In the United States, one in 600,000 blood transfusions results in death because the patient, the patient’s blood sample, or the blood container is misidentified. Most of the errors occur on nursing units, not in laboratories.
This year, about 12 million transfusions will take place. And unless we change the way we identify patients, their blood samples, and blood containers, 20 patients will die.
The good news is that we have the technology to ensure the positive identifications of patients, their blood samples, and blood containers, as we demonstrated using a prototype bar-code system.
During the past few years, several hospitals in the United States have introduced bar-code–enabled point-of-care (BC-POC) systems to help caregivers reduce errors in dispensing drugs to patients. Typical BC-POC systems include bar-coded wristbands, drug containers, and bedside scanners. These systems decrease the risk of wrong drug or wrong dose to wrong patient.
At Georgetown University Hospital, we recognized the potential of BC-POC technology for reducing the risk of human error in blood transfusions. We partnered with Bridge Medical to evaluate their prototype MedPoint-Transfusion system on our Inpatient Blood and Marrow Transplant/Hematology Unit.
Using our prototype system
Our prototype system included bar-coded wristbands, a handheld bar-code scanner, a portable blood sample label printer, and a blood bank analyzer that was bar code–enabled and fully automated. When a physician ordered a blood transfusion, our BC-POC procedure began with a nurse printing a bar-coded wristband for the recipient, using data from the hospital information system (HIS). This wristband contained three identifiers: first and last name, medical record number, and date of birth.
The nurse then generated bar-coded labels for blood samples and for the blood bank order forms. The three-step procedure consisted of the nurse scanning his or her bar-coded identification tag (electronic signature), scanning the patient’s bar-coded wristband and, using the handheld scanner, beaming the information to a portable label printer.
The bar-code–labeled tubes and order forms were delivered to the blood bank, where the blood components were matched and labeled with a bar-coded crossmatch label, designating a particular blood container for a specific patient. The results of ABO and Rh testing were transmitted from the automated blood testing analyzer in the blood bank to the laboratory’s information system. And the blood bank issued the bar-code–labeled blood components to the nursing unit with patient-specific bar-coded crossmatch compatibility labels.
Before starting the transfusion, the nurse performed the standard visual identification and verification steps to match the patient and the blood component. Then a second nurse performed a bar code–facilitated identification check. Nurses performed this BC-POC patient–blood container verification check by scanning their bar-coded identification tags, the patient’s wristband, and the blood container. If an electronic match existed, the screen projected a signal to start the transfusion; if not, an alarm sounded. After verification, the nurse returned the handheld scanner to the cradle, where it downloaded the acquired information to the HIS server.
Evaluating our prototype system
We evaluated the BC-POC prototype system in stages to ensure early feedback to the manufacturer. After the first 25 transfusions, we met to discuss any concerns of the nursing staff. We found that the main concerns were all related to the learning curve for using new hardware and software. The more the staff used the system, the more comfortable they became. The training seemed to meet the needs of the staff, but we recognized that nurses need to use the BC-POC system frequently to be proficient and confident.
We also identified technical problems. For example, the identifiers from all sources needed to match exactly. If a zero preceded the medical record number on the blood bag, but not on the identification band, the scanner didn’t accept the information and aborted the process. Also, the scanner required eye-hand coordination and practice to align the beam when scanning and printing bar-coded labels. Grocery stores use expensive multi-laser, omni-directional scanners, but our cost-effective handheld scanners operated with only one laser beam, so nurses needed more skill to align and scan.
After 50 transfusions, we recognized a need to streamline and simplify the process by limiting the number of times the blood container needed to be turned front-to-back to scan required information. We also suggested a way to limit the number of bar codes, while maintaining safety.
After 100 transfusions, we modified the software program to make it match the hospital’s patient-identification routine. After this change, nurses reported that bar-code enabled transfusions were easier and more efficiently administered.
After 25 more transfusions and more feedback from nurses, we decided that to ensure proper identification, the person who affixes the identification wristband shouldn’t be the person who scans the blood container. Instead, on admission, all patients should receive one bar-coded wristband that stays with them throughout their hospitalization. The wristband could then be used to identify patients not only for transfusions, but also for procedures such as drug administration and diagnostic tests.
The change to the BC-POC approach wasn’t easy. Some nurses weren’t comfortable learning new technology and maintaining proficiency with a BC-POC scanner. The comfort level they achieved during our intensive training diminished if they didn’t use the scanner frequently. Also, the price tag for installing a BC-POC system in our hospital was about one million dollars. For the system to be cost-effective, we’ll have to use it in other areas, such as the pharmacy, general laboratory, and radiology department. Combining software programs for several applications on one scanner platform may yield two benefits: increased safety throughout the hospital and increased use—and acceptance—of the system by all nurses.
Putting aside the questions of cost and acceptance, our experience was remarkable. We found that the bar-code readings of wristbands, blood sample labels, and blood component labels were 100% accurate. All information on bar-coded order forms was also accurate, complete, and legible. Blood bank requests to nursing units to redraw blood samples because of mislabeled tubes or erroneous order forms simply stopped. In short, we experienced increased efficiency in ordering and transfusing blood components—and, of course, increased patient safety.
America Association of Blood Banks. Standards for Transfusion Services and Blood Banks. 24th ed. Bethesda, Md: America Association of Blood Banks; 2006:55.
Dohnalek L, Cusaac L, Westcott J, Langeberg A, and Sandler G. The code to safer transfusions. Nurs Manage. 2004;35:33-36.
Sandler SG. Technology can reduce errors in blood transfusion. In: Lewis RF. The Impact of Information Technology on Patient Safety. Chicago, Ill: Healthcare Information and Management Systems Society; 2002:57-66.
U.S. Department of Health & Human Services. HHS announces new requirements for bar codes on drugs and blood to reduce risk of medication errors. Available at: www.fda.gov/bbs/topics/news/2004/hhs_022504.html. Accessed March 28, 2007.
Laurie J. Dohnalek, RN, MBA, CNA, is Director, Inpatient Oncology and Blood and Marrow Transplant, Apheresis and Dialysis, Georgetown University Hospital, Washington, D.C.
The author thanks S. Gerald Sandler, MD, Director, Transfusion Medicine and the 2 Bles Staff at Georgetown University Hospital for their roles in this research project.