Reducing error rates in the clinical lab is best accomplished through identifying, evaluating, and monitoring vulnerable areas in the total testing process, according to the tenets of the risk management process profiled in the last issue of Strategies . The second part of this two-part series will examine how the tools of risk management were applied at the University of Kansas Hospital 's clinical laboratory in Kansas City to reduce patient identification and specimen labeling error rates.
Soon after Shirley Weber, MT (ASCP), MHA, was appointed Assistant Laboratory Director at University of Kansas Hospital (Kansas City, Kan.), she set out to launch a risk management program, but she knew that first she needed to assemble the right team for the project to be a success. “You need to get caregivers to come to the table,” she explained, adding that these need to be individuals who work at the bedside, such as nurses drawing and labeling specimens, as well as the lab and medical directors, because these people know and understand the barriers and problems. “You also have to have a focus in your hospital that patient safety and quality is absolutely the highest priority,” added Weber, who is now Director of Laboratory Services. “Once everyone understands that you are doing this because you want to improve patient care, then everyone will likely be on the same page.”
This process, which was the focus of a recent workshop sponsored by the Clinical Laboratory and Standards Institute (CLSI), “Risk Management Tools for Improved Patient Safety,” is similar to other quality improvement programs in that it is rooted in standards published by the International Organization for Standardization (ISO). The world's largest developer of standards, ISO published a risk management standard for medical device manufacturers (ISO 14971) in 2000, and is currently working on one for the clinical laboratories (ISO 22367).
Applying Risk Management Tools
When Weber and her team began the first phase of the process—risk analysis—they used the failure mode and effect analysis (FMEA) technique to investigate three specific areas: patient identification, specimen collection, and specimen labeling. For each of these areas, they relied on FMEA, fault tree analysis (FTA), and appropriate flow diagrams to address the following questions: where is there failure? What is the failure mode? What is the frequency of each failure and what are the likely clinical implications of each failure? (See Box 1)
Box 1 (click on thumbnail to see larger image.)
Box 2 (click on thumbnail to see larger image.)
After analyzing and evaluating the risks, Weber and her team had a clear picture of their problem areas in patient identification and specimen labeling, and began implementing some changes or interventions (see Box 2). The first major intervention involved a new patient identification policy requiring that bracelets with patient identifiers, such as patient name, medical record number and date of birth, be provided for all inpatients and any outpatients undergoing an invasive procedure. The use of two patient identifiers is now mandated by the Joint Commission on the Accreditation of Healthcare Organizations, which accredits the University of Kansas Hospital.
Tightening Specimen Collection
In addition to revamping the patient identification policy, Weber and her team developed a new restrictive specimen acceptance policy in the lab that does not allow for relabeling of mislabeled or unlabeled specimens, except if recollection would be a risk—as would be the case for tissue specimens collected in the operating room or cerebral spinal fluid collected by lumbar puncture. In situations when relabeling is considered medically necessary, the person who collected the specimen must come to the lab, identify the specimen, and fill out appropriate documentation for the medical record. In addition, specimen tubes are no longer allowed to be labeled prior to a specimen draw.
But there were other problems in specimen collection that needed to be addressed, according to Weber. The team found high error rates when test ordering information was input to the computerized order entry system, because it required staff to translate the physician's written order into the corresponding order in the computer system. There were two antidotes to this issue. One, the computer order entry screens were updated to match the written order sets currently used by the physicians, and two, the laboratory leadership began meeting with the unit secretaries on a regular basis to address issues of confusion.
Another problem in specimen collection was using the wrong container for blood draws, an error committed primarily by non-laboratory staff, according to Weber. To combat this problem, they changed the hospital's IT database so that it communicates which tube or container was to be used for which type of test. This online database is now regularly updated.
Looking Forward to Zero Error Rates
All of the interventions introduced as a result of the risk management process proved challenging for various reasons, primarily because it required the staff to change their practices and routine methods of labeling specimens. “They were very used to doing things the old way, which included not confirming identification and labeling specimens at the bedside, but it was extremely important that they change their practice,” Weber said. “We still struggle with it today. We're much better than we were, but we continue to educate.”
But the revamped policies and other changes were soon validated. “Most importantly, the changes the team implemented made a positive impact; the labeling error rate initially declined from 0.19% to 0.024%. However, our goal is to have zero errors,” Weber explained. “When we started this endeavor, our goal was to just improve the error rate, but now we would like to eliminate labeling errors entirely. We got another team together—again, made up primarily of bedside caregivers—and created additional tools that helped us further reduce the error rate. We developed immediate notification forms so that when there is an error, we send a fax to the unit, and the charge nurse immediately investigates what happened while that person who made the error is still on duty, so that they can together talk about where they fell short on what they should be doing.”
This immediate notification system further reduced errors, but Weber believes that any error-proof method will require bar coding technology. “That's where we need to go. We're looking at a couple of different vendors right now that have solutions, but we have to find a solution that works for the nurses and caregivers,” she said.
For more information:
- The first part of this Strategies series can be found online.
- Additional information on ISO 14971 and the forthcoming ISO 22367 can be found on ISO's Web site, www.iso.org .
Julie McDowell is the Editor of Strategies. She can be reached by email