How an Iowa lab slashed its Pap test TAT
Melanie Fridl Ross
Moving standard industrial engineering and organizational development tools from the line to the lab could be just what the doctor ordered when it comes to processing Pap tests more quicklywithout sacrificing accuracy.
That’s what laboratorians at the University of Iowa Health Care’s hospital and clinics discovered after a patient satisfaction survey revealed that women thought they had to wait too long for Pap test results. The criticisms spurred laboratory administrators to scrutinize the screening process and identify areas that needed to be improved, a process that began in 1999. The approach worked:The mean turnaround time has been slashed from nine days to 3.4 days.
"Most of us think Pap smear collection, analysis, and reporting is a very straightforward process," says Sue Zaleski, MA, SCT (ASCP)HT, operations coordinator in anatomic pathology at the University of Iowa Health Care. "We discovered it is much more complex and lengthy than any of us expected."
The University of Iowa study coincided with publication of a CAP Q-Probes study, which offered a comparative database for gynecologic cytology turnaround time. The mean turnaround of the median laboratory in the Q-Probes study was 4.4 days, while the top 10 percent of the labs surveyed reported 90 percent of their results in three or fewer days. (TAT was defined as time from specimen receipt in the laboratory to final report signoff.)
"I said, ’This is great! This is fantastic!’" says Tom Persoon, MS, a clinical laboratory scientist in the University of Iowa’s Department of Pathology and an instructor in industrial engineering at Iowa who was enlisted to analyze the Pap test process in the hospital and clinics. "In addition to having an internal benchmark of what we used to do and what we’re doing now, we had the opportunity to compare our results with an external benchmark.
"We took that data, puzzled over it awhile," he says. "We identified a couple benchmarks we could show people and say, ’Look, this is what world-class organizations do. Where do you want to be in this spectrum of Pap smear turnaround time?’"
Persoon interviewed staff members to identify every step in the screening process. He talked to the physicians who collect the specimens and the personnel who transport them. He interviewed the cytotechnologists who examine the cells and the secretaries who sort the reports.
From his findings, Persoon constructed what is known in industrial engineering circles as an IDEF3 process model, a tool similar to a flow diagram. The model described 72 separate activities (units of behavior in IDEF3 terminology) that laboratory staff and other personnel engaged in. Time was a factor in 24 of these. By scrutinizing the two dozen activities that had time controls, Persoon was able to pinpoint problem areas.
So employees could visualize their place in the process, Persoon unfurled a 14-foot-long scroll with the IDEF3 model displayed on it. "Lots of people knew what their little parts were," he says. "No one had any idea what things looked like outside the four walls they did their job in. One of the key elements in this whole process was the ability for us to view this as a system and to engage in systems thinking, not look at it as one piece of actionwhat the laboratory does."
Persoon also shared data he gathered by measuring how long it took laboratory personnel to complete a particular step or group of steps. "My measurements were done unobtrusively, because people who know their work is being watched or measured will change their behaviors," he says.
Persoon conducted measurements in two ways. For part of the process, he says, personnel who performed one step in the system but were not "major stakeholders" attached a coded tag to samples as they handled them. "I later intercepted those samples and recorded time data based on the coded tags," he says. For other parts of the process, Persoon extracted from the lab’s information system the times that certain activities occurred.
In all, he studied processing times for 377 Pap tests. "I was interested in measuring the process time-the service quality. The diagnosisthe analytical qualitywas not an issue," he says.
Persoon quickly spotted a sizable work-in-progress pile that was a key culprit in slowing turnaround time. At one point, more than 400 Pap tests awaited screening.
"The work was piling up in the cytotechnologists’ queue," he says. "The slides would get stained and put into trays, waiting for the cytotechnologists to examine them. That was where the largest delay was."
Another holdup: the schedule for printing paper reports. Persoon discovered that final reports were printed in a batch at 7PM but not distributed to physicians’ mailboxes until early the next afternoonalmost 30 hours after the results became available. The physicians’ offices accounted for additional delays in reporting Pap test results to patients, the study found.
The next step was to determine why there was such a huge work-in-progress pile.
"We knew that people weren’t slacking off," Persoon says. "It wasn’t a situation of somebody not doing their job. Everybody was working really, really hard."
Periodic staff shortages were partly to blame, such as the time a cytotechnologist went on bereavement leave. "They needed to plan for the fact that such situations were going to occur," Persoon says. "That was an important concept for this group."
Persoon and Zaleski made sure staff learned to recognize normal productivity so they could compensate when needed.
"When people become aware that an event is going to affect the rate at which work is done, they can say, ’Okay, what adjustments need to be made in the system to account for it?’" Persoon says. "People need to be aware it matters. They need to know what their productivity is when everything’s working okay, and then when they run into a problem, they can sit down, put their heads together, and decide what kind of changes need to be made in the way they operate to maintain that productivity."
So Persoon employed an organizational development tool known as an action research model. In this model, people are asked to scrutinize their own actions and, if necessary, change their behavior to ensure a more desirable outcome. But first Persoon had to overcome an obstacle: Cytotechnologists feared that improving productivity might harm accuracy.
"Their first impression was: ’This means I have to work faster and harder to get this out. It means I can’t spend as much time screening this slide because I’ve got so many more to read,’" Persoon says. "We had to point out that how much time they spent reading an individual slide ought not to change at all in this process. We wanted to improve the service quality, but we would not do that at the expense of result quality."
One of the steps the cytotechnologists took was to shift some of their quality control data-gathering activities to other laboratory personnel. That freed the technologists to use their technical expertise to perform microscopic examinations.
They also began to monitor turnaround time regularly. Now, when the work in progress begins to grow, they decide as a group to redirect efforts, triaging nongynecologic specimens to the pathologists and prioritizing Pap tests, Zaleski says. Only on rare occasions have the technologists had to work overtime.
The study also found that printing reports on demand in the clinics might further reduce turnaround time.
"While the cytology laboratory, at some levels, has taken a lot of heat for turnaround time, I think these data show there are multiple points in that complex 72-step sequence where you can have delays," says Michael B. Cohen, MD, professor and head of pathology and medical director of cytology at University of Iowa Health Care. "Clinicians aren’t immune, either. They could take a while to read the report, dictate a letter, or make the phone call. Like all complex problems, there wasn’t a single thing you could focus on. But with that knowledge in hand, you can work in a prospective manner and make adjustments."
More complicated than it might appear, Pap test turnaround time has attracted its share of debate.
"This issue of turnaround time comes up periodically, but we’ve got to be careful because it’s a double-edged sword," says Dina R. Mody, MD, professor of pathology at Baylor College of Medicine, Houston, and vice chair of the CAP Cytopathology Committee. "You don’t want to rush cytotechnologists. There are major lawsuits during screening already because labs supposedly missed this needle in a haystack. And trying to get cytotechnologists to screen faster to meet a turnaround time target could be asking for more trouble." Accuracy trumps speed, she stresses, "as long as the result is reported in a timely manner."
That there is often a wait for results is no surprise, says Diane D. Davey, MD, professor of pathology and laboratory medicine and director of cytopathology at the University of Kentucky Chandler Medical Center, Lexington. Technologists are in short supply, and cytotechnologist slots are especially hard to fill. "Vacancies have increased recently," Dr. Davey adds, "and a decrease in training programs and number of graduates has compounded the problem."
"Preanalytic procedures and getting reports to charts can certainly cause delays, and those are things that should be investigated," says Dr. Davey, who chairs the CAP Cytopathology Committee. Furthermore, "technologists may be doing a lot of paperwork or computer tasks, and those could be turned over to another staff person. But a factor the lab may not have much control over is the shortage of technologists. In the future, we may have better screening instruments that can help eliminate more of the manual screening time, but that’s not a reality right now."
Meanwhile, the University of Iowa lab forges ahead. Zaleski’s next step is to survey clinicians to gauge their perceptions of turnaround time and assess their expectations.
"After all these efforts, and now knowing that we’ve dropped the mean turnaround time from nine days to around three, we’ll be very curious what the clinician’s perception is versus the realitywhat the gap is, if any," Zaleski says.
But whatever the survey finds, laboratorians at the University of Iowa have seen that industrial engineering methods work in health care, says Iowa’s Dr. Cohen.
"Big automakers want to crank out a certain number of cars daily and know every step along the way where some of the problems are," he says. "If they do that, they’ll figure out a way to speed up the production process. Our study, to me, was a great example of applying industrial methodology in a health care setting."
Melanie Fridl Ross is a freelance writer in Ocala, Fla.