Utah scraps its investigation of TestUtah’s coronavirus accuracy and will try again

(Trent Nelson | The Salt Lake Tribune file photo ) COVID-19 testing at Timpanogos Regional Hospital in Orem on Tuesday, April 14, 2020.

Editor’s note: The Salt Lake Tribune is providing free access to critical stories about the coronavirus. Sign up for our Top Stories newsletter, sent to your inbox every weekday morning. To support journalism like this, please donate or become a subscriber.

Weeks after data showed TestUtah was returning an unusually low number of positive results for the coronavirus, state officials have scrapped an experiment that was intended to ascertain the accuracy of the statewide testing system operated by Utah tech companies.

And there are new questions after patients were directed to get retested, following problems that arose during the recent analysis of more than 100 TestUtah tests.

Meanwhile, health officials confirmed, the lab that is processing TestUtah’s tests — more than 26,000 as of mid-May — was under review earlier this month by federal regulators, prompted by an unspecified request.

Utah Department of Health Director Jeff Burton said he expects to receive their report as early as this week. He has confidence that TestUtah’s results are reliable, he said, but added that state health officials plan to devise a new test to assess them.

“We’re always concerned about accuracy,” Burton said.

State lab report rejected

State health officials earlier this month said they had devised an experiment to check the accuracy of results returned by TestUtah, whose operators have secured more than $60 million in no-bid contracts in Utah and two other states.

Questions about the precision of TestUtah’s tests have been circulating since data collected in April showed that the positive rate for the symptomatic patients they tested was less than half that of other test providers. And in Nebraska, where the same companies have contracts, their rate of positive results has been less than a third of the overall positive rate.

TestUtah refused to join Utah’s other labs in a collaborative “proficiency challenge” to check the sensitivity of one another’s coronavirus tests. Instead, in a compromise with health officials, TestUtah agreed to a more rudimentary experiment: It would exchange 90 samples with the state lab and compare results.

The state lab and TestUtah’s lab, which is at Timpanogos Regional Hospital in Orem, tested the samples in early May. The state then provided those results to the hospital’s parent corporation, HCA Healthcare, which reviewed them over several days last week.

The report was being finalized late last week, and, on Wednesday, a spokesman for the Utah Department of Health said the final draft had gone to Burton “for approval.”

Instead, Burton said later Wednesday, the report will not be released.

“There are some challenges with the way the [sample selection] was done,” Burton said. “Some of those were on our part. We're going back because some of those samples were old.”

Burton said he became aware of “questions about the validity” of the sample selection a week ago.

He did not comment specifically on why the department sent the report to HCA for a week of review, and then prepared a final draft for Burton’s approval, if irreparable problems with the study’s methodology were known. He also did not comment on why the disqualifying age of some samples wasn’t identified and potentially corrected earlier in the testing and analysis.

Now the state plans to create yet another experiment to check TestUtah’s accuracy, Burton said.

“We’re going to set up a third-party, neutral review,” he said. However, he did not say why the state health department’s neutrality might be in question — or what other lab might conduct the experiment.

“It's just a fairness issue,” Burton said. “We don't know yet.”

He said he did not know when the new experiment was likely to begin. “I'm going to have to talk to my lab director about that as soon as we can. We want to get it done,” he said. “It's unfortunate this happened, but it did.”

A federal inspection begins

Burton said the state lab’s comparison with TestUtah was delayed because federal regulators were auditing the lab at Timpanogos Regional Hospital — though a health department spokesman previously said the samples for the experiment were processed by May 9. Burton did not explain why a federal audit would affect the state’s work.

It’s unclear what inspectors’ findings were, or what prompted the audit, said both Burton and a spokesman for Timpanogos Regional.

“I have heard this one was asked for by somebody out of state,” Burton said.

The companies behind TestUtah have been criticized for problems at similar testing operations in Nebraska and Iowa, though the companies are using labs located in those states to process their tests. It’s unclear whether the federal review at Timpanogos Regional is tied to those complaints.

The Salt Lake Tribune has requested records of the inspection from the U.S. Centers for Medicare and Medicaid Services, which oversees lab certification, but CMS officials said the report is not complete.

Burton said the federal inspectors did not instruct TestUtah to suspend test processing at Timpanogos Regional. “Nobody — not anybody federally or locally — told them to stop testing,” he said.

But for three days in mid-May, the processing of TestUtah’s tests was moved to ARUP Laboratories in Salt Lake City, state and hospital officials confirmed.

“The state made this decision even though the lab at Timpanogos Regional Hospital has remained fully functioning throughout the pandemic,” hospital spokesman Mike Graul wrote in a statement.

The hospital obtained equipment to automate processing for the coronavirus tests about the same time as the federal review began, Burton said. As a result, TestUtah transferred 1,005 of its tests to ARUP for analysis, both to better accommodate the regulators and to confirm the new machines produced accurate results, Burton said.

But at ARUP, new problems arose.

‘You need to be retested’

Lindsay Kuttler went to TestUtah’s Orem site for a coronavirus test on May 11, while the tests were being sent to ARUP for analysis. She’d had some brief symptoms, and there were a lot of new infections in her neighborhood, she said, apparently related to a party. Kuttler hoped to rule out COVID-19 before an upcoming doctor’s appointment scheduled for her young daughter.

But the results didn’t come. “I was expecting it after 72 hours," Kuttler said. “I had to reschedule my daughter’s doctor’s appointment.”

Finally, after nine days, she received an email. “You need to be retested," the subject line read.

“We sincerely apologize, but we were not able to process your test. In spite of the fact that we are extremely careful with these tests, once in a while we encounter a problem with individual tests.”

Kuttler’s test could not be processed because “the sample leaked in transit or there was not sufficient liquid accompanying the sample,” stated the email, which was sent from the “COVID-19 Utah Task Force.”

In fact, Kuttler’s test was not an isolated case: More than 10% of the TestUtah samples — from more than 100 patients — could not be processed at ARUP because they arrived with too little fluid in the tubes, wrote Julio Delgado, chief medical officer for the lab, in a statement to The Tribune.

The lab’s first challenge, he explained, was that the TestUtah tubes containing the swabs and transport fluid were smaller than the vials ARUP typically uses in its own analysis of COVID-19 tests. So ARUP had to develop a new methodology to use with the smaller vials that met federal regulations, he said.

“Compliance with all regulatory requirements is critical because it helps to protect the validity of patient results,” Delgado wrote. “In the case of the TestUtah referral, the specimens we received did not meet the collection requirements for our test, so our laboratory directors and technologists worked overnight to validate a method to perform the test on the specimens received.”

But some samples still could not be analyzed because their tubes had a low level of fluid; some were almost empty, Delgado said.

“The specimen volumes were insufficient for us to be able to provide accurate test results, even using the method our laboratory worked overnight to validate to perform the TestUtah tests,” Delagdo wrote.

“Cost-effectiveness was not a factor” in ARUP’s inability to analyze those tests, he wrote.

Burton disputed that, saying ARUP “just decided flat-out it was not cost effective” to process the tests that had little fluid.

“It was more than just [low fluid], though that was one of the major factors,” Burton said. “If TestUtah had been using their [own] lab, this wouldn’t have been an issue.”

A return to Timpanogos

For TestUtah’s continuing testing, ARUP offered to provide collection kits compatible with its typical analysis process, Delgado said.

“TestUtah told ARUP it required its own labels on all tubes, and TestUtah was determining whether it could attach its labels to the collection tubes ARUP provided. ARUP never heard back from TestUtah,” Delgado wrote.

TestUtah instead resumed sending its tests to the lab at Timpanogos Regional for processing after three days, state health officials said.

It is unclear why TestUtah did not immediately notify patients of unprocessed tests.

A prepared statement from Nomi Health, the software company that runs TestUtah, read: “During the time the samples were being processed by ARUP, the TestUtah call center ... began to receive complaints of test result delays, and we are disappointed that any patients experienced any inconvenience. The Department of Health quickly redirected all samples back to Timpanogos.”

But Delgado said ARUP alerted TestUtah of the untested samples within two or three days of receiving them, once lab staffers determined they could not produce reliable results from the scant specimens.

Similar problems appear to have beset the tech companies’ operations in Iowa, where officials have confirmed that some of the TestIowa samples could not be processed due to leaks and damage. And in one county, health officials said nearly 10% of TestIowa’s results were “inconclusive.”

Meanwhile, staffers for Lt. Gov. Spencer Cox said they are trying to learn why TestUtah was calling itself the “COVID-19 Utah Task Force” in emails to patients — a name similar to the Utah COVID-19 Community Task Force, which is the state’s official coronavirus response team.

The email address itself appears to be connected to Qualtrics, the Provo tech company that received a $1.8 million no-bid contract from the state to help operate the TestUtah.com online assessment.

“The task force didn’t authorize it to go out under our name,” wrote Kirsten Rappleye, Cox’s spokeswoman, in response to the email alert to patients whose tests were unprocessed. “Saying it’s from the task force isn’t correct. … It’s fair to say that if this came from a testing site we will work to get it corrected quickly.”