College of American Pathologists
Printable Version

  The next step: bringing autoverification into
  the blood bank




October 2007
Feature Story

Suzanne Butch, MT(ASCP)SBB
Theresa Downs, MT(ASCP)SBB

Now that we have automation in the transfusion service and blood bank instruments interfaced with laboratory information systems, it's time to consider adding these instruments to the core lab automation line and autoverification.

This might seem like a radical suggestion since no automated typing and screening instrument for the transfusion service can yet be added to an automation line and since autoverification is only now being accepted as a standard operating procedure in other areas of the lab. But it makes sense. Except for allogeneic stem cell transplant patients, a person's blood type doesn't change, and negative antibody screens predominate. Furthermore, samples requiring technologist intervention may hold up other normal results, delaying the time that blood components are available for the patient.

So why address this now? Because in order for instruments and software to be interfaced in five to 10 years, blood bank and transfusion service specialists need to start asking for such capabilities today. Vendors want to make sure a market exists for a new product before beginning the development process.

Federal requirements for FDA 510(k) premarket approval of blood establishment software add to the delay in adopting new features. Take electronic crossmatch as an example. The FDA just issued draft guidance on electronic crossmatch, more than 10 years after the first facilities were approved to use the technique. And LIS vendors only recently began offering electronic crossmatch features in their software.

So we pose the question, what would it take to satisfy you that instruments are ready for autoverification of results? For us, the software should compare current with previous results. If the past two times the patient has been O-positive and the patient types as O-positive today, we feel confident the specimen and patient are O-positive. There is no need to send the results to a technologist. Likewise, if the patient has always had a negative antibody screen and the current specimen's results are also negative, there is no need for technologist intervention. Just as with hematology and biochemistry, the algorithms determine whether autoverification is appropriate.

One caveat is that the strengths of the reactions used by the instrument or software to determine a patient's blood type must be controlled by the user, not hard coded by the vendor. For example, a weakly positive (1+) result may be used by an instrument to conclude that a patient is Rh positive, but the end user may decide that a 1+ reaction with anti-D is not sufficient to automatically conclude that a person is Rh positive.

Some blood bankers may be slow to embrace the concept of autoverification because of the burden of validation. Since blood bankers, overall, are a conservative group, many may feel more comfortable monitoring for problems themselves. These individuals could review a daily report of patients with positive antibody screens and typing discrepancies to allay their concerns.

One must keep in mind that autoverification with a hematology instrument will be different than with a blood bank instrument. Once a hematology result is autoverified, the laboratory is finished with it. The process is more complex for verifying a type and screen. The blood bank may be using the sample to prepare blood components or may file it away to be used in the future. How will the technologist know that testing is complete and components can be prepared? For blood banks that use electronic crossmatch, samples with one blood type need to be identified for a second blood type. The sample would have to be located and re-tested.

To convey to instrument and LIS vendors what you want their respective software to do, you should start documenting process specifications. These include the goals of autoverification, the type of results that do not need intervention, an explanation of what constitutes a problem sample, or details of how the instrument will identify a problem that needs technologist intervention. A process flow chart could help define the path of a normal versus a problem sample.

Instrument limitations will continue to determine process flow. For example, a chylous sample will influence a gel technology result but might not be a problem with a microtiter plate. Examples of questions that relate to process flow are, Will the repeat test be on the same instrument or use another test method? If the instrument on the line is in another building, and the tube is needed for additional work in the blood bank, how will it be transported? Will the blood bank be completely paperless?

Think about process specifications now and then ask for what you want and need for the future. In other words, ask for today what you want tomorrow.

Suzanne Butch is administrative manager, and Theresa Downs is supervisor, Blood Bank and Transfusion Service, University of Michigan Hospitals, Ann Arbor.