Clinical Analyzers. General Chemistry - Analytical Chemistry (ACS

---

Anal. Chem. 1999, 71, 351R-355R

Clinical Analyzers. General Chemistry Larry E. Schoeff

Department of Pathology, University of Utah, 50 North Medical Drive, Salt Lake City, Utah 84132 OVERVIEW In my last review (I1) I focused on trends in preanalytical (sample processing), analytical (chemical analyses), and postanalytical (data management) phases of automation as they merged closer to an integrated total laboratory automation (TLA) system. Changes and improvements with popular general chemistry analyzers were also highlighted. From my conclusion in that review, most of the predictions for the future have become reality or at least have a foothold in the door to the present. Analyzers have continued to perform more cost-effectively and efficiently with expanded menus of tests along with a new focus on combining measurement techniques of traditional chemical analyses and immunoassays on a single analyzer. More system and work flow integration has occurred with robotics and data management for more inclusive TLA. However, a new alternative approach has emerged for many laboratories that is more palatable than the cost-prohibitive, all-inclusive TLA; that alternative is modular, or stepwise, automation. This new approach to automation and its attendant set of definitions is reviewed below and compared to a TLA approach. Currently, the major emphasis with this modular automation is front-end or sample processing. The mainstream companies and their modular/TLA automation products are also reviewed here. Also predicted from the last review was that more in vitro diagnostic (IVD) companies will form alliances to get their instrumentation products into laboratories. This has proven to be an understatement. So many alliances, mergers, acquisitions, partnerships, collaborations, etc., have occurred in the last 2-3 years, that it is difficult to keep track of “who owns who” with the big-name IVD companies. The number of these mainstream companies is definitely shrinking as they strategically position themselves in the intensely competitive marketplace, not only with their own evolving analyzers but with the relatively new frontend automation products that complement and synergize their analyzer systems. These same companies also have a very keen interest in the major advances (as predicted) being made by smaller companies with biosensors and chip technology, which will revolutionize automated analyzers as we know them now into miniaturized and microscopic components and systems. The major IVD companies and their latest analyzer systems are also highlighted in this review. Finally, a few more predictions are offered as a reality check for the next review. NEW CONCEPTS IN AUTOMATION According to laboratorians who ranked the top 10 issues of 1998, bringing new technologies into the laboratory was rated the sixth biggest challenge to face (I2). Not only did this issue include the hottest new assays approved by the FDA in 1998, e.g., free PSA and homocysteine, but also new concepts in automation. These new automation concepts and solutions significantly im10.1021/a1999909j CCC: $18.00 Published on Web 05/20/1999

© 1999 American Chemical Society

pacted the top two issuesscost containment and personnel management. The pressures of healthcare reform and managed care along with a shrinking number of employees and larger test volumes are some of the major problems facing laboratories today. Many laboratories across the country have undertaken major reconfiguration initiatives to centralize and/or regionalize their laboratories in order to reduce costs and improve service (I3-I7). Central, or core laboratories allow much more flexibility to integrate and improve work flow and efficiency. This increased flexibility has forced laboratorians to rethink the way they purchase laboratory instruments. Traditional ways of purchasing instruments have gone the way of “fee-for-service” healthcare. Buying laboratory instruments is more complicated these days. Now, diagnostic companies are willing to make deals with buyers using terms and techniques previously considered unthinkable (I8). Much more time and effort is expended on the planning phase toward purchasing an instrument before acquisition and implementation occur. A successful transition for each new instrument often requires months of careful planning (I9). In recent years, some instrument vendors have incorporated work flow and process analysis into the services they provide for laboratories. The software tools have been quite helpful for laboratories to identify the kinds of instruments they need to reconfigure a more efficient laboratory. One such program called Prospectus is offered by Dade Behring (Deerfield, IL) and enables laboratory managers to perform comparative cost accounting across different instrument platform combinations as they scrutinize that big decision on the “right” analyzer to acquire (I8, I10). The companion piece of software by Dade Behring called Prognostics is labeled as a probabalistic event-based simulation modeling program that assesses labor and process time for various scenarios (I8). The so-called “right” analyzer to purchase these days is increasingly likely to be a multiple platform instrument that performs tests from two or more disciplines with traditional chemistry assays mixed with immunoassay capabilities. An automated stand-alone single-discipline instrument, even with significant internal automation to perform autorepeat, autodilute, and reflux testing is not flexible enough to achieve the efficiency and cost savings that laboratories must have today. Other new concepts that have taken hold are modular work cells and modular clusters, whereby two or more instruments of the same discipline (work cell), or of different disciplines, are linked by a single controller, with additional instruments added as needed. These new work flow and testing combinations of technologies and disciplines with platforms, work cells, and clusters have given rise to yet another new automation strategy, i.e., modular automation. As a much more financially feasible alternative to TLA, which is simply out of the question for many laboratories, stepwise, Analytical Chemistry, Vol. 71, No. 12, June 15, 1999 351R

modular automation can offer even small laboratories an opportunity to begin to automate their front-end processing of specimens. With ever increasing labor costs, specimen handling errors, and larger test volumes, laboratories are anxious to automate wherever possible to experience savings and improve efficiency. The most labor-intensive and error-prone part of the laboratory testing process is the preanalytical phase, or sample processing, which includes sorting, cap removal, aliquotting, centrifuging, recapping, etc. Today, laboratorians are realizing their biggest cost savings and improved productivity with multiple platforms, modular work cells, and clusters that are integrated with front-end modular automation or total laboratory automation. As laboratories evaluate each of these new concepts for implementation, the potential is great to enhance the laboratory’s value in the following areas: process efficiency, menu consolidation, labor utilization, clinical priority, and operating cost reduction (I11). MODULAR AUTOMATION VS TOTAL LABORATORY AUTOMATION A record-breaking number of companies displayed many new technologies and automation products at the American Association for Clinical Chemistry (AACC) Clinical Laboratory Exposition in August 1998. In fact, alternative solutions to laboratory automation systems offered by IVD companies dominated the exposition. Stepwise or modular automation received much of the attention there as an attractive alternative to costly total laboratory automation. Continuing with this lexicon of new terminology, a modular automated system from a single vendor allows linking of that vendor’s instruments, along with front-end and/or back-end automation operated by a single controller with additional components added as needed. An integrated modular system is another alternative whereby two or more analytical modules share a “sample- and reagent-handling” preparation and transport system. And finally, custom-automated systems and institutional automation are two types of total laboratory automation that employ multivendor instruments linked to front- and back-end components and fully automate each section of the laboratory process. Their only difference is the degree of throughput possible within each discipline (latter is highest) (I11). Modular, stepwise automation is on track to become the most common strategy in the future of diagnostic testing, since most laboratories will not be able to afford TLA (I12). A late 1996 survey of 950 laboratory managers, mostly from 200 to 400 bed hospitals, found that 16% of respondents already had automated sample preparation in their laboratories and 47% of small to midsize hospitals targeted front-end automation as their first planned automation project within the next two years. A total of 63% of hospitals over 400 beds and 62% of reference laboratories shared the same preference (I13). “Increased productivity” and “faster turnaround times” were reported as the top two reasons to automate. The top two reasons NOT to implement laboratory automation were “too expensive” and “not enough testing volume”. Whereas TLA systems automate all five steps of the testing process, i.e., preprocessing, processing, analysis, postanalysis, and storage/retrieval (I14), a more gradual stepwise approach allows most laboratories to realize significant benefits from basic and flexible front-end automation, after which time they can purchase additional components that will grow with their needs. A minimally 352R

Analytical Chemistry, Vol. 71, No. 12, June 15, 1999

configured preanalytical processor can save up to 2.5 FTE in a medium-sized laboratory (I15). Success with new automation initiatives, either modular or total has required laboratories to consolidate their workstations by reorganizing laboratory space and staffing and purchasing flexible high-throughput instruments (I16). Essential to modular automation is coordinating and optimizing individual processes to eliminate redundant steps and repetitive loops, so as to reorganize the diagnostic testing into the most efficient process possible (I17-I19). Modular systems can be sold preconfigured for most laboratories, while large TLA systems require much customization. Case studies abound with the testimonials of clinical, R&D, and pharmaceutical laboratories that have scrutinized the pros and cons of laboratory automation and have acquired modular or total systems (I10, I20-I26). LABORATORY AUTOMATION SYSTEMS The number of companies marketing laboratory automation systems has increased since the last review, due to significantly more laboratories pursuing the more affordable stepwise approach with modular components (I27). Abbott Laboratories (Abbott Park, IL) recently formed a strategic alliance with TECAN Corp. (Hombrechtikon, Germany) to market front-end automation systems (I28). The long-term goal of this alliance will be two Genesis systems to automate sorting, centrifuging, decapping, aliquotting, barcode labeling, and racking of samples. The Autolab (MDS, Etobicoke, Ontario) Materials Handling System has five modular components designed to stand alone or be integrated into a fully automated preanalytical sample-processing system. These components are as follows: Autoload, Autofuge, Autoquot, Robotic Workcell (barcoding and routing), and Autosort. Their System Console is the single controller with the laboratory information system (LIS) and can connect any number of autolab modules with the LIS through one interface. Beckman (now Beckman Coulter, Fullerton, CA) still markets the AccelNet laboratory automation network to automate sample processing. A robotic arm moves samples down a transfer lane onto adjacent components of AccelNet that barcode and route, sort, perform serum endices, load/spin/unload tubes at the centrifuge, decap tubes, load samples and run analyzer, interrupt for stats, report results, and sort for storage. AccelNet integrates two synchron CX7 Delta analyzers. Beckman’s new Power Processor (formerly the Coulter/IDS system) is a more sophisticated second-generation processing workstation that eliminates the robotic arm. It has an inlet station for loading samples into tube holders for independent movement, a centrifuge station, decapping unit, and outlet station for organizing racks of tubes for placement onto the analyzer. The LXpress system includes a Power Processor, a Synchron LX4201 with DataLink for data management, and an Access immunoassay analyzer consolidated into one workstation. With the ChemXpress option added as an online connection to load the racks of tubes onto the analyzer, the LXpress is the ultimate system for maximal throughput (I29). The Hitachi Clinical Laboratory Automation System (CLAS) by Boehringer Mannheim (now Roche Diagnostics/Boehringer Mannheim Corp., Indianapolis, IN) was one of the earliest players in the automation market and is now the second most popular system in U.S. laboratories. It automates sorting, centrifuging,

decapping, aliquotting, barcode labeling, and transport to the analyzer for a total or modular system. Up to six analytical modules using Hitachi 917 and 747 and Elecsys (electrochemiluminescence) analyzers can be linked to CLAS for their company’s answer to modular systems (I30). The largest selling automation system in the United States is the Coulter/IDS Sample Transport System (now the Power Processor) with its automated sample preparation components by Coulter Corp. (Miami, FL) (now Beckman Coulter). The Coulter/ IDS employs a “U” lane for direct sampling from a specimen conveyor belt without using intervening robotics. It was the first system to be installed in a U.S. hospital in 1995 at the University of Virginia. The Coulter/IDS had already gained widespread acceptance in Japan in earlier years. The sample preparation components are similar to those already discussed. Beckman Coulter provides an upgrade path with additional conveyor belt modules to integrate chemistry, hematology and other analytical workcells. HelpMate Robotics Inc. (Danbury, CT) is the only notable company that markets mobile robots capable of transporting medical specimens throughout a hospital, dropping them off at specified points, e.g., a sample preparation station. Joseph Engelberger, widely acknowledged as “The Father of Robotics”, is chairman of the board of HelpMate Robotics. Ortho-Clinical Diagnostics (Raritan, NJ) formerly Johnson & Johnson Clinical Diagnostics have formed alliances with Coulter and LAB-Interlink to develop and distribute automated work cells that incorporate their Vitros line of analyzers. The installation at the University of Virginia was a Vitros 950AT with the Coulter/ IDS Transport System. At AACC in August 1998, Ortho unveiled its LAB-Frame SELECT which was developed with LAB-Interlink. This system integrates two to eight of Ortho’s chemistry analyzers and immunoassay and hemostasis instruments, providing complete sample management from receipt to storage. LAB-Interlink (Omaha, NE) is number three in the U.S. automation market. Their LAB-Frame Custom Systems emphasize customization to fit the client’s clinical and technical needs. The original prototype of this sytem was the first automation platform in the United States in 1992 at the University of Nebraska. LAB-Interlink integrates desired sample preparation components with sample transport hardware and system control software. Bayer Corp. (Pittsburgh, PA) is a newcomer in the automation marketplace with its recent debut of the ADVIA Systems in August 1998. The new ADVIA Labcell is an automated track system designed for configuration as a workcell or as a totally automated system with components added over time as needed. It connects to Bayer’s new ADVIA 1650 Chemistry Analyzer. Labotix (Peterborough, Ontario) and Automed (Richmond, British Columbia) round out the notable companies of automation products in North America. They manufacture automated sample processing modules for stand-alone work cells or integrated for total automation. Until recently their installations were only in Canada. Labotix now has two systems in the Unites States: Portland, OR, and Birmingham, AL. IVD COMPANIES AND THEIR SYSTEMS As indicated above, the number of IVD companies has been shrinking. Mergers, acquisitions, alliances, partnerships, and

collaborations are the order of the day for companies trying to strategically reposition themselves in this very competitive instrument and automation marketplace. The most recent acquisition late last year was the purchase of Chiron Diagnostics by the Bayer Group (Pittsburgh, PA) making them one of the largest diagnostic companies in the world. Abbott Laboratories aligned with TECAN to enter the frontend automation market. DADE International (Deerfield, IL) merged with Behring Diagnostics (Marburg, Germany) in late 1997. This merger followed the 1995 buyout of Syva Co. by Behring and the 1996 acquisition of DuPont’s Diagnostics Division by DADE (I31). In late 1997, a nontraditional merger occurred with a predominantly chemistry instrument company and a mostly hematology instrument company, as Beckman and Coulter joined forces to form a single companysBeckman Coulter, Inc. Johnson & Johnson Clinical Diagnostics consolidated their operations with another J & J company, Ortho Diagnostic Systems, in 1998 into a single diagnostics franchise called Ortho Clinical Diagnostics, Inc. The recent merger of Roche Diagnostics with Boehringer Mannheim Corp. rounds out the remaining “big name players” in the IVD industry. According to 1998 CAP survey data, the five most popular chemistry analyzers continue to be the Synchron analyzers (Beckman Coulter), Dimension (DADE Behring), Paramax (DADE Behring), which is no longer manufactured or marketed, Vitros (Ortho-Clinico Diagnostics), and Hitachi Systems (Roche/BMC) (I32). The latest models/features from these companies and new systems/models from other companies are reviewed here (I11, I33-I35). The Synchron ALX is the newest analyzer of the Synchron CX series by Beckman Coulter. The ALX, which replaces the CX7 Delta, includes features that improve operations, reduce maintenance, and make automation connectability possible. It analyzes nine critical care chemistries in 52 s and has 100 barcoded reagents onboard the analyzer. Their newest clinical chemistry system is the LX20 analyzer, designed for the high-volume laboratory. It offers clot detection, serum indexes, drug sampling from pediatric tubes, multiple wavelength blanking, and fiber-optic technology. It has 41 methods onboard the analyzer and a throughput of 1440 tests/h. A modular work cell (LX4201) is formed by linking two LX20s and the company’s Data-Link software, thereby processing 2880 tests/h. DADE Behring’s Dimension and Paramax analyzers continue to be popular even though the Paramax was discontinued from production in late 1996. The dimension RxL was introduced in 1997 and was the industry’s first high-volume general chemistry analyzer to combine sensitive immunoassays (including heterogeneous) with routine chemistries. It performs 167 immunoassays/h independently from routine testing to maintain high throughput with over 60 methods available and a large onboard capacity that includes 10 user-defined channels. A patented cuvette system produces and seals cuvettes to minimize exposure of biohazardous materials. With its Data Fusion Systems Integrator software, up to seven Dimensions and/or ACA analyzers can be linked through a single LIS port. DADE Behring also still markets the ACA Star Analyzer. With advanced electronics and highperformance software, throughput and turnaround time have been improved. It offers the largest onboard test menu with over 90 Analytical Chemistry, Vol. 71, No. 12, June 15, 1999

353R

chemistry tests. With an ACA Plus accessory, the Star can now perform sensitive immunoassays. The Vitros 950 by Ortho-Clinical Diagnostics offers a throughput of up to 900 tests/h, ability to handle multiple reagent lots onboard, larger slide capacity, and zero maintenance electrolytes. Its expanded test menu now includes immunoassay capability. The Vitros 250 is two-thirds the size of the 950 and offers the same 40+ test menu, onboard dilution, and immunoassay testing. With Ortho’s new LAB-Frame SELECT automation platform, two to eight chemistry instruments (950AT and 25 AT), immunoassay analyzers (Vitros ECi AT), and coagulation instruments can be integrated into one modular cluster for remote operation. The Hitachi analyzers by Roche Diagnostics/BMC complete the top five systems in U.S. laboratories by leading IVD companies. The Hitachi 747-200 is for large laboratories with high-volume testing, offering throughput up to 5700 tests/h. The Hitachi 917 offers a 98-test menu with throughput up to 1200 tests/h and continuous sample loading. For small- to medium-size laboratories, the 912 and 902 systems offer throughput of 720 and 300 tests/h, respectively. The newer 902 can also function as a dedicated specialty analyzer for esoteric chemistries. For larger laboratories, up to six modules of Hitachi 917 and 747 and Elecsys (immunoassay) Systems can be joined to offer a throughput of more than 10 000 tests/h and up to 155 reagent channels. When the preanalytical CLAS platform is added, sample racks are moved between modules, and processing lanes within each module leave the track free to expedite samples to the first available analyzer. The Roche/BMC Cobas Integra 7000 was introduced in 1998 and offers fluorescence polarization, absorbance, turbidimetric, and ISE technologies. It has an onboard capacity of 72 tests and a throughput of 860 tests/h. Abbott launched the Alcyon and Aeroset analyzers at the 1998 AACC Exposition. The Alcyon 300 and 300i (integrated ISE module included) were developed by the French manufacturer, Alcyon, which was acquired by Abbott in early 1997. The Alcyon is a benchtop analyzer for small- to medium-size laboratories performing fewer than 1000 tests/day. The 300i model provides a throughput of 300 photometric and 450 ISE tests per hour with an automated cuvette handler for loading and unloading of cuvettes. The Aeroset chemistry analyzer was developed by Abbott and Toshiba Corp. (Tokyo, Japan) for larger laboratories, processing 2000 tests/h from 56 onboard reagent positions and 3 ISEs. It has paired sets of sample probes, reagent probes, and cuvettes. By late 1999, Abbott is expected to debut the Architect C8000, a partner chemistry system to the current architect i2000 immunoassay analyzer, to form a single integrated workstation capable of up to 1400 tests/h (1200 and 200, respectively). Bayer Corp. developed and first marketed the ADVIA 1650 Chemistry Analyzer in 1998 in collaboration with JEOL, Ltd. (Akishima, Japan). It has a throughput of 1200 photometric and 450 ISE tests per hour, requiring only 2 µL of sample per test and almost as small reagent volumes. It can also be connected to the ADVIA Lab Cell, Bayer’s new automated track system. Olympus America, Inc. (Melville, NY) introduced the AU400 Chemistry-Immunoassay System in 1998 for smaller laboratories with lower testing volumes, complementing its existing higher throughput AU600 and AU1000 analyzers. The AU400 has a userdefinable test menu from over 115 methods of general chemistries 354R

Analytical Chemistry, Vol. 71, No. 12, June 15, 1999

and immunoassays with a throughput up to 800 tests/h. Abbott, DADE-Behring, and Olympus are currently the only IVD companies that do not market automation systems (modular or TLA) to link their product lines of chemistry analyzers, although Abbott Laboratories will soon, due to their recent alliance with TECAN Corp. CAP Today recently published an extensive survey of individual chemistry analyzers marketed by the major IVD companies in the United States (I36, I37). These vendors responded to a CAP Today questionnaire to provide data about their instruments. Responses to over 80 questions covering very detailed specifications and features are tabulated for each analyzer. Part I (I36) was devoted to analyzers used primarily for the hospital laboratory market and included data for 17 individual instruments by 8 vendors. Part II (I37) was aimed at those analyzers for the commercial laboratory/large hospital market, with five companies responding on six analyzers. It was recognized that categorizing these analyzers into a primary market was in some cases arbitrary, as many analyzers serve both markets well and overlap is common. Articles on tips and strategies toward selection of analyzers by laboratories accompany the tables in both parts of these instrument surveys (I38-I40). FUTURE TRENDS During this review period, more and more laboratories are integrating front-end automation systems with their arsenals of analyzers. What comes next after automating sample preparation? They are creating customized systems by hooking up their specimen-processing modules with analytical work cells for chemistry, immunology, and hematology, either as stand-alone units or integrated into TLA systems (I13). IVD companies and laboratory automation manufacturers are achieving an unprecedented level of cooperation with their mergers and alliances to ensure compatible work cells, since most laboratories are opting for a stepwise, modular approach to automation. By the year 2003, North America is expected to have 150 totally automated laboratories, while Japan and Europe will have 195 and 24, respectively (I13). This reflects an increase of ∼15-fold for North America from 1997 numbers. Modular automation, however, will grow exponentially and involve thousands of laboratories of all sizes. In the future, most common tests will be performed near the patient using micro- and nanofabricated devices with userselectable test clusters (I12). Widespread use of this technology favors a strategy of localized sites of diagnostic testing using principles of modular automation. Following the rising popularity of modular and total laboratory automation, the next revolution in laboratory science will be miniaturized analytical devices using microchips the size of a postage stamp or smaller. Microchips can move fluids and cells through microscopic channels and perform assays that previously would monopolize a benchtop or even floor model analyzer (I41-I43). Startup companies working on microchips have already developed prototype instruments. They are finding business partners such as major IVD companies who have an interest in microminiaturized systems, to get the devices ready for FDA approval and to market them. The demand for these miniature devices is being driven in part by the push for more pervasive bedside testing. Current obstacles to performing assays on a microchip include fluid movement control,

simplified test protocols for a clinical setting, and development of a platform or housing for the miniature device. It would be remiss not to offer a last-minute reminder of the Year 2000 problem that may or may not be looming for clinical laboratories. Computer glitches could be lurking in packaged software, application and program source codes, data files, screens, operating systems, and BIOS chips, electronic data interchange (EDI) interfaces, devices and instrumentation, suppliers, vendors, partners, workers, customers, and patients (I44) Any device that contains embedded chips or a date and time may be vulnerable and needs to be examined. Finally, the incorporation of artificial intelligence into analytical systems using expert systems and neural networks has not evolved at the pace previously predicted (I45). Still advancing the technologies of robotics, digital processing of data, computerassisted diagnosis, and data integration with electronic records, artificial intelligence is emerging more gradually as part of postanalytical data management, having been eclipsed by the significant popularity of emerging preanalytical systems. Nonetheless, in the future the interface between automated materials handling and informatics will create a dynamic system knowledge (artificial intelligence) that can make changes and connections respond to emergencies and predict future needs (I46-I49). There are three additional references sited here that are “must read” handbooks for any laboratorian or administrator entering laboratory automation. They offer definitions and development of automation, directions of technology movement, considerations toward systems planning, and options for various automation phases (I50-I52). Larry E. Schoeff is the Medical Technology Program Director and Associate Professor in the Department of Pathology at the University of Utah’s School of Medicine. He also consults on education issues for ARUP, Inc., a national reference laboratory owned and operated by the University of Utah. He received his B.S. degree in medical technology and his M.S. degree in medical technology/education, both from Indiana University. He has been at the University of Utah for nine years and previously was an associate professor at the University of IllinoissChicago for fifteen years. His specific interest is clinical chemistry instrumentation.

(I9) (I10) (I11) (I12) (I13) (I14)

(I15) (I16) (I17) (I18)

(I19) (I20) (I21) (I22) (I23) (I24) (I25) (I26) (I27) (I28) (I29) (I30) (I31) (I32) (I33) (I34) (I35) (I36) (I37) (I38) (I39) (I40) (I41) (I42) (I43) (I44) (I45) (I46) (I47) (I48)

(I49)

LITERATURE CITED (I1) (I2) (I3) (I4) (I5) (I6) (I7) (I8)

Schoeff, L. Anal. Chem. 1997, 69(12), 200R-203R. Auxter, S. Clin. Lab. News 1998, 24(12), 4-5. Berenblum, E. Med. Lab. Observer 1998, (July), 60-64. Keitgez, P. Med. Lab. Observer 1997, (October), 54-56. Miller, S. Clin. Lab. News 1998, 24(1), 11-15. Burke, J. Clin. Lab. News 1998, 24(1), 16-19. Edwards, G. Clin. Lab. News 1998, 24(1), 20-24. Auxter, S. Clin. Lab. News 1998, 24(7), 46-47.

(I50) (I51) (I52)

Ringel, M. Med. Lab. Observer 1997, (January), 44-70. Guterl, G. Adv. Med. Lab. Prof. 1998, (May 11), 28-29. Sasavage, N. Clin. Lab. News 1998, 24(10), 18-19. Felder, R., Kost, G. Med. Lab. Observer 1998, (April), 23-27. Boyce, N. Clin. Lab. News 1997, (3), 9-10. Felder, R. Automation of Preanalytical Processing and Mobile Robotics. In Handbook of Clinical Automation, Robotics and Optimization; Kost, G., Ed.; Wiley and Sons: 1996; pp 242282. Felder, R. J. Int. Fed. Clin. Chem. 1997, (9), 56-60. Luczyk, K. Med. Lab. Observer 1997, (March), 42-44. Kost, G., Felder, R. Med. Lab. Observer 1998, (June), 46-56. Middleton, S.; Mountain, P. Process Control and Online Optimization. In Handbook of Clinical Automation, Robotics and Optimization; Kost, G., Ed.; Wiley and Sons: New York, 1996; pp 515-540. Felder, R. Overview and Challenges. In Handbook of Clinical Automation, Robotics and Optimiation; Kost, G., Ed.; Wiley and Sons: New York, 1996; pp 3-29. Guterl, G. Adv. Med. Lab. Prof. 1998, (May 11), 30-31. Wild, M. Adv. Med. Lab. Prof. 1997, (May 19), 6-7. Scypinski, S.; et al. Lab. Rob. Autom. 1997, 9(5), 229-236. Tamilarasan, R.; et al. Am. Lab. 1997, 29(20), 22-26. Farnsworth, R. Lab. Rob. Autom. 1997, 9(3), 129-134. Hamilton, S.; et al. Lab. Rob. Autom. 1996, 8(5), 287-294. Reichman, M.; et al. Lab. Rob. Autom. 1996, 8(5), 267-276. Clin. Lab. News 1997, 23(3), 11-12. Clin. Lab. News 1998, 24(10), p 28. Customer Bulletin 9095, for Power Processor; Beckman Coulter, Inc.: Brea, CA, 1998. Clin. Lab. News 1998, 24(7), 50-56. Clin. Lab. News 1998, (August 4), 1-2. CAP Surveys; Participant Summary: Chemistry (general) Survey Set. College of American Pathologists, Northfield, IL, 1998. 1998-1999 Clinical Laboratory Reference; Medical Economics Co., Montvale, NJ, 1998; pp 61-115. Neelkantan, N.; et al. J. Autom. Chem. 1997, 19(1), 9-13. Blic, A.; et al. Clin. Lab. 1997, 43(12), 1125-1131. CAP Today 1998, 12(6), 41-54. CAP Today 1998, 12(7), 55-60. Paxton, A. CAP Today 1998, 12(6), 34-40. Aller, R. CAP Today 1998, 12(6), 42. Aller, R. CAP Today 1998, 12(7), 54. Sainato, D. Clin. Lab. News 1998, 24(6), 22-23. Kricka, L. Clin. Chem. 1998, 44(9), 2008-2014. Pfeiffer, D. EXS 1997, 81, 149-160. Lafler, K., Lafler, D. Clin. Lab. News 1998, 24(7), 45. Boyce, N. Clin. Lab. News 1997, 23(1), 2-3. Kost, G. Artificial Intelligence and New Knowledge Structures. In Handbook of Clinical Automation, Robotics and Optimization; Kost, G., Ed.; Wiley and Sons: New York, 1996; pp 149-193. Lifshitz, M., DeCresce, R. Strategic Planning for Automation. In Handbook of Clinical Automation, Robotics and Optimization; Kost, G., Ed.; Wiley and Sons: New York, 1996; pp 471-496. Kost, G. Point of Care Testing f The Hybrid Laboratory f Knowledge Optimization. In Handbook of Clinical Automation, Robotics and Optimization; Kost, G., Ed.; Wiley and Sons: New York, 1996; pp 757-838. Sasaki, M.; et al. Automated Clinical Laboratory Systems. In Handbook of Clinical Automation, Robotics and Optimization; Kost, G., Ed.; Wiley and Sons: New York, 1996; pp 442-467. Kost, G., Ed. Handbook of Clinical Automation, Robotics and Optimization; Wiley and Sons: New York, 1996. Bissel, M., Petersen, J. Automated Integration of Clinical Laboratories: A Reference. AACC: Washington, DC, 1998. Liscouski, J. Laboratory Automation. In Handbook of Instrumentation and Techniques for Analytical Chemistry; Settle, F., Ed.; Prentice Hall: Upper Saddle River, NJ, 1997; pp 101-118.

A1999909J

Analytical Chemistry, Vol. 71, No. 12, June 15, 1999

355R