4-1.Construction of Laboratory Automation System by A&T

Therefore, when compared to the field independent system, the branched integration system is clearly more advanced. However, when the integration of the fields of analysis and analyzers is not completed, the degree of improvement is lower.

Separate component units
There are many LAS methods, and it is difficult even to explain everything about a single LAS. As a result, our concept of common units that make up LAS/STS will be discussed individually.

Start stocker
Due to the differences in the function of this unit among various manufacturers, this unit has several different names (start unit, start yard, etc.). The basic function of this unit is to direct arranged specimens to the subsequent line, but in fact, many functions are associated with this unit. The following are possible functions.

1. Reading specimen ID (bar code) and eliminating specimens with illegible BC
2. Reading rack ID (bar code) and eliminating racks with illegible BC
3. Prioritizing STAT specimens
4. Prioritizing the order in which racks are processed
5. Sorting haphazardly arranged racks to maximize subsequentanalysis efficiency
6. Controlling racks by analyzing the state of subsequent connecting analyzers
(standby)

A&T's CLINILOG is capable of performing the above functions, except for 5. It was determined that it would cost too much to carry out function 5 using a robot, so this problem was resolved by color coding caps and racks, utilizing trays containing racks that held specimens to be subjected to the same test (start row/multiple row), and processing racks while evaluating the priority of tests and the status of connecting analyzers (the unit was designed so that the number of rows of trays can be increased).
In addition, the ID/BC(bar code) of specimens is read only once by the start stocker. Subsequent connecting units (including analyzers) only read the ID/BC of racks, not of specimens. The information on specimens that are on a rack is ascertained from the control computer (this avoids problems associated with reading bar codes by the connecting units and analyzers).

Automatic centrifugation unit
This unit is most effective when a large number of specimens are batch processed. When specimens arrive sporadically (this will happen when specimens are promptly processed), it appears that under the current technology and costs this unit would not satisfy investment cost. To process 150 specimens per hour (fairly large laboratories), we recommend the use of four to five small manually operated centrifuges. If an automatic centrifuge is necessary for the measures of biohazardous materials, several parallel units must be arranged to handle specimens that arrive sporadically, thus unrealistically increasing implementation costs. Except for a centrifuge with an excellent automatic balancing mechanism (the same mechanism used in washing machines) manufactured by Panasonic, which is used in Toshiba STS, there are no centrifuges that can sufficiently handle sporadically-arriving specimens. (By the use of vacuum blood collection tubes, the amounts of blood in them become to be very various. And the automatic balancing become to be difficult.)
With A&T's CLINILOG, we determined that since some users will require this unit (handling a large number of specimens and/or approaching to the measures of biohazard ), we made it possible to bypass the present unit by inserting specimens (racks) that are processed by other centrifugal machines.

De-cap unit
The implementation cost of this unit is justified. There are two types of this unit: film (TERUMO only) and rubber cap. Problems associated with this unit are irregularities in the adhesive strength of films and the hardness of rubber (some are too soft).
This unit performs the following functions.

1. Opens seals as appropriate, and when improperly opened, the unit is designed to detect such mistakes (problems occur when tubes are sent to the next unit unopened).
2. Identify and eliminate already-opened specimens (in order to prevent errors).

Aliquot unit(online/offline)
A laboratory system will be very effective if the implementation of an automatic aliquot system and the integration of analyzers can be achieved simultaneously. Many think the implementation of a stand-alone aliquot system would eliminate the need for STS. This is another reason that justifies the implementation of the present unit.

While considering the rationality of the aliquot procedure, we came to the conclusion that a system should be designed so that each specimen is divided a minimal number (2 to 4) of times (integration of analyzers). If about 150 specimens need to be processed each hour, then only one aliquot machine will be needed. In other words, there is not much advantage in implementing separate online and offline aliquot machines. If the objective is to improve the test speed of sporadic processing, this degree (150 specimens/hour) of processing capacity is more than sufficient for relatively large hospitals having more than 500 beds.
The advantages of having only one aliquot machine are as follows.

1. Lower specimen volume (dead volume of each daughter specimen is reduced)
2. Use less pipette tips (only one tip is needed for each specimen)
3. Lower investment costs
4. Requires less installation space

In A&T'S CLINILOG, a design in which this aliquot unit is concentrated in one machine was chosen for the standard package (it is possible to connect a separate offline aliquot unit).

The following functions are required for the STS aliquot unit.

1. Lower dead volume
2. Detect insufficient specimen volume
3. Detect aliquot mistakes (such as fibrin lumps)

Secondary bar code pasting unit
New bar code labels (secondary bar code) are issued and pasted to daughter specimens to be used for identification. There are several advantages and drawbacks associated with this method.

Advantages

• When bar code readers (BCR) are placed properly within STS,
  each specimen can be identified at each location.
  Compared to a system (e.g., A&T's CLINILOG) in which the ID of racks containing daughter specimens (each rack is labeled with a bar code) are read, specimen management has been simplified.
  If specimens are managed in racks, then specimens in racks can be exchanged on the line.

Drawbacks

1. High costs
2. Daughter specimen containers must be large enough to accommodate labels.
   - In some cases, to perform analyses using micro amounts of specimens, a plastic sample cup must be placed in a larger container so that a label can be attached.
3. The bar code of each specimen is read several times, and since the characteristics (sensitivity) of each BCR are different, the probability of reading mistakes occurring is higher.

It was determined that daughter specimens would be managed by rack IDs in A&T's CLINILOG instead of utilizing secondary bar codes. Since the order of daughter specimens in racks will not be changed, the only disadvantage is more complicated software that manages online specimens.
The secondary bar codes for specimens following plasma separation (daughter specimens for storage) will be discussed in a section on post-processing.

Analyzer connecting unit
Analyzers and STS can be connected by one of three methods

1. A sampling probe of analyzers collects specimens on STS.
2. An entire rack is incorporated into analyzers.
3. Specimens are removed from racks and incorporated into analyzers.

The biggest advantage of option 1 is the ease of the standardization of the connection specification between STS and analyzers. If the distance between analyzers and specimens on STS, as well as the height of specimen collection are standardized, then standardized STS connecting units that can incorporate analyzers made by various manufacturers can be used (the same analyzer standard applies to rack type and single-tube type STS). Analyzers that are connected to this type of unit are required to achieve the following: (1)low carry-over for specimen sampling probes, and (2)short specimen sampling time to ensure that specimens are processed promptly by subsequent connecting analyzers.
With option 2, the same rack must be used for STS and analyzers (as long as industry-wide standardization can not be adopted, there will be many restrictions associated with this option). In addition, most current analyzers that incorporate racks can not, or require a considerable amount of time to, return specimens to STS. As a result, when this option is selected, most systems designate special daughter specimens for each analyzers (thus increasing specimen aliquot and dead volume).
Option 3 is a cap-piercing version of option 1. Hematological analyzers can be connected by this type of option. The ability to connect STS and analyzers using the same connecting unit used for option 1 is a major advantage.

In A&T's CLINILOG, options 1 and 3 are employed in addition to the following functions: (1)bypass function, (2)intra-rack specimen jump function, (3)designated rack count buffer function, and (4)retest waiting loop function following primary sampling, thus improving both STS and LAS. In some laboratories, even though STS and analyzers are connected, they do not contribute to the improvement of LAS. It is necessary to consider the importance of the above-mentioned functions. They are necessary to process high priority specimens promptly, maintain the maximum processing capacity of connected analyzers, and perform retests in a timely manner without increasing blood collection volume. One well-known manufacturer introduced a new internal transportation system of the new analyzer with a new rack drawing method capable of processing urgent tests. However, products by A&T and other companies have long offered this function.
Disadvantages associated with certain poorly constructed STS are: urgent and regular specimens are processed in the same fashion, the actual processing capacity of analyzers is reduced to about half the level indicated by their original specifications, and the blood collection volume is significantly increased to achieve automatic retests.

Terminal stocker

This unit receives racks (specimens) at the end of STS, and as with start stockers, this unit has several names. The basic function of this unit is the organization and storage of racks (specimens), but it must also perform the following functions when the unit is linked to the STS control computer.

1. Reading rack ID/BCs, and confirming their arrival
2. Sorting arriving racks (specimens)
3. Index management of sorted racks (specimens)

If racks are simply received in the order in which they arrive, then the system will have trouble identifying specimens that are needed for retesting.

Retest loop-back line
To retest specimens automatically, a loop-back line is generally constructed to return analyzed specimens (racks) to analyzers. Nonetheless, we at A&T believe that too many problems are associated with the construction of this type of line, and that at present, the implementation of a loop-back line is not cost effective.
There are currently five types of automatic retest systems.

1. Analyzers receive daughter specimens from STS, and retain the specimens after analyses so that retests can be performed as necessary (actualized by the design of analyzers).
2. Analyzers take a certain amount of specimen from each tube, and retain the specimens after analyses so that retests can be performed as necessary (actualized by the design of analyzers).
3. Analyzers take a certain amount of specimen from each tube, and retain the specimens after analyses so that retest can be performed with an instruction from higher level LIS (actualized by the design of analyzers).
4. Already analyzed racks (specimens) are retained immediately after use by each analyzer until the results of analyses or instructions from LIS are processed, and necessary racks (specimens) are returned to analyzers (actualized by the design of STS).
5. Racks (specimens) are sent to the stocking unit (terminal stocker), and as necessary, they are returned to the top of the STS for retest (actualized by the design of STS).

Hitachi 7350/7450/7600 analyzers employ option 1, which has two drawbacks:
data checks and retests are performed based on the results of tests performed by each analyzer, and the dead volume of divided specimens for retest is increased.
JEOL Bio-Majesty, a well known example of option 2, does not increase the dead volume of specimens since specimens are diluted and retained after sampling.
Option 3 is achieved with A&T's 502X. As is the case with JEOL's Bio-Majesty, 502X retains specimens by diluting them after sampling, so the dead volume of specimens is not increased, and after sending test results to higher LIS, 502X can wait for instructions from LIS for a certain length of time. Because of this function, retests can be performed based on the results of various analyzers, independent from STS.
Option 4 can be achieved by connecting Toshiba TBA-80FR/200FR or IRC(International Reagents Co.,Ltd.) COAGREX-700 (both lacking a retest function when they connected with STS) to A&T's STS/CLINILOG. As is the case with option 3, after sending test results to higher LIS, specimens can be retained for a certain length of time until receiving instructions from LIS. Nonetheless, it is necessary to purchase a rack buffer unit for each analyzer connecting unit.
Option 5 is frequently utilized by the STSs of other manufacturers. Even though retests can be performed based on the results of all analyzers in a system, the following problems have been identified.

• Retests are time consuming
  Compared to options 1 through 4, it takes longer to obtain the results of retests from the time the results of primary tests are generated.
• Greater specimen volumes are necessary.
  In general, parent specimens are looped back and divided into daughter specimens, and retests are performed using these daughter specimens. However, this naturally increases dead volume, thus making it necessary to collect additional amounts of blood.
• The implementation costs and required space for STS increase
  Inevitably, this option is the most expensive method, and requires more installation space.
  Therefore, the implementation costs of STSs fluctuate greatly depending on how analyzers perform retests.

Post-processing

• Blood clot separation and aliquot unit
• Bar code pasting unit for stored specimens
• Re-cap unit
• Refrigeration (index) stocker

These units are nice to have, but their costs are questionable. For example, it does not take more than two or three workers to perform post-processing in fairly large hospitals. Upward of \40 to \50 million can be saved by reducing the workforce by one, but if all of the above units are installed, it will double or triple costs. Nonetheless, post-processing technology will continue to advance, and the cost of an integrated post-processing unit will be reduced to \20 to \30 million in the near future. Until the cost of such a system comes down, A&T can not recommend the implementation of the above units.
In some cases, however, investment effects should not be the main concern, such as when measuring with biohazardous materials or having clear objectives other than saving manpower. We will be more than happy to assist in the planning and construction of such a laboratory, but we have never assisted in constructing a system incorporating the above post-processing units based on clear objectives (most systems are only concerned with automation, and they neglect to take into account the investment effects and responsibilities).

Analyzers generalized biochemical and immuno- turbidimetric analyzers
From the viewpoint of STS manufacturers, current analyzers that are connected to STSs are not designed considering STSs or LASs in mind. Naturally, depending on how analyzers are used (alone or connected to a STS to be incorporated in a LAS), their required specifications and functions are different. As a result, analyzers that have been designed to be used independently are not suitable for use in STS/LASs. Nonetheless, future analyzers will be designed to accommodate STSs and LASs. A&T, as an STS manufacturer, feel that future analyzers should be designed in the following manner.

Analyzers that can process multiple tests in various fields

The higher the number of analyzers connected, the higher the costs of a STS and the larger the LAS. More importantly, the higher the number of analyzers, the higher the dead volume of specimens (however, a higher number of same analyzers for mutually backup and aliquoted workload does not mean this problematical situation). If each analyzer can perform multiple tests in various fields, then the size of STSs can be reduced, the volume of blood collection can be reduced, and the entire operation can be run more efficiently.
Ironically, since the development of analyzers is intended to perform biochemical, hematological, serological and general analyses and coagulation and fibrinolysis, many believe that there will be no need for currently available large-scale STSs in the future (nevertheless, such integrated analyzers will contain an internal transportation system).

Analyzers that can run 24 hours a day and handle

sporadically arriving specimens
Opinions opposing the development of prompt, pretreatment and 24-hour tests are growing less common, and analyzers will surely be developed in this direction in the near future. Therefore, future analyzers should be able to handle specimens 24 hours a day and process sporadically arriving specimens. Existing instruments (e.g., Super Multi) that are designed to process a large number of specimens at once will hinder the development of such analyzers.

External sampling analyzers

Analyzers that can quickly catch and release specimens and retain them as necessary.
At present, analyzers use one of the following methods to catch specimens from STS.

1. A method in which analyzers directly draw specimens placed on a STS (external sampling method).
2. A method in which tubes containing specimens in racks placed on a STS are incorporated into analyzers (specimen incorporating internal sampling method).
3. A method in which racks containing specimens are incorporated from a STS into analyzers (rack incorporating internal sampling method).

The following matters should be considered when implementing one of the above methods.
(1) Degree of freedom in connecting STS and analyzers
With method 3, STSs and analyzers that use the same type of rack can be connected. In other words, since only certain racks can be used, the number of STS/analyzer combinations is limited. Therefore, many more STS and analyzers are available with methods 1 and 2.
Of course, this problem can be resolved if racks can be standardized among manufacturers.
(2) Degree of freedom in step construction or replacement (1)The same as the above. With method 3, since only certain racks can be used, the number of STS/analyzer combinations is limited.
(3) Number of daughter specimens and dead volume
At present, most analyzers that are designed to be used in method 3 incorporate racks from STSs but do not return them (in other words, specimens on the racks can not be used by subsequent analyzers). As a result, specially designated daughter specimens are necessary, and resulting dead volume must be discarded. Even when retests are not performed, about 350 µl of dead volume is necessary for a currently available analyzer (Super-Multi). If each of these analyzers request retests, then more blood must be collected from patients to prepare the necessary daughter specimens. This problem can not be resolved by standardizing racks.
Analyzers employing method 3 that can handle standardized racks will appear in the future, but will they be able to return racks to STSs? Even if they can return racks, will they be able to achieve the quick "catch and release" that is possible with the external sampling method? The answer is likely "No," so specially designated specimens will very like be required.
(4) Effect on automated retest system construction
The connection method of analyzers and STSs can easily affect how automated retest systems are constructed as well as the final function of that system. As was mentioned in the section on the retest loop-back line, it is difficult to implement an automated retest system via STS design alone. Needless to say, it is ideal for STSs and analyzers to be easily connectable and capable of performing many functions.
Therefore, method 1 is the most appropriate method (method 2 for whole blood cap piercing). Commercial laboratories performing batch processing are the only places where method 3 (rack incorporating internal sampling method) is useful.
If the following two points can be actualized, the external sampling method will certainly be superior to other methods.

(1) Quick catch and release
Even when an external sampling analyzer is used, if the specimen processing capacity of the analyzer is 30 seconds per specimen, then the processing capacity of subsequent connecting analyzers is also lowered to 30‚T seconds per specimen (120 specimens per hour). As a result, if an external sampling analyzer can not process each specimen in 15 seconds, then the analyzer must be placed at the end of STS to ensure that it will not affect the performance of other analyzers, or else daughter specimens must be designated for the analyzer. Furthermore, an automated immunological analyzer with a slow specimen processing capacity can not even be connected to the end of STS since inter-specimen carry-over must be avoided at all cost.

(2) One-dip sampling
To realize the above improvements (quick catch and release, and minimal inter-specimen carry-over), sample probes are repeatedly dipped into each specimen, a process that should be avoided whenever possible. Analyzers that release specimens in a single sampling are desirable.
Analyzers that possess a retest function and output final test
results
As described before, there are many ways to construct an automated retest system and there are disadvantages associated with each system. By taking into account such issues as real time output, specimen dead volume and construction cost, it would be ideal for analyzers to be able to perform retests and deliver final test results.
Analyzers that use micro amounts of specimens and reagents
It is meaningless to implement a LAS if collection volume and analysis costs increase with the installation of STS. In particular, to utilize an analyzer as one of the units connected to a STS, it is important to remember that dead volume and other issues must be analyzed as a part of an integrated system, not as an independent unit. Therefore, a unit that can perform one-dip sampling on specimens processed by JEOL's Bio-Majesty or A&T's 502X, and then dilute them for storage would be valuable, and such a unit would be even more useful if the amount of specimen for analyses and retests can be decreased.

Into the age of super single-line multi analyzers
Due to various reasons including the characteristics of STS, the age of super multi analyzers is finished and the age of super single-line multi analyzers is being implemented in many laboratories, except for commercial laboratories and ultra-large hospital laboratories. It seems that new analyzers are also designed to adapt to this change. In this "Super Single-Multi Age," the specimen processing capacity, multifunctionality and new functions (STS connection or retest) of existing single-line multi analyzers are being enhanced. Users who are oblivious to this change and continue to only concentrate on only batch processing will not be able to meet ever diversifying future demand.

Picture 2. A&T/502X

Transportation method (rack and single-line)
Two points should be considered when a manufacturer is deciding between single-tube transportation or rack transportation systems: Which system is better suited to current technology? Which system will be more advantageous in the future of the current technology? If an ideal system were pursued without considering currently available technology (a system that can be constructed with the current technology and reasonable costs), then the single-tube transportation method (plus a loop system) should be chosen. Nonetheless, if the single-tube transportation method were to be employed with current technology, there will be many problems with restrictions in the transportation efficiency and speed. As a result, it will take a great deal of money to resolve this problem, and if not resolved, the actual processing capacity of analyzers is greatly compromised. Since there were no A&T's analyzers that would cause difficulties with our selection of the rack transportation method, and the rack method was more effective from the viewpoints of cost and function, the rack method was incorporated in CLINILOG.
The branched integration system is the most advanced system as far as the transportation line configuration is concerned. However, we believe that this system is not without its problems; as long as a transportation line is branched, specimens must also be divided for each line. Therefore, if one specimen is promptly processed by analyzers on a single-line one after another, there is no need to divide specimen, thus greatly improving the issue of dead volume.

As was mentioned in the section on desirable analyzers from the viewpoint of STS, if analyzers can perform the following functions satisfactorily, there is no need to make branches on a transportation line.

1. Handle sporadically arriving specimens 24 hours a day.
2. Perform multiple tests in a variety of fields.
3. Employ an external sampling method.
4. Promptly catch and release specimens.
5. Minimize (eliminate) inter-specimen carry-over.
6. Retain specimens for retests.

If these functions can be achieved, analyzers should be arranged on a single-line. CLINILOG-package#1 arranges biochemical, hematological, immuno-turbidimetric, and coagulation /fibrinolysis analyzers on a single-line.

Connection between LIS and transportation control (action and information)
There is a movement to define TLA as "automation of the entire laboratory including LIS" and LAS as "analyzers + transportation system + transportation control computer system," so their relationship can be expressed as "TLA?aLAS+LIS." There are, however, still doubts; if LAS is not designed with LIS as its central component, then the system will have technical and economical shortcomings. In a laboratory, specimens are resources of continuously changing data, and depending on change, different physical distribution management methods are required (it's just logistics). It will be difficult to separate this type of physical distribution management from information management. In actual practice, there are many problems associated with the connection of LIS and STS systems that can not be resolved by simply standardizing communication between the two systems: specimen reception management, aliquot management and its relationship to work-sheets, retest instruction and processing, and quality control. In other words, a successful LAS is not merely the result of connecting LIS and STS (since the bid specification of most systems are not thorough, there is a lot of interference after starting construction). Since LASs are often comprised analyzers, STSs and LISs made by different manufacturers, users must have a clear plan that compensates for any differences.
At present, the information and control system of LIS, STS and analyzers are connected in one of the following three ways.

Integration
As mentioned numerous times in this article, the integration of laboratory fields and routine and urgent tests must be considered first in LAS implementation, and the integration of analyzers should be evaluated next. Integration of laboratory fields, integration of operation, integration of analyzers, integration of space, integration of people; "Integration" is truly the keyword for successful LAS implementation. If the resources, labor and space that is saved by integration can be used to construct a laboratory that plays an important role in medicine, then hospital laboratories will continue to exist and function.

Capital investment for private enterprises
Before investing in equipment, each plant will attempt to maximize operation by evening out the workload and reducing investment costs. In other words, if the workload can be aliquoted over time, then a system should be designed based on this even workload, thus avoiding excess investment.
In many laboratories (in Japan), LAS should complete a day's work in only three hours (including one hour of warm-up time). Specimens collected from hospitalized patients can be processed in the early morning, and none-urgent tests can be processed later. In one hospital laboratory that implemented CLINILOG, specimens that are collected from hospitalized patients are processed before 9:00 am. Therefore, the scale of both the STS and analyzers is smaller. Also, whether the labor that is conserved by laboratory integration can be used to improve laboratory services should be analyzed further.

Future tasks
A&T is not only concerned with reducing labor. Simply reducing costs by cutting their workforce will not enable hospital laboratories to compete with commercial laboratories. For example, in one commercial laboratory, six Hitachi 7450 and two Hitachi 7170 are operated by one full-time employee and one part-time employee, while another full-time employee operates two single-multi analyzers and five chemiluminescence analyzers. It is meaningless to attempt to compete with such commercial laboratories with regards to test costs. In order for hospital laboratories to survive while reducing costs, they must provide extra services such as pretreatment tests and urgent night-time tests, assisting in data analyses, and providing the latest laboratory information. This is precisely the reason why A&T is stressing the importance of saving resources (money, the workforce and space) by integration.
To achieve this goal, manufacturers must be able to provide inexpensive, highly productive systems. System packaging, adapting standard specifications, low costs and integration with LIS are issues that must be addressed by manufacturers to produce the next generation of LAS.