No one who works in a clinical laboratory wants to make mistakes.
Yet a system in force throughout the field tells us that it is all right
to slip up occasionally. That system, ironically, is quality control.
Current quality control practices hinder attainment of total
quality or error-free performance. Allowing a tolerance of, say, 1 per
cent of results outside 2 standard deviations is tantamount to accepting
1 per cent error.
The real goal of quality control should be no error at all. The
concept is called zero defect is japan. It was picked up in recent
years by U.S. industrial leaders, who call it error-free work.
Corning Glass Works, MetPath’s parent company, adopted the
approach in October 1983. It has since invested $2 million on
educational programs for management and employees, including
establishment of the Corning Quality Institute. What we learned at the
institute helped us pursue error-free work in our reference laboratory
Management by prevention is an essential part of the process. That
means examining the entire operation to identify problem areas and get
at the root causes to stop errors from occurring. We consider it
crucial, for example, that our 600 couriers not miss a single pickup in
bringing us thousands of specimens a day. If a pickup is missed, the
inclination may be to blame the courier involved. But it is more useful
to find out how the mistake came about. Then we should do whatever is
needed to help the courier avoid the problem in the future.
There are no quick fixes. Management by prevention encompasses a
variety of activities, including training, defining clients’
requirements and determining how to meet them, and having measurements
that tell the lab how it is doing. It is an unending process rather
than a program that comes into being with an inital burst of enthusiasm
and eventually fades out.
A refusal to stand for any mistakes may sound idealistic. Everyone
is human and makes mistakes, I am told when discussing quality at
medical schools. My response is how many dropped babies would you
accept in the hospital nursery? One in 1,000? One in 10,000? No one
ever gives a number because the goal is always zero dropped babies.
We apply the same high standard in our personal lives. How many
times do we accept a bank error in our checking accounts? Or the wrong
merchandise delivered to our homes? In these and many other instances,
the tolerance is zero, not 1 or 5 per cent.
As the last examples suggest, quality is among other things a
financial issue. A failure to manage quality can cost most of us in the
health care field anywhere from 20 to 30 per cent of our resources.
There is a cost of doing work wrong and a cost of doing it right.
The first cost always exceeds the second. Besides incurring the expense
of repeat tests, a reference laboratory can lose clients because of
inadequate test quality, missed pickups, or poor turnaround time.
Hospital laboratories have had captive clients in the past, but
that’s changing as a result of greater competition among hospitals
and from alternative health-care delivery systems. Each hospial will
have to pay greater attention to meeting patients’
requirements–from maintaining a reputation for good care and good
results to rating well in terms of food preparation, cleanliness,
relative quite, and other aspects of service.
To implement the error-free process, our laboratory organization
set up its own quality school and is forming quality improvement tems
(QITs) in each functional area. Everyone, at every level, is involved.
All senior officers participate in a corporate QIT (Figure I). We meet
every three or four weeks for a full day, bginning at 8 a.m.
First we listed administrative questions that the corporate QIT
team should answer, namely: Are we giving our employees specific skills
and knowledge to do quality work? What more should we do? How much of a
commitment will we make to this process? How do we encourage employees
to think about quality?
Next we worked up a list of problems, giving each a priority based
on the dollar impact that a solution would have. We assigned
responsibilities and time limits for dealing with the problems. Resource
needs were reviewed, and we kept tubs on progress toward solutions.
Plans called for us to extend the process to the entire company once we
achieved some dollar successes. We appointed a senior administrator to
the full-time position of coordinating the quality process. He reports
to the laboratory’s president.
To involve all employees from the start, we arranged quarterly
meetings, showed videotapes on the error-free process, and formed 14
QITs. The first QITs were formed in hematology. At three-day seminars,
members learned a common language and the principles of the new approach
to quality, which makes for quicker and more effective communication
with colleagues. QITs in turn create corrective action teams to solve
Any hospital or reference laboratory interested in adopting the
total quality approach should consider several requirements:
* Commitment. Obviously, a drive to eliminate all error demands
dedication. Corning’s substantial cash outlays for education
exemplify this. Our lab’s program has strong administrative
support, too. In a hospital, a single department such as the laboratory
can start an error-free process, but it is better if the concept is
applied throughout the institution.
* Education. By investing sufficiently in training employees, the
organization can build a staff that really understands the quality
process. Throughout the Corning organization, employees are trained to
help them apply quality principles in their everyday work.
* Teamwork. Individuals cannot bring about error-free work on
their own, even though personal dedication and effort are the bedrock of
quality performance. Teams work better because many qualtiy-related
problems are complex and span the entire organization. Teams can also
prioritize problems based on dollar impact.
Although everyone in the organization should be involved in the
error-free process, someone has to be in charge. There should be head
of quality and high-level management control of quality.
* Cost analysis. Proper mangement of quality makes strong
contributions to profit and cash flow. Discuss quality in Financial
terms–how much each problem is costing the organization–so that
administrators and financial officers can better understand what you are
trying to accomplish. Then they are more likely to provide the needed
At the hospital level, a 60 per cent occupancy rate compared to a
neighboring institution’s 90 per cent is certainly a dollar issue,
and it may indicate serious quality problems. Improper billing
procedures may also keep a hospital from maximizing its reimbursement.
Again, errors are costing the institution money.
* Communication. When the laboratory receives what seems to be an
unreasonable test request, it usually reflects a problem the clinician has and wants help with. Say the request is for an esoteric, difficult,
and unreliable test on spinal fluid. I have found in such instances
that what the clinician really requires is assistance in confirming or
ruling out multiple sclerosis, not an esoteric test result.
When the laboratory understands the requirement, the problem shifts
to helping the clinician choose the best workup. If this takes place at
the outset, you have the low cost of doing things right. There are no
charges for unnecessary and expensive testing.
Another aspect of communication: Employees should see measurements
of their error rates and performance. For example, we post a graphic
overview of the number of QNS specimens in each laboratory. Each
department should have a public record of its performance. Employees who
receive clear feedback on how well they are doing are motivated by pride
to work on problems.
Let’s turn now to a small problem that Dr. Michael Rosen,
manager of the RIA lab, and I addressed experimentally after we came
back from the Corning Quality Institute. It illustrates how error-free
quality control works.
Customers complained that we were reporting 24-hour urine test
results without indicating urine volume. Th requirement was to have all
such reports leave the laboratory with the volume, so we knew we had a
We found a “pulse point” for the problem in general
chemistry. Our computer-generated reports provided a clear measure of
how many urine volume figures were missing. Eighty per cent of our
24-hour urine tests had no volume listed on the worksheet.
Many clients submitted 24-hour urine bottles without measuring
volume because they expected us to do it. But not all employees who
handled specimen accession had been instructed to measure and record
24-hour urine volumes. Throughout the test process, no one questioned
or returned requisitions that lacked volume figures.
We were still so green at the error-free process that we thought we
could solve the problem by explaining the correct procedure to
accessioning employees. The next time we looked at the general
chemistry worksheets, volume figures were missing from 50 per cent of
reports–down from 80 per cent, but still far from zero. We had made
the mistake of thinking that there was one cause for the entire problem.
Digging deeper, we discovered that after measuring volumes,
employees recorded them on the bottles but not always on the
requisitions. Occasionally they also separated requisitions from
corresponding bottles. They assumed that technologists performing the
tests would copy volumes from bottle lablels onto worksheets.
We corrected this by requiring specimen handling to mark volumes on
bottles and requisitions, and to keep the two together. When
requisitions went to the keypunch office, operators now transferred
volume measurements to worksheets.
But we had succumbed to a quick fix again, because missed urine
volumes declined only slghtly to 40 per cent. The system was defective
somewhere else. Apparently, we would have to become more involved in
the overall test process and invest more time than we first thought.
If volumes were entered on requisitions but did not appear on
worksheets, they must be falling through cracks in the keypunch area.
We placed cards in front of each computer console reminding operators to
punch in volumes that appeared on requisitions. To monitor its own
performance, the keypunch group kept a nightly score of volumes that
didn’t make it from requisitions to worksheets.
Keypunch errors abated. About 80 per cent of urine volumes were
reported. This was good, but it wasn’t error-free. Could computer
programming be faulty?
The answer was yes. Keypunched volumes didn’t always appear
on worksheets. The data processing department discovered that some
operator didn’t know the computer demands four-digit entries. A
950 ml specimen had to be entered as “0950 ml.” All
measurements were now entered in this manner.
Data processing also found a programming glitch. When requisitions
included creatinine clearance indices, the computer entered the
keypunched 24-hour urine volumes into the calculation but did not post
them separately on the worksheet.
In a few weeks, programmers eliminated this deficiency. Now 95 per
cent of reports listed urine volumes. We could assert that every
specimen properly submitted for 24-hour urine study showed the volume on
Five per cent of reports were still deficient, we realized, because
clients provided unacceptable specimens. These included random urine
collections for 24-hour creatinine clearance testing or 24-hour
specimens submitted in improper containers, not graduated for
It would have been natural to push responsibility for these lapse
onto clients and consider our error-free effort complete.
But the delinquent 5 per cent meant we failed to show our clients
how to use the laboratory’s services properly. It was incumbent on
us to do a better job.
We compiled a list of all accounts that didn’t follow proper
sampling protocols. Then we instructed them, via letter or during
service calls, on how to collect specimens correctly.
It took us three to four months to reach the last stage, but
we’re not going to stop the process until every client receives a
24-hour urine volume report when it is required. That is the error-free