Specimen labeling QI project cuts error rate
Weekly feedback to ED aided improvement
In one year, the emergency department's contribution to major mislabeling events at Boston Medical Center was reduced from 47% (23/49) to 14% (4/29) thanks to a proactive quality improvement program initiated by Karen Quillen, MD, of the department of laboratory medicine. Her work was recently published in the American Journal of Nursing1.
The process involved the targeting of major mislabeling events and the design and implementation of a plan of corrective action.
All mislabeled specimens received by the transfusion service for a type and screen were recorded and classified into minor and major mislabeling categories.
Then, the major mislabeling events were tracked by origin of the specimen, and the areas with a high proportion of major mislabeling were given feedback within one week of the events.
"In lab medicine, you are always talking about pre-analytic and post-analytic variables, such as instruments, broken equipment, and so on," explains Quillen. "My best analog is the old 'garbage in, garbage out,' example. You can have the most sophisticated instruments operating at lightning speed, but if the sample going in is not for the right patient's blood, the results will be garbage."
Targeting inconsistencies
She notes that some sources estimate that as much as 70% of medical decisions are based on lab results, "so you have to be really sure you can verify that it is Mr. John Smith's blood that you drew."
Acute hemolysis (breakdown of red blood cells), primarily the result of ABO (the four main blood types) incompatibility, is an important cause of transfusion-related mortality and morbidity, note the authors.
"I targeted blood group inconsistencies," Quillen shares. "I would get blood group results based on a sample, but when I compared it to the patient's historic record, some had different blood types two months or two years, ago. We know your blood type cannot change, so one of those times it was wrong."
Targeting the ED
At the start of the research, Quillen was determined to look at the process and not at individuals. "We collected all relevant data, and only when we looked at it after a year did patterns emerge," she says. "We did not focus on the ED, but they do a lot of labeling, and from the prior data analysis we could tell where the high-problem areas were."
There was one other area where problems were high, she adds, "and we are just now seeing results."
It was important to work with the ED once it was identified as a high-problem area, she continues. "ED samples sent to me in the blood bank don't always request transfusions — sometimes it's a 'just in case' situation when they're doing a workup. But if they do order a transfusion and we don't pick [the mislabeling] up, it could potentially be disastrous."
The weekly feedback was provided to the ED by e-mails from Quillen to the nurse manager. "It would say that on such and such a date, this patient had this testing inconsistency and request that she investigate," says Quillen. "Since the error was relatively fresh, she could tell who had been working on the patient, who drew the blood, and figure out how they sent a mislabeled specimen."
Prior to this initiative, she notes, the ED would receive quarterly results, "But by then the trail was cold."
Non-punitive approach
Even though it was possible through this process to determine who did or did not do what, "We definitely took a non-punitive approach," says Quillen. "We started with the premise that these errors happened not because someone wanted them to — that it was always a system fault. Perhaps there was not enough staffing, or the printer label was broken. Our goal was to improve the process, not single out one person."
The ED analyzed its work flow and uncovered a number of potential sources for the errors.
"For example, in a busy ED, it's not uncommon to draw a lot of blood in case something happens in the evaluation; they get as much blood as they might need, so they do not have to stick the patient again," Quillen explains. However, if a computerized system is used to generate specimen labels, there will be no label for these extra tubes unless they are handwritten or manually stamped addressograph labels are used.
"We've now designed a function into the computer system so you can get extra labels when you do not have a specific test attached to them," Quillen notes.
Accurate patient identification at registration is another potential problem; for example, two patients may share the same common name.
To prevent such errors, the authors recommend requiring two hospital staff members to verify patient identification prior to phlebotomy; requiring photo ID prior to phlebotomy; or requiring two independently drawn samples for new patients in the transfusion service. New technology such as barcode or radio-frequency identification chips for patient identification also are suggested.
Could this project be replicated at any facility, regardless of size or budget? "Absolutely," says Quillen, "But they have to collect their baseline data first. While the ED was one area to focus on here, it may not be the same at another facility."
She re-emphasizes that "engaging the area you work with, and the collaborative approach to improve" will give you the greatest chance for success.
"Try not to point fingers and say someone made a mistake," she advises.
Reference
1. Quillen K, Murphy K. Quality improvement to decrease specimen mislabeling in transfusion medicine. Am J Nurs. 2006 Sep;106(9):58-70.
For more information, contact:
Karen Quillen, MD, Department of Laboratory Medicine, Boston University Medical Center, Boston, MA 02118. E-mail: [email protected].
In one year, the emergency department's contribution to major mislabeling events at Boston Medical Center was reduced from 47% (23/49) to 14% (4/29) thanks to a proactive quality improvement program.You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content