Printer Friendly

We carved out bar coding goals - and reached them.

IMPLEMENTING a bar code system as part of a laboratory information system (LIS) is a learning experience. It requires mastering new types of hardware, new programming techniques, and a new vocabulary. When l started ours in early 1991, all I knew about bar codes was that grocery stores use them.

The implementation provides a learning experience for the software vendor, too. When we were initiating our program, the vendor was still perfecting ways to make bar codes work with the other parts of the software and how to induce laboratory instruments to read the codes.

Meditech (Medical Information Technology, Inc., Westwood, Mass.), the system used at our 250-bed not-for-profit hospital, runs on a Vax computer (Digital Equipment Corp., Maynard, Mass.). We are currently using Meditech's Lab and Micro modules. We have 16 terminals, six printers, and four interfaced instruments. Meditech software is used throughout our hospital, including nursing, pharmacy, medical records, and the business office.

* Defining needs. My first objective for the system was to print bar codes on specimen labels. Almost any printer is capable of printing a bar code; however, the quality needed by most applications requires either a dedicated bar code printer or a laser printer. Since the Meditech software supports only two bar code printers, my choices were limited to those. I chose a Prodigy printer (Fargo Electronics, Inc., Eden Prairie, Minn.) because it cost less than the alternative and seemed at least as good.

The Prodigy offers two methods of printing: direct thermal and thermal transfer. Direct thermal printing, which does not use a ribbon, burns an image into the label stock. Thermal transfer melts wax from a ribbon onto the label stock. Thermal transfer has the advantages of using a slightly less expensive label stock and requiring less heat, thus endowing the print head with a longer life. Considering that each print head costs about $500, the saving is significant. Unfortunately, I found that the disadvantages outweighed the advantages. For example, the wax on the labels is easily scratched. Our hematology analyzer, a Coulter STKS (Coulter Corp., Miami, Fla.), holds the tube in a cassette and pushes it forward to read the bar code and sample from the tube. This process scratches a thermal transfer-printed bar code so badly that most of the codes will not read. I have been told that thermal transfer labels blur in ice water. Neither of these problems has turned up since we decided to use the direct thermal process.

The software required to run a bar code printer (a printer driver) is highly instrument-specific. For example, the driver that runs the Prodigy will not work with Fargo's newer model, the Prodigy Plus. Unless your software vendor tells you that a particular printer will work, don't assume that it can be made to work. Some printer vendors will assume the responsibility for providing a working interface to your LIS.

The other piece of hardware that you may need is a bar code reader or scanner. Many of the major analyzers come with built-in scanners. We wanted a hand scanner so that we could scan the label on the tube directly into our specimen reception routine. The scanner we selected, which is attached to a system terminal with a T-connector supplied by the manufacturer, draws its power from a transformer that is plugged into a standard electrical outlet.

Once attached, the scanner must be programmed by being made to read special bar codes that the manufacturer sends with them. The one we use (Scan Plus, BarCode Industries, Beltsville, Md.) lets us select a wide variety of bar code types, baud rates, and other parameters.

The vast array of bar code scanners on the market work in pretty much the same way. Unless your bar code software vendor specifies which scanner to use, pick any one at a reasonable price that seems able to withstand hard use and for which service is easily obtained. Unlike printers, scanners do not require special driver software.

The next consideration was to find label stock that would print readable bar codes and stand up under laboratory conditions. Like all laboratory computer labels, these labels had to be able to withstand temperatures from the freezer to heating blocks, stick permanently to tubes and culture plates, and not stick to gloves. No thermal label works in a 100|degrees~C heating block, which would blacken the whole label. Since we use a centrifugal slide stainer for hematology, we needed a label that would remain readable after staining. Having a wax coating on the label would do it.

Since the printer does not let us adjust the position of the paper, we needed labels with a backing made to a precise width and with die cuts positioned correctly so the labels would contain all necessary data and still fit properly on test tubes or slides. One label vendor, TimeMed Labeling Systems, Inc. (Burr Ridge, Ill.), had worked with Meditech's beta test site for bar coding and had already worked out the perfect label. After looking at another vendor's labels, we set up a standing order with TimeMed. Later, when we found that it was hard to write on the wax coating with ink, we substituted pencils or wax markers.

* Code types and parts. Many types of bar codes are available, each optimized for a different application. The UPC codes found in grocery stores are probably the most familiar type of bar code. Three types were supported by both Meditech and the instruments in our laboratory that would read them: Code 3 of 9, Interleaved 2 of 5, and Codabar.|1~ Standard Code 3 of 9 (also known as Code 39) can incorporate digits, a few symbols, and uppercase letters. Interleaved 2 of 5 encodes only digits, while Codabar can encode digits and a few symbols. I chose Code 3 of 9 because of its ability to encode letters. (Although that attribute doesn't provide any advantage to us now, it may lend some flexibility in the future.) Considerations in selecting a bar code type include the ability of a lab's software to print it and of the instruments to read it.

Bar codes have several parts. Start and stop characters tell the reader when to read. The start and stop symbol used in Code 3 of 9 is an asterisk. Bar codes can include a check digit that checks the code's accuracy. In Code 3 of 9, the check digit is the modulus 43 of the bar code number (the remainder when the bar code number is divided by 43). Most bar code labels contain a human-readable interpretation line as well.

To choose a bar code, I consulted the lab that had done the beta testing for Meditech's software. Their experience led me to select Code 3 of 9 without a check digit.

* Labels. Labels that will be used for patient specimens have to contain a lot of information, including the patient's name, location, and account number; tests ordered; specimen number; bar code number (if different from the specimen number); the type of container to use; and time when the specimen is to be collected. Space should be available for the collector to write his or her initials and the time of actual collection. Initially Meditech did not provide all of this, but by working with them we were able to add the missing information. We are now testing a new release of their lab system with the ability to custom design bar code labels.

Meditech's bar code is not the same as its specimen number. In Meditech's Lab module, the specimen number consists of a date, a section prefix, and a consecutive number that resets to 1 daily. For example, the first chemistry specimen on February 4 would be assigned the specimen number 0204:C1. The bar code representing this specimen contains six digits because that is what our analyzer requires. These numbers are generated sequentially as the labels are printed. Meditech chose this method to allow the simultaneous use of several bar code symbologies.

The biggest challenge I faced was in getting instruments to match the bar code number to the specimen. Three instruments in our laboratory are able to read patient specimen bar code: a Kodak XRC 700 (Eastman Kodak Co., Rochester, N.Y.), a Coulter STKS, and an ACL 1000 coagulation analyzer (Instrumentation Laboratory, Lexington, Mass.). The ACL 1000 uses a hand-held scanner, while the others are built in.

The first instrument to work as I wanted was the Kodak. Here again the beta site was very helpful. Using a Kodak, people there were able to explain all the settings we needed, including bar code type, length, and whether a check is used. Bar code length was particularly confusing because our code consisted of six digits and a zero, which printed because there was no check digit. The Kodak required seven digits. Meditech has since eliminated this zero so that the six-digit bar code is really six digits long. Unless the length is set correctly, the software will not match the bar code number to a specimen. Once these details were taken care of, there was little problem getting the instrument to read bar codes correctly.

Working with the Coulter was more of a challenge. The interface between Coulter and Meditech is unidirectional; that is, data will flow only from Coulter to Meditech. Our first problem was in getting the Coulter to read a bar code printed by Meditech. l learned from a telephone call to Coulter's technical service that the instrument would not recognize the bar code without a check digit. The way around this was to pull out one of the computer boards and to change a switch setting. We located the correct board and changed the switch. When we replaced the board, however, something was damaged and the Coulter was down until the service representative could stop into the lab the next morning.

The hematology staff was less than happy with me. At the suggestion of the Coulter service rep, I defined a second bar code type that included a check digit. While he repaired the damage, I created several specimens with the new bar code type attached to them. We tested them and they read. This shows that software changes are much easier to make than hardware changes.

The next problem was totally self-induced. I did all my initial testing using one brand of labels; then I selected a different brand to use. I noticed that the print on the new labels was lighter than on the original labels but did not think this would pose serious difficulties. I was wrong. Coulter's bar code scanner read less than half of the new bar code labels. It took calls to Meditech and TimeMed to figure out what was wrong and a formal written request to darken the labels. In retrospect, I should have tested the new labels right away to see how well they read.

* New instrument. We purchased our ACL 1000 and its interface while we were setting up our bar code capability. Meditech had interfaced this instrument before but had never used bar codes with it. Like Coulter's, the interface is unidirectional. Once the details of the interface were worked out, the bar codes worked easily. We assign a specimen to a position on a tray by scanning its bar code with ACL's hand scanner.

We can also use bar-coded control specimens on this instrument. We assign a bar code to each lot of control material within Meditech and place a bar code for each level on the front of the instrument. When a technologist scans this bar code with a wand, the screen shows the bar code number and indicates where the specimen should be placed on the machine. A prompt then instructs the technologist to scan the next specimen. When the last specimen has been entered, the tests are run. The technologist then verifies the test results and runs appropriate Westgard rules. While we could have followed the same procedure with our other analyzers, various reasons (too complex to go into here) led us to decide against doing so.

* Staff reaction. The initial response to bar codes by the laboratory staff was less than enthusiastic. Our specimen labels had been one of the most trouble-free aspects of our Meditech system and the staff saw little need to change. The print on the bar code labels was smaller and the wax coating on the labels often made it difficult to read. Darkening the print helped.

A typical day starts when our third-shift technologist prints our collection labels, which include bar codes. Labels for Stat collections are printed automatically; new orders are printed periodically throughout the day. The person who collects the specimen places the bar-coded label on the tube and writes the time collected and his or her initials on the label. Specimens collected by nonlaboratory staff are labeled when they come to the lab. Die cuts on the label provide smaller labels for aliquots and slides. Meditech can reprint any lost or soiled labels.

Specimens come to the laboratory and are distributed to the appropriate sections. Bar codes on specimens headed for chemistry tests on the Kodak tell us that we do not have to pour cups but can sample directly from the original tube. It didn't take long for us to realize what a time saver this would be.

Bar codes have made the Coulter easier to use as well. Without bar codes, each specimen that was sent from Coulter to Meditech had to be matched to a Meditech specimen number manually. Bar codes have made it possible to eliminate this extra step.

The critical stages in implementing a bar coding system are choosing a printer and label stock and "convincing" the instrument interfaces to recognize the bar-coded specimen. Any time a change is introduced, even if it seems insignificant, thorough testing is required to prevent the kinds of problems I had when I switched label sources.

* Real time. For the future, I am looking for a real-time specimen collection device, which I imagine working as follows. First, a bar code would be placed on the patient's wristband. (A tattoo might be better, but probably wouldn't be socially acceptable.) Each person collecting specimens would carry a hand-held terminal to which the collection batch would be downloaded along with the information on bar-coded collection labels. A bar code scanner would be built into the device, which would weigh no more than one pound for ease of use. The phlebotomist would collect the specimens, read the bar code containing the patient's ID, read the bar codes on the collection labels specifying details about the tests that had to be done, and trigger the device to "stamp" the time of collection in the computer. When the phlebotomist returned to the laboratory, this information would be downloaded to the host computer.

With the arrival of near-patient testing, such devices will undoubtedly be used to record results at the bedside. Other applications are sure to develop as the technology becomes available. I look forward to them.

James A. Sehloff, M.S., MT(ASCP) is LIS coordinator at Holy Family Medical Center, Manitowoc, Wis.


1. Kasten BL. What are all those lines and spaces? Understanding bar code technology. MLO. December 1992; 24(12): 25-27.
COPYRIGHT 1993 Nelson Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1993 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Computer Dialog; Medical Information Technology Inc., Fargo Electronics Inc., TimeMed Labeling Systems Inc.
Author:Sehloff, James A.
Publication:Medical Laboratory Observer
Article Type:Column
Date:Feb 1, 1993
Previous Article:Use patterned interviewing to reveal 'people skills.'
Next Article:Avoid the seven deadly sins of management.

Related Articles
Bolder bar codes: squeezing more information into those black-and-white symbols.
Joining the bar code revolution.
Hardware for reading and duplicating bar codes.
Bar coding: the ideal system.
Quick retrieval of analyzed specimens.
Reading between the lines: using bar-code technology is a smart way to keep track of business data.
The future: faster analyzers, better LISs, more bar coding.
Getting numbers to tell the story.
TimeMed acquires Rx products firms.
TimeMed awarded Premier contract.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters