Root Cause Analysis Used To Find Source of Errors

After validation of new chemistry analyzer, random errors trigger letters to 1,000 patients

CEO SUMMARY: It is a reminder that today’s sophisticated laboratory test systems still have the potential to malfunction in unexpected ways. This spring, following installation and validation of a new chemistry instrument, one Canadian hospital laboratory quickly recognized that a series of random errors were a sign that the instrument had issues—but not before specimens from 1,000 patients had been tested, requiring notification letters to be sent to these individuals.

ANYTHING CAN GO WRONG when installing a new lab analyzer. Hardware, software, or both can fail during the initial tests.

Recently, one Canadian hospital laboratory sent letters to 1,000 patients suggesting they should consult with their physicians about the need for retesting. This step was taken after it was determined that a new chemistry analyzer was producing lab test results that could be inaccurate.

The patient notification program was carried out because of concern about test accuracy for specimens run between May 28 and June 6 on the new lab analyzer. The episode is a reminder to all clinical laboratories that, because of the increased volume of patient testing, any small problems or glitches with laboratory testing instruments have the potential to quickly affect a significant number of patients.

Further, the root cause of the problem demonstrates how the complex technologies that come together in today’s state-of- the art diagnostic instrument systems can often generate problems that are a challenge for the typical lab’s QA/QC program to identify and resolve.

“This story begins on April 23,” stated Shane Buchanan, who is Lab Manager at 106-bed in Truro, Nova Scotia. “That is the day when the manufacturer delivered a new analyzer to measure blood glucose, creatinine, calcium, and other biochemical analytes.

New Laboratory Analyzer

“With the manufacturer’s technician on site, we began to test the instrument and validate the assays,” he continued. “After the analyzer was moved from the area where it was validated to the actual laboratory and patient testing was initiated, we had a small number of infrequent and random errors in our quality control (QC) testing. The fact that the errors were infrequent and so random made identifying the cause extremely challenging.” Buchanan did not want to identify the vendor.

“Each time we had an erroneous QC result, we re-ran our QC samples through the system and each time they checked out fine,” recalled Buchanan. “Based on the accurate QC procedures and upon consultation with the technical representative from the vendor, we attributed such problems as part of the normal growing pains involved with the set-up of a new analyzer.

“These errors were so random that even the vendor had trouble identifying the cause,” he added. “We transitioned from running 25 samples per day during validation to our full production runs of 200 to 400 live patient samples per day.

“The first time an abnormal result was seen in a patient sample was June 4th,” he said. “The abnormal tests were rerun on another analyzer and manual delta checks were performed on seemingly abnormal patients. The vendor was contacted and this was again attributed to ‘growing pains.’

“Minimal testing was performed on the analyzer on June 5th,” noted Buchanan. “By June 6th, we determined that the best resolution was to take this instrument out of service. We wanted to err on the side of patient safety and we had a second analyzer available. This is when we sent letters to patients informing them about the situation and the possible need for retesting.

Taken Out Of Service

“Once the analyzer went out of service, we identified the source of the problem,” noted Buchanan. “The machine forms its own cuvettes with two sets of film. One track carries the front film and one track carries the back. After the films fuse together, hot air is blown in to form a ‘bubble’ which is the cuvette.

“The glitch occurred when consecutive tests were run,” he continued. “If one test had a high volume of reagent, the reagent would overflow into the adjoining cuvette and affect the test results in that adjoining sample. But it wouldn’t happen every time and it happened only when the reagent reached a certain level.

“The root cause turned out to be a locking screw that was not secured properly on the cuvette assembly mechanism during a routine service call performed by the vendor,” Buchanan said. “The vendor indicated that this error had occurred in only one other analyzer in almost 10 years. This was the cause of that small and random error leading to the inaccurate results.”

The experience of Colchester Regional Hospital’s lab is a reminder that sophisticated analyzers can still malfunction in unexpected ways—and when they do, large numbers of patients can be affected.

Lessons Learned in Set-up Of New Chemistry Analyzer

FOR LAB MANAGER SHANE BUCHANAN, five important lessons were learned from the experience this spring when random errors began appearing following the installation of a new chemistry analyzer.

“First, it helps to be judicial in following all quality control (QC) steps, as our lab staff was doing,” stated Buchanan. “Second, we are working with the vendor and other laboratories to establish delta check limits, a capability that exists in our laboratory information system. Not every hospital lab runs them.

“Delta checks are one way to rule out mislabeling, clerical errors, or possible analytical errors,” he said. “Running them may have helped us to more quickly find the source of the errors with our new analyzer.

“Third, we consolidated all of our log books,” stated Buchanan. “Previously, we had one log for QC problems and one log for troubleshooting. Now, all these communications go into a single logbook.

“Fourth, we asked the vendor to work with us to improve our QC education, troubleshooting, and documentation,” he added.

“Fifth, our chemistry staff took this incident to heart,” concluded Buchanan. “Many of them felt bad about it. We emphasized that this was a random error and that they had done the right thing at every step. Because of how these incidents can affect patients, it caused our lab staff to redouble efforts to ensure errors don’t happen.”

Tags: , , , , , , , , , , , , , , , ,

Enter Your Login Credentials
This setting should only be used on your home or work computer.

×

Send this to a friend