Case closed: discrepant results at multiple sites

Create PDF

Kwong

A former colleague of Dr. Greene’s, Shiu-Land Kwong, CLS, MT(ASCP), further details this multisite rollout.

“It’s good for that first site to be really solid, so we have a good plan. It took over a year to get the first site ready. Behind the scenes, in addition to validation, there are LIS issues and other issues. There’s all of this background work in trying to make sure we’re able to duplicate it when we take off from the first site to the second one to all 21 sites,” says Kwong, regional director of laboratory compliance and risk management in the Permanente Medical Group Laboratory System in Northern California.

“We’ve been able to cut down the implementation time to about eight weeks after the site construction is complete,” Kwong says. “We’re glad to have a package to be able to roll out at each site.” The AU680 is live at eight of the Kaiser hospitals in Northern California, and another three are expected to go live by the end of the year.

Even with the same equipment, it remains a challenge to reduce or eliminate discrepancies. That requires understanding the clinically acceptable maximum imprecision and investigating matters when it is exceeded.

“It could be that calibration is overdue or there are issues with the instruments or reagents,” Kwong says. “There’s a lot of implications that this is something that needs to be looked at.”

The initial validation and rollout is essential, she adds, because “it gives us an internal reference point in addition to what the manufacturers have published in their reference data.”

And, yet, not all discrepancies can be eliminated.

“We can see this when we bring up all these instruments,” Kwong says. “We can have two brand-new instruments, yet they’re not the same. It would be great if the instrument-to-instrument variation would be minimized at the manufacturing level. However they put together the instruments, we may have a certain instrument consistently running on the negative or positive side. We do see that issue, even after we calibrate. That’s just the way it is. There are constant biases between instruments. It is fine if the bias is within the clinically and analytically acceptable limits defined by our laboratory system.”

Reducing the variation among laboratory sites and instruments is especially important in the era of patients having direct access to test results, Kwong says.

“In our system, the patients have access to their information online and are able to trend results themselves,” she says. “The patient doesn’t know where the testing was performed. All they know is, ‘This is my result.’ Therefore, it’s even more critical—when patients have electronic access to all laboratory results to look at them and trend them—that we have comparable results from all locations.”

In her AACC talk, Dr. Greene offered up evidence of how faulty calibration can contribute to greater variation, even when all the tests are performed on the same platform. Comparing how 21 Kaiser sites used 42 Beckman Coulter Access 2 immunoassay systems to perform the Accu-TnI+3 troponin I assay, Dr. Greene and her colleagues found instrument-to-instrument variation was likely the result of calibration bias (Clin Biochem. 2015;48[4–5]:268–274). Sites that recalibrated and repeated the comparison saw the between-instrument bias shift.

All of this makes the argument for standardizing laboratory equipment across sites even stronger as a way to achieve greater harmonization and more reliable results, Dr. Greene says.

“That’s what my data is showing. This is hard enough when everyone’s on the same instrument. There’s no way to do it when you’re on different instruments,” she says. “This also bolsters the argument for standardization of quality practices.”

Dr. Greene no longer has the ability to see how a change in a reagent might play out across 42 instruments as she did in her previous position at Kaiser. But she still has three University of Washington laboratories within her purview, and Dr. Greene works with colleagues around the country to gather data on interesting cases that elucidate the challenges of lab harmonization.

“Other people may not think this work is sexy,” she says, “but I find it fascinating.”
[hr]

Kevin B. O’Reilly is CAP TODAY senior editor.