Calibration Is Always Subject To Variation
Kudos to Philip Stein for having the intestinal fortitude to make the king-is-naked statement in "Avoiding Calibration Overkill" (September 1999, p. 74). Stein correctly points out what common sense dictates: Have the calibration accuracy "accurate enough for the application, but no more accurate."
I would also like to extend this common sense train of thought to calibration frequency. I always advocate that recalibration should only be done when there are data that suggest the process has gone out of control. The opposite of what I advocate would be to blindly follow a calibration schedule regardless of the condition of the process.
My position rests on the belief that the act of calibration is subject to variation, depending upon the work process involved. This point is particularly relevant in the analytical chemistry world where most calibration protocol requires the user to prepare the calibration standards. Besides wasting effort, the net effect of unnecessary recalibration is to inflate the whole process' variability (Deming's funnel experiment). Apparently this is an issue most auditors and standards writers are unaware of because they continue to insist on a calibration schedule.
ALEX T.C. LAU
Toronto, Ontario Canada
Create an Effective Customer Survey
William Robert Loomis' article "QS-9000 Customer Satisfaction Monitoring Isn't Working" (July 1999, p. 54) leads one to believe that regular customer surveys are of little or no value to small and midsize manufacturing businesses. Loomis describes instances where the surveys were poorly designed and the results were not used. Any tool that is poorly designed and not used will have poor results. Rather than throwing out the customer survey as a tool, the proper alternative is to design a good survey and build in a process for using the results.
The key steps to an effective customer survey involve:
* Asking key decision makers in the company to help construct the survey.
* Using more than just numerical ratings, for example, by following up on low ratings and asking open-ended questions.
* Building steps that lead to a response to the results into the process.
A good customer survey can be an important strategic tool that defines what a company must do to reduce the gap between strategic goals and current reality. The results can identify the processes that need to be improved, provide detail about how those processes should be improved and identify the customers in need of special attention. The lack of follow-up Loomis describes is not due to a problem with customer surveys--it is due to a problem with management. My experience in working with more than 40 manufacturing companies leads me to conclude that customer surveys are one of the best tools available to stimulate discussions that lead to process and product improvements.
State College, PA
ASQ Should Stop Perpetuating Fads
I've been a member of ASQ for several years. I also have four certifications which make me a pretty marketable person. However, I expect more from my professional society. ASQ, boasting of its membership and authority on the quality movement in this country, should act as a leader. Instead, ASQ finds parades that are already in progress and marches in front of them. The latest of these parades or fads is Six Sigma. Before that it was audits and before that it was teams. ASQ should stop perpetuating these fads even though that may cause it to go against popular opinion. A leader should have profound knowledge and do what's right, not what's popular.
In the article "Cowboy Quality" (October 1999, p. 27), Miles Maguire made statements that will outrage any self-respecting Deming disciple. It was ignorant of Maguire to compare Mikel Harry to Shewhart and his disciple, Deming. This wasn't the first time I heard about Harry, so I can say he is definitely no Shewhart or Deming. His ideas are just a recital of the old management beliefs that dug the pit we're in today.
Harry says that we can't improve what we don't measure. That statement is false. Lloyd Nelson, a well- known statistician and friend of Deming, says that the most important numbers are unknown or unknowable. What you can measure is trivial compared to what you can't measure. If Harry thinks that his client's measurements are accurate and complete, he is wrong. But how could he know?
The article also talks about how 40% of General Electric executives' bonuses are tied to Six Sigma. Rewards that are contingent upon outcome kill intrinsic motivation. Rewards only encourage people to expect more rewards and to achieve them by any means possible. But how could GE's Jack Welch know? Harry laments that the quality initiatives and metrics don't link hierarchically at this time. The truth is that this problem has existed as long as we've had quality. The metrics aren't additive. This fact isn't understandable without appreciation for systems. But how could he know?
Harry has a numerical goal of 70% annualized for the improvement of performance measures, such as defects or cycle times. Numerical goals accomplish nothing. They only encourage people to manipulate numbers or increase the costs in an attempt to meet them. But how could Harry know?
Harry thinks that total quality management is worthless because today's stockholders can't even wait for 30 days. If Harry is the kind of leader Maguire describes, Harry should be working toward making stockholders understand the need for a consistency of purpose. Harry--a perpetuator of short-term thinking, which is the most important cause of our decline--is not a leader. Deming's road to quality focuses on the fact that 93% of costs are caused by management's action or inaction. Harry's approach only focuses on 7% of the costs, such as defect rates and cycle times. But how could Maguire know?
Should Clinical Labs Comply With ISO 9000?
In "Quality System Harmonization for Clinical Laboratories" (September 1999, p 57), William I. White advocates using ISO 9000 to tie the pieces of clinical laboratory compliance together. There are requirements from the Joint Commission on Accreditation of Healthcare Organizations, the Commission on Office Laboratory Accreditation, the laboratory accreditation program of the College of American Pathologists (CAP), and state governments, as well as a few White omitted such as the American Association of Blood Banks and the Office of the Inspector General.
I completely agree with White's assertion that these requirements must be harmonized. He states that "the ISO 9001 standard's emphasis on management responsibility and a quality system would guide laboratory management to focus first on the overall system of how the laboratory operates, and then take care of the details as they fit in." Because of the huge regulatory burden, it is easy for laboratorians to become focused primarily on compliance with standards. It's apparent that laboratories need a framework that will make all of the rules understandable. Is compliance with ISO 9000 the way for clinical laboratories to go?
I submit that it's the compliance mentality that led to the regulatory burden that the laboratory faces. Laboratory managers who believe that their laboratory meets standards and, therefore, produces quality test results perpetuate the above mentality. For example, for many years, the industry widely supported voluntary standards from CAP. Most hospital based labs participated in the program and were deemed good by CAP. However, quality was sorely lacking as evidenced by the patient, provider and payer dissatisfaction that led to intense federal and state regulation.
Compliance is becoming progressively more difficult to manage. For example, laboratories that receive Medicare reimbursement now have to hire a compliance officer. It's the officer's responsibility to ensure that the government is not overcharged for services. The Office of the Inspector General can audit a laboratory at any time, with potentially devastating results. Because of problems uncovered as a result of scrutiny, more federal regulation is sure to follow. The increased regulation will lead to more scrutiny, which will lead to even more regulation.
In my opinion, the only way for our industry to break this vicious cycle is for each lab to:
1. Clearly understand who its customers and stakeholders are.
2. Clearly understand what they need and expect, now and in the future.
3. Design and redesign the systems that will consistently deliver on those expectations.
I appreciate the value of standards. However, why shouldn't we primarily promote customer and stakeholder satisfaction and prevent the need for the regulation resulting from dissatisfaction? I propose a systematic, though gradual, break from the compliance mentality. White proposes a unifying compliance system, which, I admit, is better than what we have now.
The compliance minded manager would say, "Our laboratory is registered to the ISO standard and, therefore, produces quality laboratory results." But ISO 9000 compliance alone will not bring harmony to the regulatory quagmire in the clinical laboratory. Our unifying principle must become the understanding of, caring for and feeding of the patients, health care providers and even the federal payers that we serve.
HARVEY L. VANDENBURG
certified quality manager
Harvey Vandenburg's comments are very insightful, and I agree with the spirit of his suggestions. At the same time, I think there is also value in the creation of a formal quality system for any organization.
It is easy to get lost in the 20 elements of the ISO 9001 standard, and I have seen a four part informal version that helps put everything in perspective. It bears a striking resemblance to some of Vandenburg's thoughts:
1. We must give customers what they are entitled to expect.
2. Our internal processes must be organized so they accomplish objective number one.
3. We must insist that our suppliers give us what we need to accomplish objectives one and two.
4. We must check up on ourselves to make sure the first three objectives are being met.
The value of objective number four can be quickly forgotten among the pressures of everyday work if there is no formal quality system in place. A laboratory that means well can go astray if there is no checking up action. The hard part is guiding the auditing and management review so that it doesn't get bogged down in compliance detail. The processes Vandenburg discusses will be effective in creating the improvements he wants to see.
WILLIAM I. WHITE
Quality Equals Control And Eliminates Excitement
I think Thomas Pyzdek ("Quality Profession Must Learn To Heed Its Own Advice," June 1999, p. 60) is absolutely right. The more control there is, the fewer new and exciting opportunities there are. I am a certified quality management student, and it appears to me that quality equals total control and eliminates the possibility of doing things differently from the way it's written in the manual.
All quality practitioners focus on is getting things done in the best way, and all that has done is take us back. Total control never got anyone anywhere new and exciting.
CHERRIE ANN JOSEPH
Joseph Maxwell Gunstan Enterprises
San Juan, Trinidad and Tobago
How Can One Reverse The Pro-Quality Facade?
just finished reading the article by March Laree Jacques ("The Call of Quality: Doing Right Things Right," September 1999, p. 48). In the article, she quotes Robert Peach who states, "The frustrated quality manager who is in a company paying lip service to quality can, tactfully but very methodically, use the mechanism of ISO 9000 to change that. And, he should."
Our quality manager is in that exact position. What steps should he use to reverse the pro-quality facade?
Name withheld at request of letter writer
Robert Peach's Response:
When company management is paying lip service to quality efforts, adoption of ISO 9000 can be an agent of change. I agree with those who emphasize that top management needs to be in support of any companywide effort for it to succeed-- including effective quality management. The situation of the letter writer's organization is all too common. The organization is obligated, by its customers and market, to adopt ISO 9000, but management has really not bought in to the value. In some cases, frustrated quality managers may visit the facilities of noncompetitors so their management can see the value of effective ISO 9000 application.
In response to the letter writer's situation, I think ISO 9000 has at least two built-in characteristics that, if applied effectively, can get the quality job done even without the full support of management. These characteristics may even turn the most nonsupportive managers into ISO 9000 converts.
The first built-in characteristic is that ISO 9001 contains specific requirements that leave no alternative but to establish the capability to produce a high quality product. They include the following:
* Quality requirements are defined and documented. (4.3.2a)
* Product designs assure that all quality requirements will be met. (4.4.1)
* Processes are controlled. (4.9)
* All quality requirements must be met before release. (4.10.4)
* Only products that have met inspection and test requirements can be released. (4.12)
* Corrective and preventive action procedures need to be in force. (4.14.1)
* Quality records must show that the product conforms to specifications. (4.16)
* Internal audits need to determine the effectiveness of the quality system. (4.17)
If this impressive list is not enough, the second characteristic is that ISO 9001 also requires that a quality policy be defined and documented. The quality policy must be understood, implemented and maintained at all levels of the organization. This clause in particular is relevant to the letter writer's problem. A quality policy should include a statement indicating the intent to produce a product that meets customer requirements. Lip service to quality management can be interpreted as unsupportive of such a statement.
Registration by an accredited registrar is often an important element in the ISO 9000 process. A skilled auditor will probe to establish whether each element of ISO 9001 is being met. The entire organization will know when an established quality practice is being undercut by top management through a lack of support. The alert auditor will quickly pick up this fact through interviews of line personnel while determining the degree of understanding of quality policy. If it is extreme enough, such a situation can and should be the basis for the registrar's refusing to issue a registration certificate until meaningful management support is revealed.
Incidentally, elements in ISO 9001:1994 are stronger and clearer in the ISO 9000:2000 revision. In the revision there is an emphasis on management responsibility, continuous improvement and demonstration that the quality being produced consistently meets customer requirements.
'Quality Progress' News Isn't Very Timely
It was with a wry smile that I read about the amendments to the Fastener Quality Act in the September 1999 issue of Quality Progress ("Keeping Current," p. 25). The amendments were passed by the U.S. Congress way back in June, and this just serves to illustrate how far from on the ball you guys are.
Your coverage of this important piece of quality specific legislation was woefully inadequate. As a quality professional in the fastener industry in Canada, I am afraid I looked to sources other than Quality Progress for information and guidance on this American law that will affect my day-to-day activities.
What good is it to read about something as important as the enactment of a new U.S. quality related law three months after it was passed? I think that in terms of news value and as a service to ASQ members whose professional lives are directly impacted, these amendments should merit more current coverage.
FRANK M. CONNOLLY
What Is a Negative
Lower Control Limit?
In your August 1999 issue, the article "Total Quality Management in Higher Education" (Carl B. Montano and Glenn H. Utter, p. 52) contains three control charts (Figures 3, 4 and 5) that don't make sense. What kind of control charts are these, and how were the control limits calculated? In particular, what is the meaning of a negative lower control limit in all three figures?
DONALD S. ERMER
University of Wisconsin-Madison
In the article "Why Should Statisticians Pay Attention to Six Sigma?" (September 1999, Ronald D. Snee, p. 100) the MAIC (measure, analyze, improve and control) process incorrectly appeared twice as MIAC.