Ford used to advertise that “quality is job one.” The slogan may no longer be mainstream, but quality has become job one in many industries—medicine included. Since 1999, when the Institute of Medicine (IOM of Washington, DC) published its report “To Err Is Human: Building a Safer Health System,” patient safety has become a rallying point for many in the medical community. But rather than steal a slogan, health care grabbed something else from the manufacturing industry instead—quality-management programs, such as Six Sigma.
Today, laboratories implement a wide range of programs to manage quality, varying in size, scope, and philosophy. They use the same guidelines, such as the Clinical Laboratories Improvement Act, but differ in how they achieve the objectives. “The best practice in quality management is to have a quality-management system in place,” says Lucia Berte, a laboratory consultant based in Westminster, Colo.
Berte stresses that a quality-management system, or QMS, looks at all of the laboratory processes, whereas programs such as Lean management and Six Sigma are smaller in scope. “Lean and Six Sigma can make significant improvements, but for only one process at a time. In the meantime, the rest of the lab is doing what it always does. A QMS systematically looks at all 12 of the quality essentials in the lab. A Lean or Six Sigma project may take a few months to implement. A QMS is a journey that can take a couple of years,” Berte says.
A QMS addresses every process the laboratory encompasses through 12 quality system essentials (QSEs). According to Berte, these are: organization, personnel, equipment, purchasing and inventory, process control, documents and records, information management, nonconforming-event management, external and internal assessments, customer service, process improvements, and facility and safety.
Quality control, quality assurance, proficiency testing, external inspections, equipment maintenance, and calibration programs all fall within the essentials of quality management. Lean management and Six Sigma are also pieces of a quality-management system, Berte notes.
“Lean is a way to reduce waste in work processes, making it part of the quality system essential process control. Six Sigma, which aims to reduce error, is a method for process improvement,” Berte says. The two systems are not exclusionary, but they do compete for laboratory resources.
Organization
In trying to reduce cost, improve efficiency, and/or better use resources, most labs will often look at a single process, according to Berte. “For example, a lab will apply Lean or Six Sigma to the sample-receiving process. But when the project is all done, the lab has not done anything with its other processes. The question becomes, ‘Where do they want to spend their money?’ ” Berte says.
Berte suggests the least expensive and most comprehensive way to get started is to read and understand the two most relevant guidelines published by the Clinical Laboratory Standards Institute (CLSI of Wayne, Pa): “HS1-A2 A Quality Management System Model for Health Care” and “GP26-A3 Application of a Quality Management System Model for Laboratory Services.”
HS1 incorporates government and accreditation requirements into the 12 QSEs, giving labs an idea of what needs to be done to start a QMS. GP26 takes the same QMS model and actually applies it to the technical work done in the lab.
From there, if a laboratory does not have the resources to implement an entire quality system, Berte suggests there are some essentials that are “absolutely” essential. “One of these pieces is nonconforming-events tracking, which requires every person in the lab to detect and record events that do not conform with the established process or procedures, whether or not they brought harm to a patient. If you can’t have the whole system, this element is important,” Berte says, adding that document control and process improvement are two other essentials.
Guidelines
“Guidelines, in general, are developed through different mechanisms and are based on an apparent need identified through some sort of data collection mechanism or other method,” says Devery Howerton, PhD, chief, Laboratory Practice Evaluation and Genomics Branch, Division of Laboratory Systems, Centers for Disease Control and Prevention (CDC of Atlanta).
Howerton uses the rapid HIV testing guidelines published by the CDC on its Web site. “The guidelines were needed because a new test had been released that was waived and would be available to almost any lab that wanted to offer HIV testing. This was a major change in practice, so guidelines were developed in response to the perceived need,” Howerton says. Now, these guidelines are currently in revision to eliminate specificity to that first HIV test.
CLSI has a note on its Web site that asks experts to submit their ideas for guidelines, which are then evaluated for relevance to CLSI goals, the perceived need of constituents, and potential market value. Currently, the organization is developing guidelines in 10 areas, including evaluation protocols that will address the ambiguity in equivalent quality control, or EQC.
Process Improvement
“The regulations and standards tell people what to do, but they do not prescribe how it is to be done. So the lab needs to figure out how to do it,” Berte says, adding, “That’s what guidelines provide.” She recommends that in addition to the guideline documents, laboratorians can also learn about quality management and requirements through articles in professional journals, as well as industry publications that share case studies and first-hand experiences.
Jonathan Stein, PhD, director of science and research for SpectraCell Laboratories in Houston, Tex, thinks quality-control officers make sense for large organizations. “The quality-control officer looks at a protocol and every step and asks how we can be sure this is happening and how we can identify when it isn’t,” Stein says. His lab employs a quality-control officer, and this same person acts as SpectraCell’s liaison with inspectors and accreditors.
Stein believes that the key to quality management is metrics. “Some of these management programs, such as Six Sigma, were originally designed for manufacturing. Mostly, they are attempts to codify fairly good ideas, but they are very broad: define, measure, analyze, improve, and control. That sounds great, but then you have to look at what you do. Do you have a metric? Do you know when it works? Do you know when it doesn’t work? That’s essentially the approach the programs take. You have to make that fit your methods or protocol,” Stein says.
Process Control
And you should do so before a problem arises. “It’s sometimes hard to develop a program when everything is working, but you want to do it in advance. There is a tendency to ignore the processes that are working or that have never failed up to this point. But you don’t want to ignore them, because if you’re not monitoring them, it’s hard to troubleshoot when or if they do fail,” Stein says.
SpectraCell did just that, looking at all of its processes, from washing tubes to labeling samples to accessioning plates to dealing with orders, with quality and metrics in mind. Stein offers its liquid-handling dispenser as an example. The lab cultures cells in 96-well plates, into which the system aliquots 200 microliters of solution. “The question was, ‘Is the system doing that uniformly and if not, when and why?’ ” Stein says.
So the lab bought a plate reader, determined the variability within the sensitivity of the reader, and set up a process to run lots to determine proper functioning. “Previously, the technician eyeballed the wells to see how equal they were, making it dependent on the tech rather than a straight numerical analysis,” Stein says.
Personnel
Labs that rely on human resources to provide aspects of quality management, such as quality control, will find themselves at a disadvantage. “A certified lab has to rely on process and not personality. What you really need are well-defined procedural manuals that include every step of your process,” Stein says.
He also suggests that an annual review of all of the lab’s procedures is important. “As technologies change, you can have drift, and you don’t just want to rely on knowledgeable employees,” Stein says.
There is a quote in the Institute of Medicine 1999 report that says, “Medical error is a failure of process.” “Although that paper was talking about medical error, lab error is also a failure of process. The most common errors are due to variations in the way the process is conducted in the lab,” Berte says.
Different people have different ways of doing the same thing, which leads to variable results. Berte notes that this is typically more complicated than just simple pipetting. “It’s the outcome of the whole process and not just the test results that are variable,” she says.
“Being consistent in applying quality-management rules is the hardest task for any quality procedure,” says Fred Meier, MD, division head, system laboratories, Henry Ford Medical Group Laboratories (Detroit).
The QMS Journey
Yet, consistency is important at all phases of the laboratory process. QC looks at the most consistent part of what a laboratory does, which is the actual analytic phase. “But most of the mischievous variations in lab testing come before and after the analytic phase. For instance, before, we have the selection of tests, the identification of the patient, and the collection and transport of adequate specimens. After, we have verification of results, their timely transmission, and appropriate reception by the proper person. QC looks at just the testing in the middle—the generation of the result—which is the most consistent part of the whole process,” Meier says.
Meier believes the responsible thing for laboratorians to do is to try to decrease the amount of variability in the preanalytic and postanalytic phases, which QMS systems have at their core. Slowly, these concepts are also becoming standardized.
Berte notes that US guidelines are slowly incorporating requirements defined by the ISO 15189:2003 standard, which addresses a significant portion of a QMS. “Organizations such as the Centers for Medicare and Medicaid Services [CMS of Baltimore], the administrators of CLIA, COLA [Columbia, Md], and the College of American Pathologists [CAP of Northfield, Ill] are slowly incorporating requirements from ISO 15189 into their own standards and requirements,” Berte says.
The National Committee on Quality Assurance (NCQA of Washington, DC) has not incorporated ISO standards, but it does include performance metrics in its guidelines. The agency accredits health plans and providers, who interact more directly with consumers than labs. Which laboratory tests to run fall within its purview.
But what really sets NCQA apart from other medical accrediting agencies, according to Gregory Pawlson, MD, MPH—the organization’s executive vice president—is that NCQA makes its performance measures transparent to the public. “We require health plans to perform a consumer assessment and use that score in our rankings,” Pawlson says.
Performance indicators consist of clinical-performance measures, including misuse, overuse, and underuse. Results are made available to the public. The organization offers online report cards and has published rankings in publications such as US News & World Report. The organization has not incorporated ISO standards, but Pawlson believes they can be complementary to what NCQA does.
As more organizations adopt higher standards and even public-performance measures, laboratories may be introduced to processes they have not previously had to complete, and they will need to learn what those are and how to perform them. But ultimately, the benefit is higher quality and better care, improving patient safety and the bottom line. Whichever QMS laboratories employ, it is important that they have one. Notes Stein, “It isn’t very helpful to have buzzwords to slap on your process—it’s better to have a good process.”
Renee DiIulio is a contributing writer for Clinical Lab Products.
|
“We are seeing customers interested in having someone come into the lab and define a program to achieve quality objectives for efficiency, compliance, and risk management,” says George Rodrigues, PhD, ARTEL’s senior scientific manager. He notes that market leaders are moving toward higher quality standards. Currently, quality standards guiding liquid delivery are best captured by the College of American Pathologists (CAP of Northfield, Ill) checklists. Rodrigues found more than 100 CAP documents related to pipettes, but the guidelines leave a lot of latitude for laboratories to determine the frequency with which pipettes are checked. International Standards Organization (ISO) 15189 also addresses liquid delivery, and Rodrigues notes that the document is expected to have a greater impact on US controls. But the ISO guidelines avoid specifics as well, saying calibration should be done frequently enough to maintain a state of readiness that achieves quality objectives, according to Rodrigues. “There are two pieces of the equation in evaluating liquid-delivery risk. One, is my equipment functioning properly or is it outside of my tolerances? And two, what are the consequences if my equipment is functioning incorrectly?” Rodrigues says. Rodrigues uses PSA and coumadin as examples. “The consequences of a bad PSA analysis can require invasive follow-up, such as a biopsy, and coumadin needs to be closely controlled,” he says. So it’s important to identify the critical liquid-handling steps and control them with a high degree of reliability. Factors that influence accuracy include operator variability and pipette quality. “There are a few manufacturers that make professional-grade instruments and a number of manufacturers that make less expensive pipettes that don’t hold up in the long run. But this factor may not be considered by some laboratory quality systems,” Rodrigues says. Operator use accounts for roughly 60% of related errors, Rodrigues says. “We also know—from a study where they went into biomedical research centers and hospital labs to test the pipettes—that in a typical lab, up to 30% of pipettes can be performing outside of specifications,” he says. Laboratories can respond with proper training, regular maintenance, and calibration checks. “One of the most cost-effective things a lab manager can do is train the pipette operators and qualify them with a proficiency test,” Rodrigues says. He notes that drawing up too much too quickly or using too much tip-installation force are common errors. Pipettes should be stored properly (not left to bang around in a drawer), and if they appear dirty, they should be cleaned. “If there has been liquid pulled up inside the pipette, it can dry and create a crystallized residue, which results in pipette failure,” Rodrigues says. Calibration can be accomplished with an outside service (either coming in or by sending the pipettes out); other facilities do it in-house. Rodrigues cautions that gravimetry—weighing pipettes on a balance—works for larger volumes, but not for smaller ones due to evaporation and vibration issues. For these, photometric measurement, where a colored solution is read with a spectrometer, works better. ARTEL uses rRadiometric Photometry™, which uses two dyes to control for sources of error. Rodrigues claims this process is five to ten times more accurate than single-dye photometry. An anonymous quote notes that, “To err is human, to really foul things up requires a computer.” Though automation has generally served to reduce errors in the lab, it too requires quality management. Rodrigues suggests that laboratories can improve the accuracy of robotic liquid delivery by optimizing the factory default settings. “We did an experiment the other day with a robot. The factory default settings had a 13% CV. After optimization, including aspiration rates, air gaps, and other parameters, the CV was brought down to 0.5%,” Rodrigues says. According to Rodrigues, some labs are attuned to these capabilities and others are not. Some rely solely on the CAP checklists and proficiency testing to meet a minimal compliance point. Rodrigues suggests that labs need to go beyond mere compliance to ensure quality in care. —RD |
Good to the Last Drop 
Consequences vary depending on how critical the analysis is. “For some tests, it’s important that the quantitation be accurate because the treatment or follow-up may involve significant patient pain or cost,” Rodrigues says. In cases where the difference between the normal population and the disease population is great, accuracy may be less of a factor. But in instances where the window of control is narrow, inaccurate results can be catastrophic. 
