Lab Week Q&A - Bar Codes

 Part two of yesterday's question was about bar code labels.

Wait, are the colors of caps consistent across labs? I seriously always just figured it was an internal thing, and the bar codes on the sides were the important parts of communicating info to other labs if blood had to be sent out there.

 

While the cap color is extremely important in blood collection, because the quality and usability of the specimen depends immensely on how it was collected and stored prior to testing, it doesn't tell the lab what tests to run. The bar code, tied into the Laboratory Information System (LIS), has that job. The way it normally works is that a test is ordered by a physician (usually physically entered into the computer system by a nurse or a member of administrative staff working at the nurses' station), the computer puts that information into a bar code format, which is printed out and brought to the patient's side so the blood can be drawn and the tubes labeled with those bar codes. The bar codes usually list the test and the approved tube color in human-readable language, which is helpful for the staff acquiring the specimens.

Once a tube gets to the lab, the bar code is scanned, acknowledging receipt of the specimen. This time stamp is important because hospitals take turnaround time, or the time it takes from receipt of the specimen to release of a result, very seriously. The scanning is sometimes done manually by an accessioning tech, and then brought to the appropriate area of the lab, but in most larger hospitals, tubes are brought to an automated track system, where they are scanned, sorted, centrifuged if necessary, and brought around the lab where they're sampled directly off the track without a tech ever needing to touch the tube.

Automated track system (Beckman Coulter)

An automated system like that is great for many reasons. It can improve the speed and throughput of the lab, because nothing needs to wait for a tech to bring it to the next instrument. It's good for safety, because any time you can eliminate contact between a person and a blood specimen, you've eliminated a chance of exposure to a bloodborne pathogen like HIV or Hepatitis B. It's got its drawbacks, however.

One big issue is that the instrument doesn't look at the tubes. It only sees bar codes. If a specimen somehow sneaks by without getting de-capped, the specimen probe on the instrument will smash into the cap, expecting a clear path to the serum. The probe has a liquid sensor, and only starts aspirating serum when it feels the top of the liquid. It's expecting the tube to be full - if it's a very short draw, there's a risk that it will keep going and suck up the gel separator in the tube, gumming up all the tubing in the instrument and shutting it down for cleaning. That's why our lab had this sign posted by the inlet of the automated track:

Inlet of the automated line, with warnings!

Also, because all the instrument knows is the bar code, if someone puts the wrong label on a tube, there's no easy way to know unless someone is very careful about checking every tube before it gets put on the line. Ideally, they would all get a check, but when the workload is very high, some can get missed. Normally a label error is immediately apparent, like when a lavender tube, which is supposed to be diverted to hematology, chugs along the track towards the centrifuge to be spun - it's easy to spot because it's shorter than the chemistry tubes. If it gets all the way to the instrument without raising suspicions, say, if a tall pink top is labeled with a Chem-7 bar code, the results will cause the tech to pull the tube off the line and have a look at it. The potassium will be so high as to be incompatible with life!

Bar codes make it easier for the lab staff, because it saves us the trouble of having to program all the tests in manually for every single specimen. It saves us a lot of time when the instrument can read the orders and get the tests done. However, when specimens are sent out to other reference labs for more esoteric testing, they can't read the barcodes produced by our internal computer system. We have to fill out paper requisition forms and then send the tubes off (or poured-off serum) in special packaging, sometimes on dry ice, sometimes at room temperature, whatever the reference lab requests. Some hospitals have the system computerized and can fill out online requisitions and print out reference-lab-specific bar codes for the specimens - because the reference lab will have their own computer system and won't be able to use the original bar codes. It's like our bar codes are in German, but the other lab needs them to be in Finnish. We can either give them all the information on paper and they'll make their own Finnish labels when they get the specimen, or we can use their software to do that ourselves. That requires a special arrangement between the two labs, and not everybody has the resources to do it.

Even with the help of the bar codes, we still need to order things manually sometimes when tests are added on later, or if tests need to be repeated, but the only time the lab does everything manually is when the computer system goes down.

Two bad things can happen with the computers. The hospital system can go down, or the lab system can go down. In our facility, it was two separate systems, so although they could speak to each other, either one could die independently, leaving us with two different kinds of messes. If the hospital system goes down, the floors send down the tubes with patient-information labels that look like the patients' hospital armband. They include a "downtime" order form where they can check off all the tests they want run. Then we can order the labs on our end, getting bar codes, sticking them over the other labels, and proceeding as usual. The instruments still know what to do, the results still come onto our computer screens for analysis, and we can call, or print out and fax, the results to the floors. That's tedious, but not terrible. When the lab system goes down, it's chaos. We log receipt manually and program tests manually. We keep reference sheets handy to confirm panic values and normal ranges, and we fend off calls from the floors asking why everything is taking so long. We keep piles of instrument printouts so we can later manually enter everything into the computer. Some instruments will allow you to re-send the data once the system is back up, but not all of them. We love bar codes and computers. Very, very much.