College of American Pathologists
Printable Version

  Finally, microbiology labs getting a lift from IT


CAP Today




May 2012
Feature Story

Anne Paxton

Visionaries, it is said, dream things that never were, and ask, “Why not?” But if they asked that question about complete IT solutions for the microbiology lab a few years ago, pathology informatics experts could give a pretty long list of reasons.

Beginning with the fact that microbiology is very different from clinical chemistry and hematology. “When they started bringing information technology to labs, clearly most of the companies started with clinical chemistry and hematology, and microbiology was kind of left behind or was an afterthought,” says Jacques L. Baudin, vice president, Commercial Operations and Information Technology, BioMérieux.

Not until recently did LIS companies develop software modules for microbiology, he says. “And these were called a paperless solution, and they started to interface some devices.” But even the interfacing of microbiology devices has always been a complicated task, and expensive, because in microbiology you have many more steps in the process and each of the steps lasts longer, Baudin says. “When you measure turnaround time, you are not looking at minutes. You are looking at hours. The workflow, too, is much more complex because you have to make different decisions at every step, and you have quantitative, semiquantitative, and qualitative data,” making data structures difficult.

Despite promises of the paperless lab, it may have appeared that modest connectivity, plagued by gaps in functionality, clunky manual procedures, and workarounds, was about as much IT as the microbiology lab could hope for. But that IT outlook has changed, industry experts and users in the field agree. Just as laboratory automation in microbiology is expanding beyond automated streaking of plates, and the demands of infectious disease control are increasing, microbiology labs are becoming far more efficient and useful through advanced IT solutions.

As microbiology supervisor for the Chester County Hospital and Health System, West Chester, Pa., Christine Biggs has found the Becton Dickinson EpiCenter Microbiology Data Management System, installed about three years ago, to be a powerful tool. The laboratory, which does 110,000 to 120,000 tests per year, has the EpiCenter interfaced to its blood culture and ID/AST instruments. “Eventually we’d like to have all of our instruments connected,” she says.

The BD EpiCare expert system, which EpiCenter uses in setting up template rules, is helpful, Biggs says. EpiCare contains a library of template rules that can be customized using demographic information coupled with results to generate alert messages. For example, “Certain drugs should not be administered to anyone under 18, and we set up rules to not report those drugs. Or with what I call our ‘Gorillacillin,’ very broad-spectrum antibiotics, those are not reported on the regular medical/surgical units. They are released only if the patient is in one of our ICUs.” Another rule: If an organism is resistant to vancomycin, then release the results for linezolid, but if it is sensitive to vancomycin, then do not release. “It’s very customized for the drugs in your pharmacy and for the way your ID and antibiotic stewardship committee people want things reported,” Biggs says.

EpiCenter contains all the Clinical and Laboratory Standards Institute rules already enabled on the system, and Biggs finds these invaluable. “For example, it will not let me report a first-generation cephalosporin on an isolate from spinal fluid, because it doesn’t cross the blood-brain barrier. It prevents an inadvertent reporting of a dangerous bug/drug combination based on the site of infection.”

Biggs is also a fan of EpiCenter’s alert messages, which pop up daily on a range of matters. “If you see a certain resistance pattern, like a vancomycin-resistant staph, that you don’t want released until it’s verified by the E-test method, it will print that message on the report for that patient. So rather than have sticky notes all over the lab that never get read, we get alerts that are patient-specific for a particular organism, name, and drug and body site.”

It’s also helpful in prompting reports to health departments. “I have one alert set up if there is a patient with salmonella or shigella. It reminds the tech to do a subculture, and it sends me an e-mail that tells me we have something that needs to be sent to the health department. Anytime there is a VRE, the infection prevention team gets an e-mail in real time,” she says. An alert also informs the clinical pharmacy and infection control departments when there is a multidrug-resistant organism or a MRSA in the ICU, which helps the hospital move immediately to be sure the patient is isolated and on the appropriate drugs. This has become particularly important since Chester County Hospital, like many others, is coping with MRSA rates of more than 50 percent.

EpiCenter also generates antibiograms on demand. “They will show what our gram-negatives look like, what our gram-positives look like, so the antibiotic stewardship committee can incorporate that information into the clinical pathways. We can also let our ED know the top 10 organisms we see from septic patients, so they can make an informed decision on empiric therapy. It’s really powerful stuff.”

There are myriad other ways to drill down and compile the data, she says. “We can print out anything we want about our community and what organisms are circulating in there. For example, I could do a trending for vancomycin MIC in Staph isolates. Is that MIC creeping up and up? Do we have to start monitoring it more closely? Is there a particular unit or particular area from which we’re getting patients with these higher MICs? It makes a little more work for me, but in a good way.”

The capability has real-time clinical implications, she adds, because the health system’s antibiotic stewardship committee set a policy this year that a verified MIC of 2 in a Staph aureus isolate must cue a call to the clinician to warn that treatment failures have occurred when a serious infection has been treated with vancomycin alone. “That is one of the alerts I have printing on the report that goes to the bench,” she says.

Clinicians are pleased with this capability. “The orthopedic surgeons might call and say, ‘How are we doing with staph infections in joints?’ And I can search EpiCenter based on physician or surgical unit and say your staph rates are low and only 30 percent MRSA.”

The same tracking has been helpful in slashing blood culture contamination rates. “We send reports to the ER in which we drill down to the individual phlebotomists, and using those they’ve been able to reduce contamination rates to below two percent, which is phenomenal for a big unit without dedicated phlebotomists. There’s a module that monitors the volume of blood in your blood culture bottles, so I can also use the EpiCenter data to tell them which people are not drawing enough volume of blood.”

But the kind of data management EpiCenter supplies is urgently needed in microbiology for other reasons, Biggs believes. “The CDC, the Joint Commission, and everyone else are on the same page: We have to improve our antibiotic stewardship program, because if we don’t rein in the amount of antibiotics we’re giving, we’re going to lose the battle. That’s the bottom line: antibiotic stewardship. And it’s totally data-driven.”

Simply being able to eliminate manual procedures through an IT solution has been a great boon for Clinical Pathology Laboratories of Austin, Tex. It installed Siemens’ LabPro Connect in the microbiology laboratory two years ago, and it’s brought much smoother operations, says microbiology supervisor Rhonda Brown, MT(ASCP). Before LabPro Connect, the laboratory processed 800 panels a day on its eight MicroScan instruments, and the system wasn’t designed for all the units to be connected.

There were two interfaces with the laboratory information system and individual interfaces between each MicroScan instrument. “To upload and download orders, we would have to share the two LIS interfaces with all eight MicroScans. We had to time the interfaces, and if we exceeded the capacity of one instrument, we’d have to manually transfer the panels to another instrument,” Brown says. “When the panels had exceptions, the techs would have to go to the instruments and resolve them. We’d have 17 techs doing that at the same time, so it caused bottlenecking at the computer.”

With the LabPro Connect software and LabPro Database Computer, all of those tasks can now be done at each technologist’s bench. The system allows technologists to work with ID/AST data at their workstation, even if the computer is being used for panel loading or unloading or maintenance activities.

This has greatly streamlined processes in the lab. “We’re a high-volume laboratory, and anything that can keep us from having to move from one location to another saves time,” says Dan Hardy, MD, medical director of microbiology.

The reporting of results is the same, but now that the laboratory has quicker access, it can get the panels out faster, improving turnaround time. Another feature Dr. Hardy appreciates is the ability to work with ID/AST data at the benchtop. “We don’t report every drug on the panel; we give a subset, so sometimes clients call and ask for additional antibiotics.” With LabPro Connect, laboratory staff no longer have to walk to the instrument; they can bring up that information from the panel immediately at their workstations. “Is ceftriaxone sensitive or resistant? You can get that information to the doctor immediately.”

Softmic, made by SCC Soft Computer as a Softlab module, is the LIS microbiology module in use at University Hospitals Case Medical Center in Cleveland, which has seven hospitals and a busy outreach program. The Softmic system has allowed the microbiology laboratory to be paperless since 1999, says laboratory manager Katharine Roman, MT(ASCP)SM, CIC(CBIC). “Softmic is a total, hands-on, at-the-desk-level workflow management system, which interfaces with seven different HIS systems here. It’s a very powerful system for microbiology, in that from the time the sample arrives in the lab until results go out, it guides your workflow.

“There’s no such thing as a paper work card,” she explains. “Softmic keeps it electronically. Keypads guide auto-ordering of additional ID and sensitivity testing; you simply put in the isolate, keypad select what drug panels or tests you want to perform. Softmic orders the panels or tests, adds on the billing, and downloads the information to the analyzers.” The analyzer prints appropriate bar codes, and the results upload back to Softmic—all with a few keystrokes.

“We have customized rules on our MicroScan instrument to indicate testing discrepancies,” Roman says. “For example, if the oxacillin and cefoxitin screens are discrepant, an alert triggers and additional work is done to resolve the discrepancy. Once the results are in Softmic, reporting is customizable. It’s virtually limitless.” Result reporting can be controlled by standard deviation rules, which are customized for the drug/bug combinations the lab wants to report, Roman says. “You can test one drug and say ‘show the results for these drugs that are equivalent,’ or you can prevent any antibiotic from reporting based on those rules. For example, we would not use ciprofloxacin on a pediatric patient, so for certain patient ages these results would not report.”

“In addition, Softmic auto-results and autoverifies our blood cultures, so as information comes from the analyzer, cultures are resulted and verified as ‘no growth’ without ever touching the bottle.”

Alerts are also set up in Softmic to prevent errant reports. If someone is reporting a result that the rules suggest is ridiculous, doesn’t match a previous result or requires additional testing, a customized alert will pop up on the screen to flag it for double-checking.

Roman is enthusiastic about Softmic’s automated batch resulting. “Our cultures are set up on worklists according to how the microbiology lab is laid out. All plates are set up with bar-coded labels, which allow us to scan the plates, eliminating entry errors. With automated batch resulting, you attach results of ‘no growth’ to hundreds of samples at a time with a few keystrokes.”

Softmic’s specimen tracking has been useful, Roman says. “We have a very large outreach business, and all the work from our six community hospitals comes here to our consolidated microbiology laboratory. With specimen tracking, we can track samples for all distant locations as well as one floor away, making sure everything arrives as it should, and, if not, determine where in the system it currently is. Specimen tracking forces accountability for sample management every step of the way.”

Softmic’s full-spectrum epidemiology reporting is the best she has worked with. “It’s at the top of my list. In literally minutes I can pull years worth of data, export it immediately to Excel, and turn out reports in any fashion I want.”

With the help of Softlab, she has just ended manual reporting to the Ohio Department of Health and the CDC. “We had been manually logging into the state health reporting system and printing off forms to mail, and that information will now flow directly to the state health department electronically.”

When her laboratory converted from Cerner Classic to Cerner Millennium PathNet in 2005, Claudia Atherton, MT(ASCP), microbiology system manager for Legacy Health System in Portland, Ore., chose to build many of the new rules herself, especially for the scripted workups. PathNet provides for online ordering of procedures, specimen management, and result review and verification for specific microbiology workbenches including bacteriology, virology, mycology, mycobacteriology, and parasitology. “The normal procedure would be for everybody to get the starter package for the organism database, but we chose to ‘build’ every organism and every response ourselves,” Atherton says. So despite lacking an IT background, she became essentially the data-management specialist as well as the manager in the microbiology laboratory.

She’s been happy with the customized alerts she’s been able to build from behind the scenes. “With the Cerner system, you can set up your panels, you can set up your own rules for susceptibility testing, and your own interpretations.” For example, “One of the norms in the microbiology lab is if you have the same specimen in the same patient on the same type of culture and they had a previous susceptibility within the last three days, there’s really no reason to repeat that susceptibility, as it’s not going to change for the same organism. So a little pop-up window will advise the technologist in these cases so they won’t even order a susceptibility.”

For purposes of management, the system allows her to go in and define what is abnormal, and produce reports to look at daily. “Originally I had all the MRSA results on there, but as we got more and more MRSA around, it took up too many pages. MRSA is not truly abnormal anymore, so I don’t have susceptibility flagging anymore.” Correlation reports are another feature Atherton appreciates when preliminaries are not correlating with Gram stains. “These will correlate your Gram stain to your report, so if the Gram stain came up with bacteria and the report is saying ‘no growth,’ you can take it back and say ‘this has been reviewed.’”

Atherton regularly uses PathNet to run statistical reports such as antibiograms, epidemiology, and infection control candidate reports. “The infection control candidate reports are great because I set the system to include general lab tests as well as microbiology, so we have all our reportables that we have to send to counties and states, and we can review them daily.”

Automatic no growth reporting, or ANG, is a feature she loves. “We used to have to do a batch report on testing that was no growth, and that’s still available. But for certain tests, after so many hours it’s been in an instrument, ANG just sends a no-growth message, so we don’t have to physically go into the work card on those. ANG is primarily set up for the 3,000 blood cultures her lab does each month, but is now also used for AFB and fungal cultures that tend to take a long time—in the case of fungal cultures, up to six weeks.

Another of Atherton’s favorites is the capability to customize patient demographics. “When you’re actually in the microbiology results entry, you have a big top section under the accession where you can put in any demographics you want the tech to see. We have standardized this in microbiology.” For example, “We have a procedure when you’re working with urine cultures that you’re not going to work up any organism unless there are so many white cells or organisms greater than 100,000. We have a rule that pulls those counts out of the urinalysis results so they don’t have to go back and find them.”

The microbiology lab has become much more productive using the rules and scripted workups. “A lot of people I’ve talked to have been very leery about using scripted workups, thinking that the technologists won’t like to be led by the nose on how they’re supposed to work up things, but even my most experienced techs absolutely adore them because they allow them to really get through the work. This is where it helps that I was doing the ‘build’ for the system, because I built it the way the tech thinks, not the way a computer person thinks.”

When Allen Foulks, MLT(ASCP), microbiology manager at Professional Clinical Laboratories, first arrived at the Fort Worth, Tex., laboratory in 2001, it had an old program called Revelation. “It basically only sent out reports, and you had to cue each report.” More recently, the lab has struggled with four different interfaces for its eight Vitek 2s, causing shutdowns of the LIS interface box three or four times a day. “In the evenings after ordering had taken place, we had to manually send demographics to each of the PCs from the Vitek 2 systems, and that was probably an hour to an hour and a half of the technologist’s time.”

But there have been major improvements since the lab’s installation of BioMérieux’s “intelligent microbiology middleware” solution Myla, which went operational in November. “We were in the process of acquiring BioMérieux’s automated plate streaker when we took a look at the Myla brochure and realized it was something we needed,” Foulks says.

“We wanted to bring all of our instrumentation through one collective box, so that we would be able to chop and dice and analyze information, and with our current LIS we don’t have that capability. Myla allowed all the instruments to be connected directly in just one interface.” All BioMérieux instruments are connectable now to the browser-based application, and other vendors’ instruments will be connectable soon as Myla is designed to be an open system.

Setup of the Myla was simple, Foulks says. “Their instrument software already comes programmed with typical Advanced Expert System rules, so we programmed a few of our own for salmonella, enteric, and how we are going to report antibiotics, and we spent no more than an hour to set that up.”

Myla’s dashboard feature allows the 10 users in Foulks’ lab to monitor all testing activities in the lab at their individual workstations, including key quality indicators and guidance for test processing decisions. “The techs can display or minimize the dashboard on their workstation as well as the high-tech remote viewers. So if they need to look at advanced expert rules, they do it sitting at their bench. There is no line waiting for review, and we have no problem about overload. If there are other issues they can’t resolve, they can call me, and I’m even able to access any of that from my house.” As manager, Foulks can use Myla to track what kind of cultures the laboratory is getting, their quantity, and many other indicators, he says.

The “manager dashboard” provides a real-time picture of what’s happening in the lab from a lab supervisor or manager point of view, says BioMérieux’s Jacques Baudin. Right now, he notes, it’s common to have a blood culture instrument outside the lab in a satellite location. “If a specimen turns positive in the middle of the night, nobody may be in the lab to see what is happening and you wait until morning to get the information. That’s the typical practice. What we can do is provide tools to alert a technologist in real time, which could potentially save several hours.” Studies have shown an impact on survival rates if a positive blood culture is acted on as early as possible. Providing the relevant information to the right person in real time is also important in fighting hospital-acquired infections, he adds. “We have a lot of programs around trying to fight resistance; we call it being smart. It’s really just providing the relevant information, and that’s why IT is so important.”

Another important role of Myla, Baudin says, is to use the key performance indicator information it collects to improve the lab’s productivity. “With the dashboard, you are able to monitor what is happening in the lab, and the idea is to say, if I change one parameter or if I change the way I work here, I can measure in real time its impact on the laboratory’s performance. So it’s really a productivity tool to use IT to be much more efficient in operating the lab.”

But BioMérieux sees advanced IT like Myla as integral to—in fact impossible to separate from—the emerging high-tech microbiology lab that will have Full Microbiology Lab Automation (FMLA) as well, Baudin says. “If you just try to automate without changing the workflow of the lab, you may not get the benefits you are looking for. As an example, Myla is the backbone of the Vitek MS and the integrator between the ID performed on the mass spectrometry device and AST performed on the Vitek 2 device. The real idea is you cannot provide an automation solution in microbiology without a brain, and beyond the robotics and specimen management, you really need to address how to manage the workflow and the information flow.”

A subsequent challenge, after reporting a hospital-acquired infection and witnessing that you have one, he says, “will be how we get to the point where we can prevent and reduce the rate of HAIs in a hospital or network of hospitals, and this is where we are looking at the data produced by this instrumentation. There is still a lot of work to be done on this, but if you consolidate this information with other types of health care information, you can really make an impact.”

Information management will be even more important as microbiology moves into digital imaging of culture plates. “After streaking and putting the plates in the incubator, the idea is then, rather than review the plate manually with the human eye, computer imaging assists in standardizing plate reading review and allows you to see things you cannot see with the naked eye.” The objective is that this will shorten the biggest drawback of microbiology testing: turnaround time, he says. “We are working on algorithms to make the incubation process smart, to shorten the time it takes to observe microcolonies. But the first step is to be able to identify if something is growing on a plate or not. We are also doing clinical studies at the moment to determine how to automate and speed up, as much as possible, the differentiation between negative and positive specimens.”

At this relatively early stage of adoption, advanced IT has already become central and essential to the microbiology lab, Baudin says. “Directly or indirectly, the microbiology laboratory needs to serve in real time more than clinicians through the LIS. It needs also to help with infection control, pharmacy, government reporting, and so on. So we need to talk to a lot of stakeholders, and the way you present information and how quickly you provide it to them is very important.”

Anne Paxton is a writer in Seattle.