Printer Friendly

Robotics: boon or bust in the lab?

Robots are already performing intricate, useful work in clinical labs, but laboratorians and manufacturers can

develop systems that

will do even more.

There's a marvelous future for robotics in very limited, welldefined areas of the clinical laboratory.

That forecast-starting out full of bright hope, becoming somewhat restrained along the wayparallels the kind of passage many of us have gone through in this area of research and development. Our initial expectations were higher than the results we have settled for.

My enthusiasm, and much of it remains, was tempered by the difficulty of finding useful clinical laboratory problems for robots to handle. For example, many simple procedures have already been automated in the clinical chemistry laboratory. When we seek a work problem to tackle in clinical chemistry, we tend to pick something difficult. And just as it will be difficult for a technologist, so too will it be difficult for a robot.

Robots have been defined as computers capable of doing physical work. One largely manual clinical laboratory area where they can make a contribution is accessioning, from delivery of the specimen to the laboratory to presentation of the aliquoted specimen to an automated analyzer. This includes comparing labels on specimen tubes with a list of patients or tests, centrifuging the specimens, opening up the tubes, and pouring the serum or plasma specimens into cups designed for the analyzer.

Most laboratory errors occur during accessioning. The technician's paradox comes into play: Anyone intelligent enough to do the job is too bored to do it right. The job is also time-consuming and hazardous, subject to the spread of aerosols and breakage.

For these and other reasons, such as the nature of the tasks, accessioning is a good place to introduce robotics. Our clinical chemistry laboratory has taken some steps in that direction.

In a project that began two years ago, we developed a workstation built around a microcomputer-controlled Mitsubishi RM 501 robot arm (see photo, page 32). It was designed to accept a rack full or partially full of evacuated specimen tubes; lift the tubes out of the rack; open the centrifuge, load the tubes in, making sure the centrifuge was balanced; close the centrifuge; and run it for the appropriate period of time. Then the robot would go through the process in reverse, so that the rack would have the tubes back in the same order but centrifuged.

It's quite a problem, from a robot's point of view, to find and pick up tubes that vary in size and sit at different angles in a rack. To overcome that problem, we built a special gripper for the end of the robot arm and a rack holder that stabilized the tubes.

Through the action we designed for it, the gripper also avoided becoming entangled with test tube labels that have partly peeled away, exposing sticky surfaces. A technologist gets around that easily because of the beauty of the human hand; a robot can have a much harder time with it.

A weighing algorithm helped the robot quickly balance the tubes as they were loaded i to th centrifuge. This 'involve elaborate mathematics on the distances the tubes were placed front the center of the centrifuge arm.

The system worked well, although it had minor reliability problems-one serious error in 1,000. The main drawback was the robot's slowness. Our design aim was 160 specimens per hour, but to achieve that, the robot needed a second centrifuge to load during the 10-minute spin of the first centrifuge.

We decided not to go any further. Besides requiring a second centrifuge and extra peripherals for weighing and other functions, expansion of the system would have called for a fair amount of redesign and programming. The payback period-how long till savings from replacing part of one technologist's time matched development costs-would have been four years. And there was nothing proprietary about the system, so we could not market it to other laboratories.

We reviewed our experience with the project and decided that the most serious problem had been the robot's difficulty in dealing with a centrifuge. That led us to invent a single-tube centrifuge that makes loading much easier for the robot: All it has to do is address a single point in space, one hole.

The device is fast, processing a specimen in one to three minutes. With our present design, using four centrifuge modules, we can achieve a throughput of 116 specimens per hour.

These centrifuge modules are also smart. They can read bar code labels and examine serum for such characteristics as icterus, hemolysis, and turbidity as the tube is spun. The tube is centrifuged on its vertical axis and does not move in space, which facilitates examination by sensing devices. Information thus sensed can be transmitted to a computer with a resident expert system that checks its internal files and states whether or not the specimen is suitable for the tests ordered on it.

To us, this is an important step forward in the pre-processing phase of testing. Everywhere else in the laboratory, checks and balances have been introduced through technology to determine whether or not the analysis is accurate. Accessioning, on the other hand, is often staffed by the least experienced medical technologists, who are supposed to make critical decisions on the suitability of specimens.

Our system of centrifuge modules is completely enclosed. It enhances laboratory safety at a time of great concern over the threat of AIDS, hepatitis, and other health hazards. We hope to have a manufacturing prototype of the system by 1990.

Meanwhile, in another testing area, our chemistry laboratory has already reaped cost-saving dividends from robotics. This involves a system we bought from Zymark Corp., Hopkinton, Mass., four years ago and programmed to perform an estrogen receptor assay (see photograph, page 32).

The assay was one of those uncommon procedures in clinical chemistry well suited for robot handling. It was highly manual, consisting of many boring steps that did not require much in the way of intelligence but did demand a great deal of attention. The technologist had to handle radioactive material, and he or she had to do very precise pipetting of small volumes.

Today, a technologist spends a half-hour setting up the robot work table with materials for a complete run. The robot program knows where every item is on the table.

As in the past, the technologist also prepares the cytosol from tumor tissue. The cytosol is then presented to the robot, which adds the buffer and aliquots out the appropriate amounts, adds the radioactive hormone, adds cold hormone for competitive-binding purposes, does the mixing and the incubation, adds dextran-coated charcoal, centrifuges to take off the remaining fraction that has not been bound by the dextran-coated charcoal, aliquots that into vials, adds liquid scintillation fluid, screws caps onto the vials, and finally puts the vials into racks.

The technologist lifts the racks off the robot table and places them into the beta counter for counting. While the radioactive levels in the assay are relatively low, the technologist nevertheless benefits from avoiding some exposure in steps now handled by the robot.

When we acquired the robot, we had been performing most of the estrogen receptor assays for management of breast cancer cases in British Columbia. Technologist time freed up by the robot on these tests enabled us to add a progesterone receptor assay, which physicians had long sought. In two years, added revenue from the. new assay paid for the purchase, installation, and programming of the robot.

A laboratory planning to introduce a similar robot-run test today would not have to duplicate our major programming effort. Zymark has developed program modules to operate different robot peripherals-such as vortexers, cappers, pipets, and syringe pumps-and a user just puts the modules together. Technical representatives also provide engineering and applications assistance.

What Zymark and other companies in the robotics field cannot do is sell a ready-to-operate clinical laboratory testing system-at least not without obtaining approval from the Food and Drug Administration. The manufacturers have been reluctant so far to undergo the lengthy, time-consuming FDA approval process. They are instead committing the bulk of their resources to pharmaceutical and industrial analytical laboratories, where robots are well established.

Another problem inhibiting growth of robotics in the clinical laboratory field 'is the imprecise quality of disposables. The robot itself is fairly precise-it has a high degree of repeatability. But if you present it with a pipet tip that is bent or off-center or with a warped cap for a tube, the robot cannot do anything with it. Technologists easily make adjustments for such deformities with their hands. Robots lack that ability, so disposables will have to become more robot-friendly.

From a personnel standpoint, robots give interested technologists an opportunity to acquire programming expertise and skills that were characteristic of the autoanalyzer technology of the 1950s-skills in such areas as mechanics, hydraulics, and electronics. These skills withered in the 1970s when more advanced automated analyzers left laboratories largely reliant on service representatives.

Robotic systems tend to be modular (see Figure 1). You can take them apart and put them together in different ways to make them do new things. It's possible to design a robot layout that could perform one job in the morning, another job in the afternoon, and yet another job at night. Not everyone in the laboratory would need to know how to program robots to do all that, but the medical technologist who acquires those skills will become more and more valuable.

Interestingly, a study I conducted through questionnaires and interviews at a variety of robotequipped laboratories disclosed less fear of the changes that take place after robot implementation , than other investigators have found among blue-collar workers at automobile plants and in other manufacturing industries.' I probed four areas: 1) whether laboratory personnel fear they might one day lose their jobs to a robot, 2) how much lab personnel know about robots, 3) management concern for employees during robot implementation, and 4) expected job changes.

In three of those areas, my study uncovered a significantly more receptive attitude toward robots than others had reported for industrial settings. The only poor score came in the area of management concern for employees (the score was poor overall, but there were big differences between laboratories). This probably reflected a poor presentation of planned robotics to the laboratory staff and a lack of consultation with employees on how best to implement the changes.

Contributing to the problem was the increased stature of the technologist most directly responsible for the robot. Suddenly, this person became a center of interest, spent more time with the boss, had new responsibilities, and was engaged in interesting activity unavailable to others. Compounding the resentment to all this was the work that the rest of the staff had to pick up because the one technologist had to spend so much time with the robot.

I have indicated some of the obstacles that lie in the way of robot introduction in clinical laboratories. Widespread clinical laboratory use will come when manufacturers of robot systems have met the demands of the market in pharmaceutical and industrial analysis labs and begin to look further afield.

We will see robots in clinical laboratory areas that are now la bor-intensive, such as the alytic steps in accessioning, centrifuging, and distribution of specimens to analyzers. Laboratorians will have to clearly define those areas where robots rather than automated instruments would be most useful.

Manufacturers will have to develop a greater number and variety of programmed modules so that it doesn't take much effort for laboratorians to assemble and reassemble robot workstations. Robots should be flexible automation. They should be able to do different jobs at different times. And if you figure out a better way to do a robot's jow-with, for example, an automated instrument-the robot should not become obsolete. You should be able to turn around and reprogram it to do some other useful job.
COPYRIGHT 1988 Nelson Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1988 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Godolphin, William
Publication:Medical Laboratory Observer
Date:Nov 1, 1988
Previous Article:New Medicare physician pay plan could affect labs too.
Next Article:Upgrade your hiring process with this questionnaire.

Related Articles
Leveling the playing field: the economics of robotics in the hospital clinical labs.
Lab automation still a robust market.
Thermo Electron to acquire CRS Robotics. (Executive Briefing: news, trends & market intelligence for instrument executives).
Lab automation: prospering from productivity gains.
Lab automation: pharma comes to the rescue.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters