Labs weighing pros, cons of micro TLA

By some signs, it now appears, the winds are shifting. New instruments, economic forces pushing consolidation, and the impending changes from the Affordable Care Act may combine to nudge clinical microbiology to the automation milestone that has eluded it so far.

For example, Dynacare Laboratories, a joint venture between Laboratory Corp. of America and Froedtert Hospital in Milwaukee, is in the final steps of administrative review for new microbiology automation instruments. “So we may be one of the first to have total automation in the U.S.,” says Nathan A. Ledeboer, PhD, D(ABMM), associate professor of pathology at the Medical College of Wisconsin and medical director of clinical microbiology and molecular diagnostics at Dynacare.

The microbiology laboratory has already seen the benefits of acquiring in 2012 a Copan Diagnostics WASP: Walk-Away Specimen Processor. “When we initially brought the WASP in, we intended to use it for urine testing. As we expanded our molecular menu and needed to start performing molecular tests on second and third shift, we needed some way of accomplishing that volume on the off-shift without adding staff. Having all our urines now directly on the WASP is a wholesale change in itself,” Dr. Ledeboer says.

“Among our clients, there were some that had incompatible containers, with metal caps that couldn’t be gripped by the instrument, but through a multidisciplinary working group we are working to standardize containers. I also think our streaking is much more reproducible with the WASP, and the technologists end up subculturing far less often than when they were doing manual streaking. That means we’re turning our cultures out, in some cases, a day or so faster.”

MALDI-TOF mass spectrometry is one of the first steps needed to automate microbiology, Dr. Ledeboer says. “MALDI-TOF has truly been a game changer. The reason was that it standardized our identification method. I can now use a variation of a MALDI-TOF extraction in a very standardized manner across bacteria, gram positives, gram negatives, yeast, Nocardia, and in the very near future we’ll have mycobacteria.”

“The Kiestra MalditofA or Copan MALDI picker or bioMérieux MALDI picker all do the same thing. They automate that entire process of spotting target plates, adding matrix, and all the technologist has to do is put that plate into the instrument and read it. And that was a major change.” Digital plate reading is another key advantage of automation because it obviates the need to have somebody move plates around the lab, Dr. Ledeboer says. “The beauty of high-resolution imaging is you can see all those detailed components without having to touch the plate.”

The Toshiba SCARA robots are also part of what make the WASP possible, and in Dynacare’s experience, they have been extremely reliable. “The only issue we’ve had with implementing WASP was a bad printer that had to be replaced, and Tarzan, our larger robot, broke a finger and had to get a new one. The company’s been very, very responsive whenever we’ve had problems.”

Dynacare Laboratories, which has about 50 percent of its volume in outreach testing, decided to move beyond specimen processing and look into a TLA solution for personnel and revenue reasons, Dr. Ledeboer says. “If we didn’t connect our WASP to an automated line, we’d have to replace two people the WASP is currently supporting. In addition, we’re looking at revenue enhancement opportunities whereby we can free up an additional person on a couple of shifts, then bring in a couple of molecular tests we badly need from a patient care perspective but haven’t had the staff to perform.” One of those is HIV genotyping, he says. “We’ve been consistently sending those out, and they’re extremely expensive.” Another candidate for insourcing is automated HPV testing.

As Dr. Ledeboer explains, microbiology TLA extends automation from the front end to the analytical phase and can bring significant benefits to the laboratory through its flexibility, scalability, and interchangeability. With continuous incubation of plated media, the laboratory can achieve a smooth workflow. Since plate images are stored digitally, review of growth over time and examination of patients’ culture history can be done independently of staffing levels at any given hour. Because of TLA’s greatly simplified workflow, studies have shown it can significantly increase productivity, measured by the number of samples processed per full-time-equivalent technologist, and facilitate labor savings or redeployments.

Of the systems on the market, the most established is the Kiestra TLA, which has been installed in at least 38 laboratories since 2006, using an open architecture. Its modules, linked by a conveyor/track system, include the specimen processing and streaking modules (SorterA, BarcodA, and InoqulA TLA), plus ReadA incubators with digital imaging equipment and ErgonomicA workbenches. Kiestra is enhancing the TLA system with instruments to automate microbial identification and antimicrobial susceptibility testing using automatic colony picking by the MalditofA instrument.

The bioMérieux Full Microbiology Lab Automation system begins with the Previ Isola automated plate streaker, which is connected by a conveyor/track system to the smart incubator system (SIS), available in CO2 and non-CO2 atmospheres, and to image analyzers. The FMLA’s Myla software links the system’s components, including the Vitek MS (MALDI-TOF instrument).

Like the FMLA and Kiestra TLA, Copan Diagnostics’ WASPLab connects its microbiology specimen processors to incubators that assign each plate a unique address or “shelf,” and to an image acquisition station, with the option to carry out MALDI-TOF target plate seeding with either the Bruker MALDI-TOF plate or the bioMérieux Vitek MS plate. An Inpeco sorting station and track allows microbiology tube specimens to be routed to the WASP­Lab. The first WASPLab system was installed in a microbiology laboratory in 2012.

NorthShore, a four-hospital system that has about 35 FTEs doing 300,000 microbiology tests per year, is already planning to start installation of a BD Kiestra system in 2014 as part of a major laboratory renovation, Dr. Thomson says. “Right now, we have automation similar to what others have—the semiautomated blood culture instruments. The plates go to a manual workbench where the technologists do the identifications. But organism identification in the microbiology lab is very quickly switching to MALDI-TOF, and that instrument can readily automate the identifications.”

“So with the Kiestra, the smart incubators will provide high-resolution images of the plates, and the techs will have a computer screen workbench; they’ll select all the colonies to work up and antibiotic tests to be done through the screen in front of them, and the conveyor will take the plates right to wherever they have to be worked.”

Automation will add standardization that laboratories haven’t had before, Dr. Thomson notes. “The procedures will be computer precise. Every plate is inoculated perfectly and in the same way, and incubation will always be for the prescribed amount of time. The plate won’t be photographed for analysis until it’s been in incubation the minimum time. You’re not going to have a technologist working two overnight shifts a month who is not streaking as well as someone on the day shift. When you look at a high-resolution computer screen, you can magnify those plates and see what needs to be done with a particular colony. I think it will be far better than what we can do manually right now.”

One of NorthShore’s additional incentives to acquire TLA was to make the laboratory more useful to doctors and better for patient care, Dr. Thomson says. “The quandary is that all of the equipment is expensive. Labs are already moving off site from hospitals, and spending a lot of money on this automation is going to accelerate that because of economies of scale. We in the lab, and the clinicians we serve, have argued for years that a remote lab is not good for patient care; it removes the ability of laboratorians to communicate effectively with clinicians.”

But with digital imaging and TLA, he says, “we can finally offer exceptional patient care because there will be a picture of everything that’s in the computer. So when we issue a report, we’ll have the option of not just saying ‘staph,’ but also showing a stain and showing the bacterial count on the plate. And the clinician or caregiver or whoever is interpreting it will once again be able to see this is a pathogen or a non-pathogen, and whether it’s associated with inflammation, and this will enhance their ability to interpret the laboratory result. To me, that is one of the wonderful things that sells me on all this automation.”

In England, Dr. Thomson notes, “To try to save money, they’ve declared they’re going to centralize all laboratories. I’ve told the microbiologists there—who happen to be infectious disease clinicians as well as microbiologists—that with this laboratory automation I think they can do the same job they’ve done in the past from a remote site. They can take the lab report on an iPad or other device and at the patient’s bedside pull up all the smears and plates and look at them right there, just as when the lab was attached to the hospital. So I really think it’s going to revolutionize the way medical microbiologists here and abroad do their job.”

Kaiser Permanente SCPMG Regional Reference Laboratories plans to choose a TLA system for two of its laboratories in the first quarter of 2015 and is considering all three vendors, says Susan M. Novak-Weekley, PhD, director of microbiology for the North Hollywood, Calif., reference lab. After a demonstration of TLA’s capabilities at the American Society for Microbiology meeting a few years ago, “Several of us within Kaiser realized at the same time that we should be heading to total laboratory automation. We’ve always tried to look at what we can do better. Acquiring total lab automation for microbiology was just a matter of time.” The Kaiser lab has had automated plating systems since 2004.

In fact, with the bioMérieux Vitek MS MALDI-TOF having been approved by the Food and Drug Administration in August, Kaiser plans to acquire a MALDI-TOF within the next couple of months. “It’s going to revolutionize bacteriology,” Dr. Novak-Weekley says. “We were unable to bring it in before because it was labeled as research use only, but now the system will replace other automated identification systems and produce a result in a fraction of the time.” Once the technology is validated internally, she says, it will be used as it is in other laboratories that are using MALDI-TOF already—to identify pathogens from blood cultures.

One of her key concerns in considering full automation in a high-volume laboratory is the robustness of the automation. “Mean-time-to-failure data is important when considering the purchase of any piece of automation in the laboratory, something microbiology has not had to concern itself with as much as other departments, such as chemistry.

“Mean time to failure can be somewhat deceiving as well, in a way, because there’s a big difference between a failure in a lab that’s putting through 200 specimens a night versus our lab, which would be putting through 4,000 or more specimens a night. So it’s critical to get that data from the vendor and talk to other users to assess how the instrument is going to perform under a similar set of circumstances.”

The TLA vendors’ analyses project a hefty return on investment, either through reducing FTEs or reallocating staff. “We have a laundry list of tests we need to internalize that are now going to outside reference labs. Celiac serology is number one, then beta 2 glycoprotein, CMV viral load, Quantiferon, some of the hepatitis tests, HTLV when the assay is available, and perhaps some other molecular tests. Literally seven of our top 10 send-out tests are microbiology-related assays, since I am also responsible for serology and molecular infectious disease testing as well.” Total laboratory automation will allow the reallocation of staff in a way that will, it is hoped, save jobs and reduce send-out test costs at Kaiser.

But very few labs are the size of Kaiser’s, and return on investment analysis has to be linked to a laboratory’s size and what’s going to fit in and be most effective, Dr. Novak-Weekley emphasizes. “When you speak of TLA, it’s kind of a catchall phrase for labs like ours. For smaller or medium-sized labs, the question is, will they have a track system, will they have specimens going out of the plating instrument into incubators? Laboratory automation in microbiology can mean different things for different labs. The question is: How does the automation fit into a laboratory based on its size and sample volume? I think there are components of this that will be integrated into every laboratory at some point, that are going to be staples for any lab.” For TLA vendors, the challenge will be to develop modularity that will benefit small and midsize labs, not just the large laboratory like Kaiser. “You might not need a plating instrument, but I think everybody will need a smart incubator. Everyone might not need a MALDI picking station for plates, but they’ll need digital microbiology and MALDI-TOF.”

She expects microbiology labs to see increases in volume from the Affordable Care Act. “That’s what we’re anticipating, that our patient base will increase along with the workload in the laboratory. Automation will allow us to open up paths to those new patients without relying on increasing staffing,” Dr. Novak-Weekley says. So it’s a good moment in the nation’s health care history to consider an automation solution. “Sometimes the water’s cold—but you’ve just got to jump in,” as she puts it.

Taking on TLA is a significant move, since specimen processors run about $300,000 to $400,000 while fully automating a high-volume lab can easily cost $2 million to $5 million, Dr. Ledeboer says, noting that change management is always a challenge. “As laboratories, sometimes we don’t know what we’re going to be doing on any given day, and if you’re not sizing appropriately with your instruments, you’re not going to get savings. You start adding significant expenses by building in a lot of redundancy or a lot of extra capacity on the instrument.”

Microbiology TLA is likely to have a relatively small marketplace of 500-bed hospitals and larger, Dr. Ledeboer believes. “I think the national labs will be looking at this too and will likely automate, but they are a for-profit business and will want to see the return on investment faster.”

Anecdotal data in the United Kingdom, where some TLA systems are up and running, are showing improvement in overall laboratory efficiency and reduction in cost, but “I think peer-reviewed studies are really badly needed in this particular area,” he says. Most hospitals’ financing for automation systems will typically come from their bottom line. “It’s difficult to reagent-rent a piece of automation. Are you going to take your blood plates and transition them from costing 20 cents to $3 or $4? So you have to find capital, and when you’re talking $2 to $5 million, you’re in the price range of a new MRI scanner or a new Da Vinci robot for surgery. The minute everyone’s trained on that, it starts saving money. Those devices generate a much faster return on investment than the microbiology lab.”

Reducing the number of employees or full-time equivalents is not something a lot of laboratories like to do. “But in all honesty, reducing FTEs or expanding workload is the only way you can make TLA work,” he says—and it may not be as bad as some predict. For one thing, the chronic shortage of technologists, new research suggests, may ease due to some technologists’ deciding to delay retirement. In addition, he believes that automation embraces technologists for their knowledge and skill set. “It’s not a wise use of a technologist to have them plating urine cultures or chasing plates around the lab. We need technologists for their ability to read and interpret plates and make decisions on what needs to be worked up.”

The Affordable Care Act will add new factors to the return-on-investment equation. “The change in reimbursement that’s coming is going to bring a significant alteration in the laboratory landscape, and I think a lot of people don’t know how to manage it. My initial feeling is it’s going to promote another round of consolidation, because there’s so much pressure on the lab to do good quality on the test and do it at low cost, and you can’t do that by pushing things out to the patient, although you can certainly improve turnaround time and clinical meaningfulness that way. For some tests like Group A strep PCR, there’s a lot of value in moving them closer. For mycobacteriology, on the other hand, we’re going to wait two weeks for the organism to grow anyway, so pushing it closer to the patient may not make sense. So it’s really going to be a balancing act.”

How TLA will mesh with the increasing role of molecular testing is an open question, Dr. Ledeboer believes. “None of these automation systems at this point have embraced any form of molecular testing, and I think that’s an area of opportunity for these systems. For example, BD has its own thermocycling system, BD MAX, and bioMérieux just bought BioFire which has a lot of potential as a near-patient diagnostic. What if bioMérieux’s FMLA could both set up cultures and molecular tests?” Integrating molecular makes a great deal of sense, he points out. “We’re already managing that specimen to perhaps set up plates for respiratory culture or other types of specimen. Why can’t we ask our automated testing system to pipette out the specimen to set up a molecular test?”

Alternatively, he notes, there is at least a question as to whether it makes sense to automate an old technology: culture. “We’re in the 21st century in an era of metagenomics. What is the impact of high-density digital PCR—and more importantly, next-generation sequencing—on culture over the long term?” For those contemplating TLA, when the future may be in the direction of NGS, the answer depends on how long your time frame is, Dr. Ledeboer says. “I do think automation will be a significant, beneficial tool for the next five years, but we’ll absolutely be using more NGS and more mass spectrometry. So in five years my answer might be very different from my answer today.”

Automation still has its critics. Steven D. Dallas, PhD, director of the microbiology laboratory at University Hospital in San Antonio, isn’t quite on board with a full-scale embrace of automation in microbiology. “Robots and digitizers are impressive, mesmerizing, and fascinating, but they’re overly complex. I’m not a retro-grouch,” he insists. “I just want the automation to do more than impress me with robotics; I want it to be clearly superior, and I’m not convinced it is.” He calls MALDI-TOF a “clearly superior technology compared to biochemical identification,” but says it can be implemented without TLA robotics and imaging.

His own laboratory, which does about 15,000 billable microbiology tests a month, is located in a 500-bed county hospital where the focus is on cost-avoidance rather than earning revenue. While some laboratories appear to be jumping to automation via WASP and Kiestra, and he has seen a bioMérieux Isola Previa in the laboratory of a hospital with only 250 beds, he questions the investment calculations. “If you’re an administrator trying to get the best payback on an instrument, it had better be running all of the time. But in many reference laboratories, the majority of microbiology testing comes in on the second shift because that’s when the doctor’s office closes.” Dr. Dallas is not sure whether automation can deal effectively with surges in volume and with idle time.

He also doesn’t see how automation can shorten the turnaround time of culture. Even the most complex robotic system on the market, he says, still streaks only one plate at a time or reads growth from one culture at a time. “We are starting to read cultures on two shifts, so we’ve shortened our turnaround time that way. But I’m not convinced at all that automation of preanalytics decreases turnaround time.”

Demonstrating return on investment with microbiology automation is a challenge, he adds. “If your administrator says here’s an instrument, show me the CPT codes I can bill with front-end automation, there aren’t any. So the only way you can save money is by putting someone out of a job.” However, in a reference laboratory with sufficient specimen volume, “automation might fit better because you can re-purpose your technologists to do things that really require a brain.” Back-end microbiology automation such as MALDI-TOF can be justified more easily, he says, because there’s a CPT code associated with the identification of each organism.
Another under-recognized risk is that microbiologists could find themselves in the position of radiologists, watching imaging work drift away through outsourcing. “Be careful what you wish for because, theoretically, you could have a technologist reading your Gram stains and cultures from China or India. There are terabytes worth of information on a digitized Gram stain slide, but I think we’ll see more outsourcing of anything that is image-based because it can be viewed on a server that anyone can access from anywhere, any time of day. So we could lose control of our microbiology altogether.” For these reasons, he says, laboratories need to think more about the upstream and downstream implications of what they do.

Most crucially, Dr. Dallas points out, automation of microbiology is still just automation of a 20th century technology—placing a specimen onto a Petri dish. Even though robots do it much more consistently than humans, he says, “it’s inaccurate to call any of this ‘full’ or ‘total’ automation because it does only culture-based bacteriology, which is 60 to 80 percent of what we do.” He envisions a truly disruptive technology as the next step, perhaps involving “nanotechnology or microfluidics or small tabletop platforms instead of robots with arms like ours—something we haven’t imagined yet.” Microbiology automation has a long way to go, in his view. “Robots need to be small and scalable, and they need to automate new technologies, not old ones.”

Dr. Thomson agrees that TLA puts small microbiology laboratories at risk of being phased out. “One thing that seems constant every year is squeezing more and more money out of the system, and the small, unaffiliated, freestanding lab is not going to be financially competitive. TLA is going to make it even less competitive, as will the Affordable Care Act.”

However, Dr. Thomson finds himself converted from a skeptic of microbiology automation to an enthusiastic advocate. “I was cautious a decade ago, even five or six years ago, before we got glimpses of what TLA was really going to be. But now that it’s appeared, I think it’s going to be great.” He hopes the instrument manufacturers will get together to make sure that the automated lines by one company can handle automated PCR instruments and other devices made by other companies. “We in the profession need to convince the manufacturers to opt for open systems, because we know from longstanding experience that no one company makes the best of everything. So far the companies haven’t taken that step.”

“The important thing is for everybody to make sure they don’t put their heads in the sand and ignore automation,” Dr. Thomson emphasizes. “It’s a wonderful opportunity for us to change 180 degrees the way we’ve been doing things for a long, long time. There’s a large potential for improvement, and we all have to be part of the process so the end result is excellent microbiology that helps patient care, costs as little as possible, and moves us forward.” 
[hr]

Anne Paxton is a writer in Seattle.