CAP TODAY’s annual laboratory automation survey is designed to help laboratories that are considering investing in laboratory automation systems understand the strengths and potential limitations of today’s LAS options. Many lab automation systems have purchase prices in excess of $1.5 million, and thus a purchasing mistake can not only be financially disastrous but also set back the clock on laboratory expansion projects that are needed to maintain laboratory profitability. Though there are numerous examples of successful LAS installations, there are also examples of laboratories that did not realize the full benefits of automation. For example, a reduction in turnaround time can just as easily become an increase in turnaround time if all the process variables that need to be controlled are not considered.
The laboratory automation data published on pages 30 to 48 in the print edition of CAP TODAY—vendors’ answers to CAP TODAY’s questions—are the best source of laboratory automation specifications to consider when making a purchasing decision. For those who have not yet studied laboratory automation, understanding the implications of these specifications can be daunting. Deciding whether and what to purchase requires a working knowledge of process engineering, laboratory science, and informatics. So that you can more fully understand and appreciate the vendors’ responses in the following tables, I will explain why CAP TODAY asked the questions it did of the participating vendors. (My words in italics match those in the left column of the tables, which are shortened versions of the actual questions.)
A laboratory automation system consists of many modules and subcomponents that are integrated with laboratory instruments to function as a well-coordinated system. The main components of an LAS are an input area for preanalytical sample processing, a transportation system, a queuing system to queue specimens before analysis, a buffer to hold specimens until they are finalized, and a storage and retrieval system. The LAS should be managed by a centralized computer system running a process management software program that requires minimal operator effort and provides a seamless gateway for verified laboratory data to be presented to the hospital information system.
In the tables, the first year installed will tell you how long the vendor has been in business, which invariably relates to time spent working out system bugs. Automation products that are available addresses the automation components found in an LAS, listed in the order in which they are integrated from the beginning to the end of the automation line. Process control software is essential to manage the complex data that an automation system generates, representing status flags of devices, location of specimens, and instrument status. Automation centrifugation is occasionally selected as an offline function because laboratories that adopt Lean processes can generally perform manual centrifugation more rapidly and efficiently than an automation system. Most customers purchase the decapping module because decapping can lead to aerosol generation and repetitive stress injury. The storage and retrieval module is an absolute requirement for efficient specimen buffering in an automation system, since not all specimens can be assayed in a linear fashion. (Buffering modules should not be confused with long-term specimen storage devices [automated storage and retrieval available].) Aliquotting is present on some systems but not on others because the goal of a Lean laboratory is to eliminate the need for aliquotting. Specimen recapping and sealing is an essential tool because it eliminates a boring laboratory job that leads to repetitive stress injury. Lab automation systems that do not track specimens within the system create a challenge for laboratories that wish to report specimen status in real time to customers, and for laboratories that wish to quickly locate misplaced specimens.
System architecture addresses a key question in laboratory automation today: Does the customer wish to have its vendor deliver a complete solution or to select its own analyzers based on specific analytical needs? Fourth-generation automation systems will allow customers to mix and match automation components and analyzers. Adequate automation support will be needed to help customers integrate their laboratory information system with their lab automation system. Laboratories should look for software systems with elegantly designed user interfaces and common database architectures (for example, SQL). To test the intuitive design of the interface, prospective buyers of automation should visit a working automated laboratory and try to operate the software without prior training. Incorporating many of the LIS functions into the LAS process controller has its benefits. Thus, software features (from patient demographics to evaluates validity and releasability of results) allows laboratories to determine where the patient and sample-based information will reside, such as patient demographics, rules engine, reporting, Internet connectivity, help screens, quality control data, statistical reports, and auto-validation. The remaining features listed (from specimen tracking to supports CLSI standards) focus on how the process controller manages the flow and tracking of specimens to include retrieval from storage for repeat/reflex/dilution/add-on. Flexibility of hardware configurations is addressed there again as well.
In factories, which have managed complex automated processes for decades, it is known that the use of a process control architecture is the most critical component of an automated system. Laboratories should be cautioned that not all LISs will support an efficient interface with the process controller. Therefore, some laboratories strongly consider either purchasing an LIS that supports process control or purchasing a process controller that already includes many of the traditional LIS functions. Process controllers with integrated LIS functions will appear on the market soon.
The number of systems installed in North America, Europe, and Asia will give you an idea of the global penetration of each vendor’s market. Vendors with relatively few installations in any one market segment may be willing to offer the purchasing laboratory a price discount in order to develop a local demonstration site. Once installed, automation systems should give longer periods of error-free and maintenance-free performance compared with clinical analyzers. Thus, local service may not be as critical for automation systems. For example, the transportation system is usually the most robust part of an LAS and may last up to a decade or longer.
The basic dimensions of an automation component are important from a flexibility standpoint (modular HW/installed options/device can operate in track and manual mode). Long track systems offer little flexibility in tight spaces and limit the selection of analyzers that might be interfaced if lateral floor space is required near the track (most systems accommodate parallel or perpendicular attachment of analyzers). The importance of vendor adherence to the CLSI standards (Auto 1-5) cannot be overemphasized because they are the current optimal specifications that have come from years of trial and error and represent the opinions of many notable experts.
Versatility in handling many tube sizes has long been debated. One approach to process efficiency is to standardize the size of tube your customers use to send blood and body fluids to the laboratory. This results in the greatest institutional savings and greatest gains in process efficiency, but not all laboratories can achieve this degree of influence over their customers. Thus, laboratories may not wish to limit the tube sizes accepted by the LAS (containers device accommodates). Pediatric tubes are rarely accommodated directly; they are generally poured off into a smaller conical tube placed in the top of a standard 5-mL or 10-mL test tube. Once the tubes are placed on the automation system, conveyance throughput is important because it can become a rate limiting step over which the laboratory has no further control (a conveyor system usually cannot be speeded up once it is purchased and installed). Intelligent specimen rerouting will make it unnecessary for a technologist to intervene in the process of deciding where specimens have to be moved in subsequent processes (automatic rerouting for reflex/repeat/dilutions).
The impact of the LIS on laboratory support services is addressed in the transportation system section (from modular HW to required maintenance). The industry is moving toward modular hardware designed around a plug-and-play configuration. Generally, floor-mounted systems are the easiest to install, but past successes with wall-mounted systems pioneered by the late Dr. Masahide Sasaki (in his legendary laboratory at Kochi Medical School, Kochi, Japan) resulted in a space-efficient system. Some systems can be operated as standalone units, termed task-targeted automation, or will function only in track-mounted mode. Versatility in equipment design will allow the automation to be redeployed should the laboratory’s needs change (can operate in track and manual mode).
A useful feature of some automation systems is the placement of all utilities in the space underneath the track. When utilities are integrated, the finished product is efficient and visually appealing. Analytical systems that eliminate the need for water supplies and waste drains can reduce laboratory renovation costs considerably (required utilities). Automation systems usually require only weekly preventive maintenance. However, the replenishment of tips for aliquotters and caps for recappers may require scheduled daily maintenance (required maintenance). Look for systems with self-washing probes to save material and labor costs.
Another point of industry debate is the efficiency of single- versus multiple-tube carriers (carrier type). Single-tube carriers allow the greatest flexibility in routing specimens and optimizing system throughput, but no significant degradation in system performance is seen with multiple-tube carriers. Laboratories are advised to use a process simulation tool to determine if the choice of multiple specimen carriers will influence turnaround time significantly. Similarly, a scalable system will be able to accommodate a growth in laboratory business. One of the great benefits of an automation system is the ability to increase laboratory productivity by up to 40 percent without having to incur higher labor costs.
As indicated earlier, automated centrifugation may be considered a luxury because attentive technologists can often outperform an automated centrifuge attached to a track. However, in laboratories experiencing labor shortages, with technologists often diverted to other important tasks, an automated centrifuge eliminates the bottlenecks associated with specimens waiting to be loaded or unloaded. Since most commercial automated centrifuges are limited to throughputs of between 250 and 400 tubes per hour, two systems are required for laboratories with large peak demand. The more versatile automated centrifuges will adjust their spin times based on specimen need (greater times or speeds to obtain platelet-free plasma).
A key component of the ideal automation system is an automated input/accessioning device. Countless errors can be avoided and much time can be saved by automating the task of entering each specimen into the LIS or process control computer. However, the full benefit of an automated accessioning unit can be realized only if specimens are bar-coded at the patient bedside. If a manual accessioning system is required in the laboratory, an automated input/accessioning device is still useful for sorting those tubes that need centrifugation and other preanalytical tasks from those that may be shunted directly to the analytical process (for example, hematology tubes).
Automated decapping will be a welcome addition to any laboratory. Versatility in cap removal is a key selling point for a decapper because tube heights and diameters vary and there are both pressure fit and screw cap tubes. Engineers of decappers have achieved relatively high throughputs; therefore, only one unit is generally required for most LAS lines. Look for systems where the disposal chute (where the contaminated caps are deposited) and the waste bin are easy to clean and maintain.
Automated sorting can reduce processing time by routing tubes efficiently to the appropriate analytical station. Furthermore, sorting tubes into analyzer-specific racks makes it possible for the technologists to move these racks manually and quickly to distant processes. User-programmable sorting by priority (stat versus routine) or specimen destination (hematology versus chemistry/immunoassay) can be useful for optimizing throughput. An added feature that can automatically perform quality assessment is a specimen integrity monitor, which is usually integral to today’s analyzers. However, it would be beneficial to flag and divert specimens with quality issues (excessive hemolysis, icterus, and lipemia) before they get beyond the sorting function. A system available in Australia, the Pathfinder, is one of the first commercial preanalytical processors to incorporate a machine vision system for sample inspection.
The next step in the process, sample aliquotting, is best avoided altogether. Expanding a single sample tube into three aliquots essentially quadruples the number of sample handling tasks, sample inspection tasks, and medical waste. One way to avoid alliquotting is to draw each blood at the time of phlebotomy so that it may be directed to an analytical station once it enters the laboratory. Thus, minimal aliquotting is recommended. However, an aliquotter can make the preparation of send-out tubes (destined for commercial labs) more efficient. Aliquotters should be able to accommodate a large number of input tubes (automated aliquotting dimensions) as well as create and apply daughter tube labels with bar codes and human-readable information.
Attaching analyzers to automation systems is where most of the contention arises during the installation of an LAS (instrument interfaces). The instrument must be mechanically interfaced to the LAS to gain access to the specimen. Furthermore, each analyzer must be electronically interfaced to the LAS. Not all LAS vendors will accommodate a full bidirectional rules-based instrument control interface. Using a partial or unidirectional interface between an analyzer and an LAS, however, will result in a compromise in throughput and centralized knowledge of how the analyzer is performing. With regard to mechanical interfacing, most analyzers will accommodate the CLSI Auto 5A interface standard, which defines how the analyzer must be engineered to perform point-in-space sampling of specimens temporarily positioned by the conveyor system. Robotics arm interfaces still exist for some analyzers, and despite their rather large size, they allow robust and error-free movement of samples to the analytical systems.
Postanalytical processing is almost as important as preanalytical processing for improving laboratory efficiency because up to 20 percent of laboratory labor can be spent on postanalytical tasks. Recappers ensure long-term specimen integrity in the refrigerator or freezer. Foil-based recappers allow recapping material to be stored in large quantities on a roll, making it unnecessary to replenish caps regularly. Furthermore, foil-based systems allow piercing of the foil lid for postanalytical reanalysis and then replacement of yet another foil cap on top of the previous recap. Once tubes are recapped, then automated refrigerators can store and retrieve up to 10,000 tubes in one device (automated storage and retrieval). Though it’s an expensive option, the convenience may be justified in laboratories where frequent reanalysis is necessary.
In summary, certain automation functions are essential for maximum efficiency while others may be considered luxuries. LAS selection should take into account the current and future needs of the laboratory. Given the relatively simple nature of automation systems, they should provide laboratories with many years of trouble-free service and a significant improvement in laboratory efficiency and patient safety.
Dr. Felder is professor of pathology and director of the Medical Automation Research Center, University of Virginia, Charlottesville.