Always dependent on a large number of repetitive, manually intensive tasks, the laboratory department was among the first in healthcare to look to automated equipment for solutions. First it was instruments--some more intelligent than others. Then came laboratory information systems (LIS) to support electronic data entry, storage and retrieval. The preferred technology now, according to Arden Forrey, PhD representing the American Clinical Chemistry Association of Washington, DC, is a clinical information management system (CLIMS).
According to Robin Felder, PhD, professor of pathology at the University of Virginia in Charlottesville, a whopping 80 percent of the numbers generated in a typical medical facility still come from the laboratory. Too often, however, the laboratory generates numbers but few value-added services. "If we embrace automation technology, the laboratory can become an information specialist and manage most of the information generated in healthcare facilities," he advises.
Political and regulatory upheavals have taken their toll on the laboratory. Government regulatory compliance rules have placed new responsibilities and new restrictions on the laboratory doorstep. After the fraud and abuse muddles of the past decade, laboratory reimbursements now are tied to "approved" and "medically necessary" diagnoses and test codes. Meanwhile, the laboratory must rely on cooperative clinicians to provide correct diagnostic codes. A computer-based patient record should eventually assure proper coding, but in the meantime, little short of goodwill can efficiently get those codes to the order entry form.
The government is serious about compliance with its regulations, however, and the laboratory is beginning to see new regulatory compliance IT tools to address compliance. James Small, MD, PhD, associate pathologist and director of IS at UniPath PC, a specialty reference laboratory in Denver, hopes vendors will incorporate such functionality into product updates, but sees little value in a third-party product because access to the data is the important element. "CLIMS vendors need to recognize that users must be able to retrieve information in ways the vendor never envisioned during the design process," he says. Larger vendors are just now beginning to offer the powerful and flexible query tools that users need to address unpredictable problems such as queries for compliance with legal regulations.
Many hospitals are trying to regain lost ground following periods of rampant outsourcing. In addition to bringing back more in-house test procedures, many also are adding services--often teaming with other members of the care team to provide outreach and specialty services. Such tactics are resulting in more direct competition with reference laboratories, long dominant in the outpatient market. And as business edges blur, so do IS needs, resulting in increased markets for hybrid clinical information management systems.
Hybrid hospital/commercial systems may be evolving more rapidly, but overall, changes in laboratory systems are occurring slowly, says Don Lyons, CIO of Bio-Cypher Laboratories of Sacramento, Calif. The director of clinical chemistry at Montefiore Medical Center in Bronx, N.Y., Herb Rose, PhD, agrees. Following extensive research while planning his new laboratory, Rose sees few bright lights--and those he does see are from relative newcomers in the market. By and large, he gives most well-known laboratory systems poor scores. For the two systems Rose thinks capable of standing up to the workloads in today’s laboratory, the common thread is development origins: Both began by developing systems for the reference laboratory. "Their development approaches included the integration of external sites," Rose says, "and that’s where everybody else falls apart."
Closely tied with the greater healthcare IT picture, many CLIMS and laboratory automation trends follow those in other segments of the organization. As Lyons redesigns and rebuilds the IS platforms at Bio-Cypher, the second largest reference laboratory in California, he’s considering the price/performance equation, new operating system and hardware architectures for his large operation. The laboratory can take advantage of other healthcare trends that include interface engines, multi-tiered database engines, HL7 application interfacing and advanced lower level connectivity designs based on open standards such as the TCP/IP protocol. There also are many developments unique to the laboratory. Among them, Lyons cites better specimen management and tracking, specimen processing robotic engines, comprehensive bar coding of specimens and optical archival of requisition information with interfaces to the CLIMS and financial systems.
Maturing and more reliable technologies also are good for decentralization, says Lyons. Internet Protocol supports better remote device connectivity and advanced networking technologies capable of carrying more traffic. Specimen management, too, is benefiting from advanced courier and specimen management tracking algorithms. Although browser technologies have not yet been deployed to any extent in the laboratory, Lyons believes that this will be the next serious development wave for the CLIMS.
Speeding up the line
Laboratory automation systems (LAS) are a reality, but haven’t exactly been embraced. "There is great opportunity for labor savings as well as for improved quality and safety," says David Clarke, application engineer for Labotix Automation Inc., Peterborough, Ontario, Canada. That may be true, but it’s the labor savings part that most workers hear first. Felder thinks it’s a major obstacle to widespread adoption. "What most laboratorians don’t realize is that there are hundreds of tests that the laboratory could be doing but can’t afford to do because it’s too busy with nonautomated tasks," he says. "By automating, simplifying and streamlining, the laboratory can not only do a better job, it can add even more diagnostic tests."
Most LASs on the market provide front-end automation for specific specimen preparation and sampling functions to automate many of the monotonous, labor-intensive tasks associated with testing such as specimen uncapping, bar coding, aliquoting, centrifuging, test routing, storage and retrievals. At the opposite end of the spectrum is the fully automated total LAS, a turnkey system that can perform most tasks without human interaction. Modular total laboratory automation is somewhere in between, allowing the laboratory to begin with a basic configuration and add automation over time. Only a few vendors are in the market at this time. They include Coulter Corp., Miami; Labotix Automation, Inc., Peterborough, Ontario, Canada; MDS Inc., Toronto; and Roche/Boehringer Mannheim, Indianapolis.
Front-end automation for specimen preparation and sampling is getting a better reception than either total or modular total, but just barely. There may be savings from fewer and less costly laboratory workers, but most facilities still have difficulty justifying the investment. "At this time, there is no clear cost benefit to robotics versus an efficient, well-run manual laboratory with good machinery," says Small. Labotix’s Clarke thinks it’s still too early to accurately assess cost benefits.
A few organizations are investing. However, they tend to be the same ones that have been early adopters of other new technologies--organizations such as Kaiser Permanente in Oakland, Calif., and Duke University Medical Center in Chapel Hill, N.C.
According to Clarke, no CLIMS vendors are building or buying LAS systems but most recognize the future opportunities. "The LAS demands an automation-friendly CLIMS," he says, "and some are a lot better than others." The Labotix system, for instance, requires the CLIMS to be able to track individual specimens. While a few CLIMSs have such built-in capabilities, most can only track specimens on a patient basis.
Rose thinks equipment manufacturers will bring more acceptable automation solutions. These companies now are announcing modular instrumentation that doesn’t require tracks, he says. Such a machine will be big and, in most cases, the laboratory will have to be remodeled, but it will need only one person to run it. Announcements of such units now mean that it will be two years before they are fully tested and ready for action. Rose notes that it will require a lot of software, a lot of computer horsepower and a lot of engineering to link the modules together, but most of the manual work will be automated. (See Case in Point, page 76.)
But don’t expect Rose to buy the total automation packages now available. "It is a mistake because total automation never, ever pays off," he says. "Total automation manufacturers often combine front-end and back-end automation in a package. Employee savings sound good, but if you break out component costs, pay-back on the front-end is 12 to 18 months. On the rest of the automation, it’s nine years."
Weighing the options
In addition to his work in a specialty reference laboratory, Small works in a large hospital in a single vendor environment. A single vendor operation has some advantages in interconnectivity, says Small, but no one vendor offers the best of all modules. There is always some compromise. "Too many vendors have patched clinical applications onto a financial system in order to tout a healthcare information system (HIS)," he says. That wouldn’t be a problem except that it is often numbers-proficient accounting that gets the last word. "And accounting often buys the HIS with the best accounting and billing package, leaving clinical systems out in the cold," says Small. He emphasizes the importance of including clinical people in the decision-making process. "If you don’t," he warns, "you will end up hampering the clinical work force every hour of every day."
At some point, Small continues, many laboratories will face the problem of trying to merge operations of two different and incompatible systems. Migration to one of the systems or to a new system is an option, but many times, interfaces are politically the most palatable solution. He says, "Interfacing is a wonderful concept but difficult to implement well. Whenever I see the word ’interface,’ my blood curdles." Achieving functionality through interfacing is not only problematic, it’s costly: In Small’s hospital, interface problems are second only to network and communications problems for causes of computer downtimes.
Interfaces between different CLIMSs also create a working problem. If you have two different information systems, says Small, you really have two different laboratories. Since the CLIMS determines much of the workflow and the language used in the laboratory, the use of two different systems stymies cross-working. He says, "Training a new technologist is about 10 percent technology and 90 percent computer data entry."
Tracking with the CPR
The decreased turnaround times that automation can enable won’t make much difference clinically unless the test results can reach the caregiver fast,which brings us back to standards and connectivity. Pay attention to standards, warns Lyons. Connectivity protocols such as HL7, TCP/IP and ethernet are essential in a heterogeneous environment, he says, and will enable the inter-connectivity and record exchange requirements for the computer-based patient record.
"If the laboratory is to play a significant role in generating and interpreting results for diagnoses, technology is key," says Felder. The 12-minute window during which the physician sees the patient is the critical time period to receive the data, interpret it and get it back to the point of care. With technology, it’s possible, he says. And by moving testing nearer the patient, involving clinicians in simplified testing and focusing generating information for direct patient care, Felder is building a solid place on the patient care team.
The distributed testing logistics at the University of Virginia may make Felder more acutely aware of remote connectivity issues than many of his colleagues who are consolidating and centralizing. (See Case in Point, below.) Among other research efforts to improve clinical laboratory operations, Felder and his team at the University of Virginia’s Medical Automation Research Center have moved to reducing data entry requirements in the next generation of system. The center’s new connectivity tool will rely heavily on the Web and will minimize, if not completely eliminate, interactive data entry.
Business process automation also is a key part of Lyon’s plan for his commercial reference laboratory, but "the client will drive the design process--not the other way around," Lyon says as he confronts his current challenge--a system to transmit data from the laboratory database with connectivity to all Web-enabled physicians. In his view, "No longer will your system be an island unto itself, rather it will be a part of the lawn of digital architecture that is slowly surrounding our world. It is rapidly transforming the world of medicine to benefit the ultimate consumer--the patient."
DISTANTLY RELATED TO JAVA, THE TOOL command language (Tcl, pronounced "tickle") was developed 10 years ago as a joint project between Sun, University of California at Berkeley and John Ousterhout, PhD. Although Tcl lacks the name recognition of its cousin, the Association for Computing Machinery (ACM) presented the 1997 Software System Award to Ousterhout for his work with Tcl and its toolkit (Tk). Presented in May, the award recognizes the development of an influential software system.
One of the first embeddable scripting languages, Tcl can "glue" together complex systems and has dramatically changed the creation of custom and extensible applications. Its companion, Tk user interface toolkit, simplifies graphical user interface construction. Petar Stojadinovic, project manager at Chiron Corp. in Emeryville, Calif., thinks Tcl shines as a tool for laboratory automation. It is particularly good at enabling computers to talk to robots and machines, he says. If you are using Beckman instruments, chances are good that you are already using Tcl.
"The beauty of Tcl," says Stojadinovic, who also teaches the language, "is that it was intended to be simple." Unlike Java, Tcl is not just for developers. And because it is a scripting language, experience in programming is not necessary. "If you want to keep it simple," he maintains, "you can use Tcl. All you need is a manual and a table of commands you want to issue."
For more information on Tcl: http://www.tclconsortium.org
Bedside Blood Banks
SPECIALIZED AREAS IN THE LABORATORY present unique challenges to IS. In the blood bank department, a ruling by the Food and Drug Administration (FDA) means that its computer systems are now regulated as medical devices. But not only did the FDA decision compound compliance requirements, most vendors with blood banking systems exited the market. Few choices remain for prospective buyers of the specialized systems, and for those with donor programs, the choices are particularly restricted, notes James Small, MD, PhD, associate pathologist and director of IS at UniPath PC, Denver.
The FDA ruling has resulted in an even wider separation between the blood bank department from the main clinical laboratory. Robin Felder, PhD, professor of pathology at the University of Virginia in Charlottesville, calls it a feudal kingdom. "The blood bank department is a golden opportunity for automation," says Felder, "but it is politically the most difficult to automate." Some processes can begin to be automated--witness the first wave of automated cross-matching--but trying to consolidate blood bank into the central laboratory to take full advantage of the integrated system is tough. "Eventually I would like to see blood banking at the patient bedside," he says, "but technologically it will be one of the last places to target."
Another area in the laboratory that may require special data management is tumor marker tracking--especially if the clinical laboratory information management system is short on online storage space. Few systems have the capability to capture and maintain the patient demographics and specimen results for extended periods, says specialty software developer David Rhoads, president of David G. Rhoads Associates, Inc., Kennett Square, Pa. And this is a problem when tracking changes over months or even years.
Lab terms defined
Clinical laboratory information management system (CLIMS). CLIMS is a management system that facilitates the receipt of all information necessary for ordered laboratory procedures, the delivery of laboratory results to the caregivers and storage of the information for future use. It supports the data management requirements for all sizes and types of clinical facilities. CLIMS shares common information management functionality with the laboratory information management system (LIMS) but, as a clinical system, includes integration with the computer-based patient record (CPR) domains. Eventual linkage with the LIMS is expected.
Sometimes used to describe a second-tier relational database management system that supports more efficient accumulation and reporting of clinical data such as cumulative graphics, utilization studies, encounter and Health Plan Employer Data and Information Set (HEDIS) reporting data.
Laboratory automation system (LAS). System for automating laboratory tasks. The modular LAS automates only specific functions or tasks; the total LAS can perform most laboratory tasks with little or no human intervention. Modular total laboratory automation combines aspects of both systems. Once a core system is configured, the laboratory can add automation stations over time.
Laboratory information management system (LIMS). Used in pharmaceutical manufacturing, general analytical and environmental chemical measurements, LIMS is subject to a separate standard (ASTM E-1578) than the clinical laboratory system. LIMS and CLIMS share a common core information management functionality and eventual linkage between the two is anticipated.
Laboratory information system (LIS). Traditional name for what is now being termed a clinical information management system (CLIMS). It is sometimes differentiated from CLIMS by its focus on the day-to-day production side of laboratory data management.
CASE IN POINT
University of Virginia, Charlottesville
While much of the industry is consolidating, integrating and centralizing laboratory services, laboratory services at the University of Virginia in Charlottesville are moving out--out to nursing stations, clinics and the patient’s bedside.
Robin Felder, PhD and professor of pathology, says, "Point-of-care healthcare delivery will replace everything else. Eventually it will be in the home, but until we get there, healthcare must provide laboratory, pharmacy and other support services at the clinic or at the hospital bedside." According to Felder’s data, such a distributed system costs about half that of a centralized system, with most savings in labor. Moreover, the organization makes most efficient use of its most valuable commodity--physicians.
In addition to the more common handheld testing devices, Felder’s laboratory has been aggressive in placing table top instruments and even floor model analyzers near patients. Any member of the care team--including physicians, nurses and ward clerks--can submit specimens for testing. The only qualifier for running specimens is literacy, he says.
Long distance QC
Historically, maintaining quality control when nonlaboratory personnel perform testing procedures has been difficult at best. The trick, Felder says, is to link the entire episode to the professional in the laboratory who monitors the event. Instead of transporting the specimen to the central laboratory--with all the inherent problems of lag time and the quality control issues of packaging, transporting and repackaging--the process goes directly from the patient to analysis. "Electronic linkage over a local area network or the Web allows the laboratory to maintain quality by transporting the information rather than the specimen," Felder says.
More sophisticated and more comprehensive tools for on-site analyses are in diagnostic manufacturers’ pipelines, he says. The new wave in microfabricated analytical engines not only will offer a wider range of high quality testing (glucose, coagulation, lipid analysis and a 10-panel chemistry panel are already a reality), but more affordable analyzers to the point-of-care. Smart analyzers with built-in calibration and QC parameters already self-manage many of the QC processes.
"The key is the informatics to ensure that the entire episode is captured and monitored by laboratory professionals," Felder says. The laboratory still participates in each event, but they participate electronically and at a distance. "We need to raise the level of professionalism of laboratorians to the point where they are a member of the decision-making team," he says. "We want to sit at the bedside while the event is happening and help the physician make a final decision."
Felder also is director of the Medical Automation Research Center (MARC) at the University, the only academic medical automation group in the world (http://labautomation.org/Marc/home.shtml). Among other research efforts to improve clinical laboratory operations, the MARC invented the Remote Automated Laboratory System (RALS) in the early 1980s that is now licensed and used exclusively by Roche/Boerhinger Mannheim of Indianapolis. Felder and his team now are focusing on a next-generation system that reduces data entry requirements. Now under wraps pending patenting processes, its new connectivity tool will rely heavily on the Web and will minimize, if not completely eliminate, interactive data entry.
Felder claims big cost benefits from automated laboratory systems such as RALS. "Once you remove labor from the equation," he says, "the cost benefit is dramatic. Pay back periods for the hardware and on the informatics are usually less than a year. "What more could you ask for," he queries, "than higher quality, faster turn-around times and a one year return on investment?"
For more information on laboratory automation: The Association for Laboratory Automation (http://labautomation.org).
CASE IN POINT
Montefiore Medical Center, Bronx, N.Y.
Two years ago Bronx, N.Y.-based Montefiore Medical Center, one of the largest nonprofit healthcare systems in the U.S., began to shift its focus from hospital inpatients to serving an outpatient population in neighborhood-based family care centers. Now, with a 32-clinic network of satellites, Montefiore physicians offer evening office hours for working families and take walk-ins seven days a week.
Centralized laboratory operations are an important part of Montefiore’s ability to manage a distributed clinical environment efficiently. Herb Rose, PhD, director of clinical chemistry, says the laboratory maintains its own fleet of couriers and vehicles that collect specimens from the clinics every day between 10 a.m. and 10 p.m.
Sheer volume is enabling Montefiore to add more specialty testing--and important revenue streams. "Without volume, the laboratory cannot become efficient," Rose says. Although the days of free-wheeling physicians ordering scores of laboratory tests are over, Rose maintains that the laboratory is still one of the hospital’s cash cows. Government intervention and regulations have just made it more difficult.
Montefiore’s centralized laboratory receives 1,200-2,000 tubes of blood a day. "In order to become efficient and effective," he says, "I have taken the laboratory and turned it into a factory. What do efficiency and cost savings mean? Use the least number of people to do the most amount of work." Developing clinical pathways and workflows for specimens isn’t feasible in such a high volume setting, he says. "We get a tube with a bar code label, period. We don’t have a chance to ask any questions."
Montefiore also is subscribing to the new core laboratory concept where chemistry, hematology and some parts of microbiology are merged into one operation in one location. The hospital’s new laboratory now is under construction. When completed, Rose has planned to automate everything that can be automated. "The planning process is tedious," he warns, "but taking the time to do the detail work is more important than the concept."
The laboratory eventually will link to a computer-based patient record--a migration that has already begun within the medical center. Rose does not plan for his laboratory systems to be part of a single product family in the medical center, however. His computers will be separate and interface for connectivity. "There is no way that one product can be used to do both," he notes. "Our volume would bring such a system to a crawl."
Charlene Marietti is senior technology writer at Healthcare Informatics.