Interviews

An interview with Dr. Jan Krouwer, PhD, an expert on Risk Management, FMEA, and many other topics.

This month we are privileged to talk with Dr. Jan Krouwer, PhD, FACB, an expert on Risk Management. Dr. Krouwer has over 20 years experience in the medical diagnostics industry. For most of this time, he managed an R&D consulting group that focused on evaluation protocols, reliability improvement methods, statistical tools, and software development for data analysis. He is now an independent consultant with Krouwer Consulting in Sherborn, MA.

Dr. Krouwer has been at the forefront of critical discussions in laboratory quality. He is the author of Assay Development and Evaluation: A Manufacturer's Perspective and Managing Risk in Hospitals, both from the AACC press, as well as the creator of an Integrated Fault Tree FMECA software program. Recently, he served as the AACC's Expert Access for “Equivalent QC: Some myths about internal monitoring systems.” Finally, he writes his own online essays at http://www.krouwerconsulting.com

Q: What is the short definition of Risk Management?

Dr. Krouwer: Risk management is a process of identifying and providing mitigations to reduce the likelihood of problems.

Q: Why is Risk Management a good thing?

Dr. Krouwer: In the lab, many quality activities don’t account for all possible problems. For example, quality control doesn’t address patient sample mix-up problems. Risk management is good because it is all encompassing. Done correctly, all potential problems are exposed.

Q: If we aren't doing Risk Management in the lab right now – what are we doing? How do we characterize what it going on now vs. what is going to happen as Risk Management is embraced?

Dr. Krouwer: Risk management is being done in the lab now – albeit often on an informal basis and often not according to risk management standards. As more people are exposed to what is being done in other industries, the quality of risk management in the lab will improve.

Q: Based on recent industry news and scientific press, it looks like we are heading into a new era of Risk Management for the lab. What impact do you think that will have on the way a laboratory works?

Dr. Krouwer: This remains to be seen. Quality programs (such risk management) have a history of being embraced, and later set aside. Often, a useful element remains with the organization.

Q: Where does Risk Management fit in a world of ISO, Lean, Six Sigma, Patient Safety, EQC, etc? If one has to choose between one management technique, is Risk Management the most important?

Dr. Krouwer: ISO provides standards. I have avoided the buzzword “Lean” and am not sure what this means. Six Sigma is a collection of tools which includes risk management tools. Patient Safety is a movement, which uses risk management as one tool. EQC is a specific proposal related to quality control.

Risk management can be considered the most important tool because as said previously, it encompasses all possible problems within a lab. Besides tools, organizational culture is important.

Q: How do you envision the practical steps of implementing Risk Management in the lab?

Dr. Krouwer: This cannot be answered quickly. The first question is what are the lab’s quality goals. Unfortunately, many quality goals discuss getting 95% of the results right. For a lab that reports a million tests, that could mean that the lab could meet its quality goals with 50,000 wrong results a year. Compare this to the Veteran’s Administration allowed frequency for the most severe errors, which is once in 5 to 30 years. A recent paper showed that many errors are present in labs: (Nevalainen D, Berte L, Kraft C, Leigh E, Pacaso L, Morgan T. Evaluating laboratory performance on quality indicators with the six sigma scale. Arch Pathol Lab Med 2000;124:516–519). So a lot of work has to be done on goals. Once this is complete, tools can be selected and training provided to implement tools.

Q: Will laboratories need to do Risk Management on a test by test basis, or instrument by instrument, department by department, etc?

Dr. Krouwer: The answer is all of the above. Labs will also have to go beyond their own department. In the University of Washington, a woman was unnecessarily treated for cancer due to incorrect hCG results. The treatments included chemotherapy and two surgeries, based on 45 hCG results run during the treatment course and all hCG results were incorrect. Finally, the correctness of the assay was questioned. Better communication between the lab and clinicians is a risk mitigation step that is (partly) outside of the lab.

Q: What's the typical time frame for a Risk Management cycle on a generic laboratory process (if a typical time frame exists)? How many people need to be involved – and from what parts of the laboratory?

Dr. Krouwer: Risk mitigation should start when a process is designed and end when the process has been retired from use. At design, potential problems are identified and risks reduced through mitigations. When the process is launched, actual problems are observed and corrective actions are implemented to prevent recurrence. The people who should be involved are the people involved in the process plus a quality person to facilitate the effort.

Q: Will Risk Management require external consultants? Does everyone need to be trained (or “belted”) in Risk Management?

Dr. Krouwer: The main qualifications for someone to lead a risk management program are a keen interest in analyzing data and the ability to work with people, especially as a facilitator. This could perhaps be the person responsible for QC in the lab, with some additional training. The other people in the lab would require some training – for example it is important that errors be reported and in some cases this will mean human errors. This also means that the right culture and policies must be present in the lab. Whether people will be certified and how useful this would be remains to be seen.

Q: Will Risk Management be certified? How will the laboratory know if they've adequately completed the Risk Management cycle ?

Dr. Krouwer: This is an important question. Although certification / accreditation is important, preventable medical errors do not occur in unaccredited hospitals. This implies that the accreditation or certification process leaves something to be desired. When I worked in industry, our most successful quality programs were initiated from within and not part of regulatory requirements, which were extensive (FDA inspections, ISO 9001 certification, assay submissions to FDA). The challenge is for the regulatory people to make requirements more like self initiated quality programs with measurable goals. Otherwise, people will do the minimum required to pass inspections. A risk is that the limited funding available will be used to satisfy regulatory requirements rather than to support self initiated quality programs.

Q: It seems like that Risk Management failures occur, but the reason those failures happen is that the users didn't do their Risk Management correctly? Can it be said that "Risk Management doesn't fail – People fail to do Risk Management properly" ?

Dr. Krouwer: This is correct. Risk management is a generic term and can mean anything from almost no activity to an extensive and effective program. How this plays out remains to be seen. There is an ISO standard for risk management for medical devices (14971). This is a so called “horizontal standard” meaning that principles are explained but unlike a “vertical standard” no detailed specific procedures to follow are recommended such as those in a CLSI standard such as EP9 (how to perform a method comparison to estimate bias). Horizontal standards are often called “flexible” with the appeal that one can pretty much do anything and say that one is conforming to the standard. So vertical standards would be useful for lab risk management.

Q: Will the choice of scale for frequency, severity, etc. effect the final evaluation of the risks? What is the ideal scale? Is there a “realistic” scale that should be adopted and standardized for healthcare?

Dr. Krouwer: These scales are important and relate to goals. There is not enough money to either fix all observed problems or to reduce the risk of all potential problems and so scales are an important part of populating a Pareto (means of ranking importance). The Veterans Administration scale could be a good starting point. A problem for labs is that lab test result errors do not directly harm patients. By the way, similar questions have been avoided by standards organizations. For example, two groups (CLSI EP20 and ISO 15196) gave up on the task of proposing medical error limits for lab assays. However, the new proposed waiver guidance from the FDA requires that one count error rates in each zone of a Clarke type glucose error grid (for the assay under consideration). This guidance will get things moving about scales.

We thank Dr. Krouwer for taking the time to answer these questions.