Automation in the clinical laboratory

Background

Automation has dramatically changed both the analytical and nonanalytical aspects of clinical laboratory operations. Automation of laboratory test procedures began more than 50 years ago, but nonanalytical automation—including conveyor systems, interfaced analyzers, and automated specimen processing and storage—began in earnest in the 1990s. Today there is a wide selection of automation options designed to improve the quality, throughput, and efficiency of laboratory testing.

Content

This chapter covers automation from both nonanalytical and analytical perspectives. Historical contexts are provided. The discussion of preanalytical automation includes a review of labeling, barcoding, and portable wireless labeling systems along with the use of pneumatic tube systems and mobile robots for transport of specimens. Single-function robotic systems and multifunction systems for specimen processing are discussed. Some of these systems have pre- and postanalytical capabilities. Total laboratory automation (TLA) systems are discussed extensively. Several TLA systems include postanalytical functions; thus the chapter also addresses storage and retrieval systems. The second half of the chapter discusses automation from the analytical perspective, including specimen and reagent handling on analytical instrumentation, as well as common measurement approaches used by automated analyzers. The chapter concludes with area-specific considerations of how analytical automation has impacted all subdisciplines of laboratory medicine.

Specimen identification

Typically, the identifying link (identifier) between the patient and specimen is noted at the patient’s bedside. Maintenance of this connection throughout (1) transport of the specimen to the laboratory, (2) subsequent specimen analysis, and (3) preparation of a report is essential. Several technologies are available for automatic identification and data collection purposes ( Box 29.1 ). In practice, automatic identification includes only technologies that electronically detect a unique characteristic or unique data string associated with a physical object. For example, identifiers such as (1) serial number, (2) part number, (3) color, (4) manufacturer, (5) patient name, (6) medical record number, and (7) accession number have been used to identify an object or patient through the use of electronic data processing. In a clinical laboratory, labeling with a bar code has become the technology of choice for purposes of automatic identification.

Specimen preparation

The clotting of blood in specimen collection tubes, their subsequent centrifugation, and the transfer of serum to secondary tubes require a finite time to complete. If performed manually, the process may result in a delay in the preparation of specimens for analysis. To eliminate the problems associated with specimen preparation, systems are being developed to automate this process.

Specimen delivery

Automated methods are often used to deliver specimens to the laboratory instead of more manual methods (e.g., phlebotomist transport, courier service, nurse delivery). Pneumatic tube systems have been used in many hospitals, but mobile robots can also transport specimens and other laboratory supplies within a laboratory or to the laboratory from other locations in the facility.

Single-function workstations

Some laboratories have gained efficiency, improved quality, and lowered turn-around time by implementing an automated system that performs a single task. This automation is also referred to as task-targeted automation . Examples of such automation include automated centrifuges, decappers, recappers, aliquotters, and sorters. Such systems began to be developed in the 1990s and early 2000s, when TLA systems were relatively expensive and thought to be only affordable to reference laboratories and the largest hospital laboratories. Today these systems continue to be developed and implemented because they offer affordable solutions for specific customer needs. Box 29.3 lists a number of vendors offering one or more single-function workstations. Fig. 29.1 is an example of a sorter, one of the different types of single-function workstations that are available. However, a laboratory could use a sorter for both preanalytical sorting for different testing areas and postanalytical sorting for archival storage.

Multifunction workstations (automated specimen processing)

Although the manual operations carried out in a specimen processing area may look simple, considerable complexity underlies them. Consequently, specimen processing has been one of the most difficult areas of the clinical laboratory to automate. Each specimen passing through a specimen processing area has to undergo a series of operations, including (1) receiving the specimen; (2) inspecting it for appropriateness (labeling, container type, temperature, and quantity of specimen); (3) logging the specimens into the LIS; (4) recording the date and time of collection and phlebotomist ID, if not already entered; (5) labeling the specimen with an accession number if not already there; and (6) separating urgent and stat specimens from routine specimens. Also, specimens may have to be sorted for centrifugation and aliquoting and may have to be sorted or otherwise prepared for the appropriate laboratory station.

An example of a stand-alone specimen processing system is shown in Fig. 29.2 . Such systems place processed specimens into racks that must be transported manually to the testing areas, with some exceptions. Some of these systems are approximately the size of a large, automated analyzer; others may be slightly larger. They may be a good choice for laboratories (1) with daily workloads of 500 to 1500 specimens, (2) with space limitations, or (3) that desire ease of use with different analyzers from different vendors. Some laboratories may choose to use multiples of a stand-alone specimen processing system based on higher daily volumes or to automate postanalytical archiving in addition to preanalytical specimen processing.

Box 29.4 lists vendors offering multifunction or specimen processing workstations. Although there may be some variation in the functions that are included, these systems will typically (1) receive incoming specimens, (2) sort, (3) decap, (4) aliquot, and (5) label aliquot specimen containers with bar codes. All may be interfaced to the laboratory’s LIS. Some systems may also include automated centrifugation. Some of the systems sort into instrument-specific racks for analyzers from a number of different vendors. In addition, some users apply these systems to aliquot and sort reference or “send-out” testing, saving considerable time in locating the original specimens after testing in their own laboratory.

Total laboratory automation systems

Several manufacturers offer integrated or modular automation systems that use conveyor belts to connect preanalytical specimen processing and other functions directly to analyzers. Such systems may also include functions such as postanalytical storage or sorting of specimens or aliquots to be transported to low-volume testing areas. A list of automation vendors offering TLA systems can be found in Box 29.5 ; Fig. 29.3 illustrates a large TLA system with a variety of connected nonanalytical and analytical modules.

In addition to the nonanalytical functions listed in the preceding section, TLA systems typically add (1) conveyor transport; (2) interfacing to automated analyzers; (3) more sophisticated process control; and, in some cases, (4) a specimen storage and retrieval system. All systems are of modular design, allowing the customer to choose the modules or features that should be included. Some systems use an open design, which permits interfaces to analyzers from a variety of vendors, but other systems are of a closed design and are interfaced only to the vendor’s own (or a limited number of other) analyzers. Closed systems typically do not have process control software that is independent of the instruments or system; rather, the automation process control is integrated to work with the vendor’s analyzers.

To achieve maximum effectiveness of an automation system, process control software should be able to read the specimen’s ID bar code and obtain information from the LIS about specimen type and ordered tests. This process control software is usually referred to as the Laboratory Automation System (LAS), not to be confused with the actual hardware of the laboratory automation. Readers are referred to the CLSI standard on automation communications (AUTO03-A2) for further information on the LAS. The software should then determine the processes the specimen requires and the exact route or course of action for each specimen. It should be able to (1) calculate the number of aliquots and the proper volume for each specimen depending on the tests requested, (2) route the specimens to preanalytical processing equipment and then to analyzers, (3) recap the specimens after testing, and (4) retain the specimens for automatic recall. The software should be able to monitor analyzers for in-control production status and automatically make decisions if a test is not available. Specimen integrity for hemolysis, icterus, and lipemia should be performed. Having the automation system do specimen integrity checking, if possible, may be more efficient than having it conducted farther downstream in analyzers. However, automatic spectrophotometric methods are available on many modern analyzers. Rules-based decisions from analysis at multiple wavelengths may correct the effects of interferents for some assays and alert the laboratory on others. Finally, most process control software should include (1) “autoverification,” which is validation of analyzer results by making rules-based decisions that flag exceptions for technologist review, and (2) “autoretrieval” of specimens for repeat, reflex, and dilution testing. Assistance with designing an autoverification system can be found in the CLSI Guideline AUTO10-A.

Although most of these systems are restricted to handling specific types of specimen containers, they are capable of processing much of the daily workload of a large clinical laboratory. A few laboratories with daily workloads as low as 600 to 800 specimen tubes have justified these systems because of a shortage of technical help, but typically these systems are designed for laboratories with workloads of 1000 to 25,000 specimens per day. In addition to process control software and an LIS interface, each of these systems incorporates some or all of the following components:

• 1.

  Specimen input area: This is a holding area where bar code labeled specimens are introduced into the system.

• 2.

  Bar code reading stations: Multiple bar code readers are placed at critical locations in the processing system to track specimens and provide information for their proper routing to various stations. Alternately, bar code readers may only be required at limited locations in the system if each transport carrier (puck) has a unique ID such as a radiofrequency identification (RFID) chip, and the specimen’s bar code ID is “married” to the puck’s ID when first placed in the puck.

• 3.

  Transport system: Segments of a conveyor belt line that move specimens to the appropriate location.

• 4.

  A high-level device to sort or route specimens: A device that separates specimens based on processing requirements or order code and passes them to the transport system or to a system using racks. A high-level sorter is often used to separate specimens that require centrifugation or other processing steps from specimens that do not or to route specimens into completely different pathways within the total automation system.

• 5.

  Automated centrifuge: An area of the specimen processor in which specimens requiring centrifugation are removed from the conveyor belt, introduced into a centrifuge that is automatically balanced, centrifuged (refrigerated or at room temperature), and then removed from the centrifuge and placed back on the transport system.

• 6.

  Level detection and evaluation of specimen adequacy (specimen integrity): An area where sensors are used to evaluate the volume of specimen in each specimen container and to look for the presence of hemolysis, lipemia, or icterus. These sensors can be located on one of the chemistry analyzers on the system and data shared across the laboratory or can be located in one dedicated area.

• 7.

  Decapper station: An area or device in the automated system where specimen caps or stoppers are automatically removed and discarded into a waste container.

• 8.

  Recapper station: An area or device in the automated system where specimen tubes are automatically recapped with new screw caps, stoppers, or an alternate air-tight closure.

• 9.

  Aliquotter: A machine that aspirates appropriately sized aliquots from each original specimen container and places them into bar-coded secondary specimen containers for sorting and transport to multiple analytical workstations. The LAS generally instructs the aliquotter as to how many aliquots of what volumes are required based on the bar code ID of the specimen.

• 10.

  Physical interfaces to automated analyzers: A direct physical connection to an automated analyzer that permits the analyzer’s sampling probe to aspirate directly from an open specimen container while the container is still on the conveyor or that may robotically lift the container from the conveyor and place it onto the analyzer. Whereas some automation systems interface only to their own brand of analyzers or to a limited number of systems, other automation systems use a so-called open design that complies with CLSI standards and permits interfaces to a variety of automated analyzers.

• 11.

  Sorter: An automated sorter to sort specimens not going to a conveyor-interfaced analyzer or workstation. Such a sorter typically sorts into 30 to 100 different sort groups in racks or carriers. In some systems, the racks may be specific to certain analyzers for convenience. Some automation systems may use a sorter as a temporary holding area for specimens that may require a repeat test (see take-out stations).

• 12.

  Take-out stations: Temporary storage areas for specimens before or after analysis. The take-out station may be the same as the sorter described earlier, where specimens are sorted for manual delivery. However, it may also serve as a holding area (stockyard) for specimens awaiting autoverification of results in case a repeat test is required.

• 13.

  Storage and retrieval system: This unit may serve the same function as the take-out station or stockyard—that of holding specimens after analysis in case a specimen is necessary for a repeat test—but it has one major difference. These units are typically refrigerated and hold many more specimens (3000 to 30,000) than the typical take-out station or stockyard. Depending on daily workloads, the laboratory may be able to retain up to 1 week’s worth of specimens for possible repeat or additional tests. Specimen containers are loaded and retrieved with a robot. Some storage systems include automated specimen discard directly into biohazard waste bins after a predetermined storage duration has been reached.

Conveyor belts

One main feature of integrated TLA systems is the use of conveyor belts. Ordinary industrial conveyor belts have been used successfully when only transportation is required. However, in the past, when conveyors were integrated with other robotic systems to automate pre- or postanalytical functions, this technology had difficulty handling the large variety of specimen containers found in the clinical laboratory. To increase the variety of types of specimen containers that are carried on a conveyor belt system, specimens are placed into specially designed carriers that fit on the conveyor belt line. Sometimes known as “pucks” or “racks” (depending on whether they carry individual specimens or groups of specimens), the carriers have receptacles for variously sized tubes, generally ranging from 13 × 75 mm to 16 × 100 mm—sizes that are consistent with the CLSI Standard AUTO01-A. The industrial, motor-driven conveyor belts used in typical LASs have transport rates of a few hundred pucks per hour up to 2000 per hour. In addition, linear synchronous motors and magnetic pucks are now being used in our clinical laboratory ( https://www.aruplab.com/news/06-22-15 ). This technology, which is considerably faster than present conveyors and has virtually no moving parts, may become the next generation of specimen transport for clinical LASs.

Two types of conveyor belt systems are typically used as illustrated in Fig. 29.4 . Fig. 29.4A depicts a loop conveyor that has a single module for both input of new specimens and removal of completed specimens. The flow around the loop takes specimens to the processing and analytical modules. Specimens may be sampled directly by the analytical instrument while on the conveyor, or a robot attached to the workstation may remove selected specimens from the conveyor for analysis. Fig. 29.4B depicts a unidirectional conveyor in which specimens are inputted at one end of the belt, flow past various processing and analytical modules, and arrive at the opposite end where they are removed. Analyzers access specimens in the same way as the loop conveyor. Depending on the vendor, both of these conveyor types can have bypass lanes that enable specimens not designated to stop at certain modules to bypass them entirely and proceed to other modules. This approach has the advantage that it does not require that specimens be aliquoted because specimens pass by all workstations where tests are performed. If specimens are robotically removed from their carriers on the line for testing, systems for queuing empty carriers to return the tubes back to the conveyor and to identify which specimen is in each carrier are required.

Transfer of specimens from the conveyor belt to the laboratory workstation has been implemented in various ways. For example, some manufacturers have equipped their laboratory instruments with devices to move specimen containers from conveyor belt systems onto the analyzers, and others have installed the ability for the analyzer to sample directly from specimen containers remaining on the conveyor. In practice, the automation system requires a device that stops the tube in the exact location required by the analyzer for sample transfer and verifies and transfers the tube’s bar code identification to the analyzer. The CLSI Standard AUTO05-A specifies a common “point in space” that is this exact location for all analyzers and automation conveyor systems.

Depending on the particular automation system, specimens for “STAT” testing may or may not be processed and analyzed at speeds that are faster than if manually handled. Laboratories that are purchasing LASs are advised to learn how each system being considered handles STAT specimens.

Automated specimen storage and retrieval

The automated capability to store and retrieve specimens on demand and with readily known exact storage locations is an important aspect of automated specimen delivery systems. In addition to automated storage and retrieval options in some of the integrated systems described earlier, several automated or semiautomated options for storage and retrieval, as well as LIS modules and PC-based software systems, permit laboratories to track trays or racks of specimens in their own freezers or refrigerators. Some large reference laboratories have adapted large storage systems commonly used in other industries into their laboratory settings. One such large system, in use in the authors’ laboratory (ARUP), is depicted in Fig. 29.5 .

Evaluation of requirements

Any consideration of total or modular laboratory automation should start with an evaluation of requirements. Such an evaluation begins with mapping of the current laboratory workflow from the arrival of patient specimens through completion of testing and reporting of results. Box 29.6 lists potential workflow steps that should be mapped. Mapping of material (specimen) flows and data flows is directly related to process flow and assists the laboratory in determining process steps that (1) are bottlenecks, (2) waste labor, and (3) are prone to error. Workflow mapping thus enables the laboratory to better identify what steps should be considered for automation.

Some laboratorians use an “80% rule” in guiding decisions about automation. Clinical laboratories have many exceptional tests, specimen containers, and handling situations. Nevertheless, if 80% of the daily workload of specimen containers, covering most routine handling situations, can be standardized and automated, the laboratory will achieve a dramatic reduction in its labor usage and costs, which should be sufficient to justify the investment in automation and the planning and evaluation time involved.

After the laboratory’s existing workflow has been mapped and its requirements have been identified, alternative solutions are considered. Consideration of lean processing and efficiency can help to ensure that automated solutions ultimately improve operations. Vendors are invited to make presentations and to host visits of the laboratory management team to other laboratories where vendors have successful installations. It is important at this stage to focus on the requirements identified by workflow mapping and to not allow the vendor to try to sell equipment that may not be necessary. Key points to remember when a laboratory is considering either new automation or an addition to existing automation are listed below. Although not every circumstance may require assessment of each of these points, the list is a useful outline of a systematic approach to automation that has worked well in the authors’ laboratory and others.

Problems of integration

Building a highly integrated laboratory generates many potential problems. Most vendors of clinical LASs prefer customer settings in which the integrated analyzers are their own brand. However, many laboratories may prefer to use analyzers from multiple vendors, including the automation system vendor, making integration of instruments and robotic devices from different manufacturers necessary. Decisions must be made concerning which device will be the master controller and which vendor will develop the software that provides overall control of the automation scheme. In addition, individuals or firms that will be responsible for configuration of the automation to the geometry and production schedule of the laboratory must be recruited and trained. Over the past decade, automation vendors have gained considerable experience implementing integrated systems. Although some clinical laboratories have had unsatisfactory experiences, integrated systems are becoming more commonplace.

Readers are referred to the CLSI standard AUTO03-A2 and, in particular, to the Functional Control Model (Section 5.2), which describes the relationships among the LIS, LAS, and various devices. Most often, the LAS has the requisite process control software to support automation. The functional control model, which is depicted in Fig. 29.6 , supports analytical instruments that may be physically attached to the automation system and analyzers that may not be attached but are still interfaced to the LIS. The model does not give dominance to the LIS or the LAS but rather allows for essential information flow in either direction to make the most efficient use of the strengths of each system. Although not specified in the AUTO03-A2 standard, some vendors have implemented redundant clustered computers as their LAS to protect customers from downtime associated with failure of a single computer.