim out of the laboratory to justify a track system," Dr. Kahn says. Since Beckman Coulter was interested in conducting a beta site clinical trial, the timing was right for Loyola to automate its preanalytical functions, he notes.

The project has been successful, bringing an annual personnel savings of about $100,000. "We’ve been able to take on an increased workload with a slight decrease in staff due to some shifting of positions," explains Dr. Kahn. "But the incorporation of front-end automation and creation of the core lab have only succeeded because of the hard work, professionalism, and commitment of the technical staff, the core lab manager, and the technical specialists. They’re the people who have pitched in, suffered through rapid change, worked overtime, and really taken it upon their shoulders to make the system work."

In fact, Loyola is unique in having the only core laboratory in an academic medical center in the Chicago area, as well as the only robotic workstation in an academic institution, says George Krempel, Loyola’s administrative director of anatomic and clinical pathology. He stresses that improvement in turnaround time was a key goal of the restructuring and automation. "We noticed that 20 percent of our TAT is associated with the preanalytical phase of testing, and if we were going to shorten our TAT, we knew the instrument could only process so much, and the window of opportunity fell in the preanalytical processing area, which is labor-intensive."

The improvements in service delivery are owed as much to the decision to develop a core lab, Dr. Kahn believes, as to the incorporation of front-end automation. "They were both integral, but each had an independent impact," he points out. "If we had done only one or the other we wouldn’t be nearly as far along as we are today." Now Loyola plans to take automation to the next level, by going to a limited track system capability that will extend all the way from the front end through the analytical phase to refrigerated storage, he explains. "We very much have taken a stepwise approach to incorporating automation."

Moves to automate may be tied in less obvious ways to a health care system’s decision to set up a core laboratory. Lab-InterLink’s Urwin points out that when consolidation is in the wind, the laboratory with the best image is more likely to be a survivor, and automation has a way of brushing up any department’s image.

Bearing out this phenomenon, Kaiser was surprised by the media reaction to its installation of preanalytical automation in April 1998. There had been a recent scandal about a technologist in California who was reusing needles. "So there was a big scare in the news at the time about transmission of infectious diseases," Dr. Pawlick recalls.

"The day we had an open house to go live, we had several TV stations and newspapers and a formal news conference. But when the press came in the room and saw the instruments, one of the reporters said, ’Oh, this must reduce the spread of infectious disease,’ because there’s a big plastic lid that pulls down over the whole operation. Once the specimens are in there, all this racking and decapping is actually contained inside a glass casing. So that was the biggest play on the news that night-not the automation."

Taming adaptation costs. Along with their more modular approach, most vendors of preanalytical systems are assuring customers they don’t need to convert their analyzers to achieve compatibility. Abbott Laboratories, partnered with Tecan, is also preparing to launch an open automation system that can adapt to any company’s instruments. The Genesis FE (Front End) Workcell is based on the concept of configuring preanalytical, analytical, and postanalytical "islands" in the laboratory.

"When we analyzed the laboratories’ workload, we found the front-end labor-intensive tasks took up to 65 percent of hands-on time, while transporting samples from the accessioning area was only two percent," says Matthew Noble, Abbott’s integration solutions marketing manager for the U.S. "Those who didn’t embrace TLA didn’t want to spend enormous capital and restructure their entire laboratory just to accommodate a very large automation system. Our mission was to impact the 65 percent, where there tends to be more potential for errors, mislabeling, incorrect aliquoting, and lost samples."

Noble believes that the Abbott/Tecan system, which is in place in two clinical sites, is different because it was intended from the start to be flexible. In contrast with other manufacturers, which have track systems that they have modified and scaled down for North America, Abbott "basically had a clean slate to develop a system with front-end processes in parallel rather than in a series, as in the track system," Noble says.

Also taking a workcell approach, the University of Virginia developed and conducted clinical testing on a versatile device that is being installed now at Duke University Hospital in Durham, NC. The CoagAutoLink robotics device, which performs preanalytical processing for coagulation specimens, is unique, Dr. Felder says, in that it allows specimens to be delivered in racks of 50 tubes by a medical technologist, by conveyor belt in TLA, by mobile robot, or by individual tubes that can be placed stat directly into the instrument.

In time-and-motion studies, the University of Virginia found that workcell automation completed preanalytical tasks in about 48 minutes with less variance than humans, who took from 24 to 105 minutes. However, the robotic workcell was unable to process improperly labeled specimens or determine if tubes were filled properly. Those features may be added in the future, Dr. Felder notes.

At Kaiser Permanente in Berkeley, the laboratory saw a significant reduction in FTEs in the preanalytical area along with a radically reduced throughput time and improved accuracy, Dr. Pawlick says. The number of repetitive-motion injuries has also fallen. Because of the hand and arm motions required, "Occupational injuries in the laboratory are fairly high for carpal tunnel syndrome," he points out. "We didn’t seek the system to solve that, but it became very apparent that it would alleviate the human injury problem."

From the standpoint of mechanics, the system installed by Catholic Health System in Buffalo has performed as expected, but staff acceptance is a different matter. "People aren’t sure how it fits yet or how it will affect them," says Catarella. "I wouldn’t say there’s resistance, but there is hesitation due to unfamiliarity with such a high level of sophisticated equipment."

Simplified logistics. Centralized Laboratory Services in Long Island City, NY, was given a fairly typical charge in 1998. "Like most laboratories, we were being asked to do more with less," says CLS’ technical director, Joseph Stauffer, PhD. With its workload ranging from 6,000 and 7,000 specimens per day, CLS started looking for ways to increase its efficiency and productivity through automation. But the laboratory’s physical layout complicated the automation issue.

"We have a rather old building, which would have required a lot of reconstruction for any track system," Dr. Stauffer says. "Our building has a lot of walls and wooden floors, and since most of the line systems out there are very heavy, they need really level floors. So we were looking for something that minimized the amount spent on reconstruction but also had flexibility, since our business was growing, and we needed to add new pieces on without major expenses."

CLS purchased Roche’s sorters and aliquoters. "From our perspective, TLA really didn’t fit, because our specimens come in in large batches," Dr. Stauffer points out. "Line automation was really conceived for a continuous flow of specimens. We needed something with a very high throughput; that’s why we picked individual modules. The stand-alone sorters are faster than the sorters you see on lines, and the aliquoters were faster too." Another advantage: The stand-alone units are mobile and wheeled in rather than fixed to the floor.

"We actually didn’t make any space alterations except electrical modifications and piping in an air pressure system that’s used by instruments themselves to manipulate the arms on the sorters and aliquoters," Dr. Stauffer notes.

At CLS, the staff savings that have been achieved through attrition are significant. Nontechnical laboratory aides are employed to operate the five pieces of equipment in accessioning, and Dr. Stauffer expects the accessioning staff ultimately will be cut by about 50 percent. The machines are "very easy to use, and the staffing required to run them is very small," he notes.

Unbolting and financing automation. Often, cost is not the only barrier to TLA, says Clarke, Labotix’s marketing manager. "Sometimes you don’t have the space for it." At three sites where Labotix has installed preanalytic automation, the specimens are delivered to other floors. In one laboratory, the company solved the delivery problem by building an elevator. In the laboratory at Kaiser Permanente in Portland, Ore., a rail-based Translogic carrier system was installed, while at Kaiser in North Hollywood, a robotic cart transfers specimens from floor to floor.

But one of the key advantages of preanalytical systems is not having the instruments on line, he emphasizes. "With TLA, if you change instruments, you have to change the interface between the automation and the instrument. But preanalytical automation allows the laboratorian to choose the best of breeds. You can mix and change instruments on fairly short notice."

That flexibility was one of the features that attracted Kaiser Permanente in Berkeley to a modular preanalytics system. "The process of TLA is very confining," Dr. Pawlick maintains. "In my view, once you bolt a track system to the floor, you suddenly find the driver of all future lab activity is the bolted-to-the-floor conveyor belt, and not your agile mind."

Dr. Felder sees financing issues influencing the direction of the market for preanalytical systems. "It gets a little more complicated with these systems," he notes. "The manufacturers’ bread and butter has not been mechanical devices, but reagents. That’s where they make their money. A preanalytical processor does not have a revenue stream. It has no reagents, and there are many disposables available on the open market. So it’s not going to be a real moneymaker when compared to analyzers. But it will give companies a strategic advantage they can use as a way of leveraging the sale of additional products."

Catholic Health System in Buffalo does have an arrangement for its Beckman Coulter product that is similar to a reagent rental. But most of Beckman Coulter’s customers are leasing the systems. At some institutions, McHugh points out, there are separate capital budgets for process improvement within the hospital. "Some of them have been able to acquire the systems by actually tapping into a new re-engineering budget."

Financial issues aside, Dr. Felder is optimistic about prospects for preanalytical automation. "It’s one embodiment of automation that’s useful to a broad range of people. It doesn’t have to be reconfigured and tailored to each site." Eventually, he predicts, a preanalytical processor will become a common device throughout the laboratory industry. "The automated device will resemble a box that is rapidly installed by the vendor, training will be provided, and the device will operate with the ease and reliability of a clinical analyzer," he says.

Nevertheless, Loyola’s Dr. Kahn cautions, automation has its limits. "Mislabeled specimens, inappropriately drawn specimens, occasional specimen clotting if you are dealing with serum-those are preanalytical errors that are always going to stop you dead in the water, whether you’re automated or not. There are hardware as well as electronic components in these systems that can help you identify these kinds of occurrences and track them down, but you’re not going to make them go away."

Anne Paxton is a freelance writer in Seattle.