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.

The lab that is processing TestUtah’s coronavirus tests kept specimens at the wrong temperature, analyzed leaky samples against its own policy, and didn’t verify the accuracy of a machine it was using, according to federal inspectors.

Meanwhile, the lab’s director was “unavailable” for an entire month while the lab more than tripled its workload and undertook a testing process unlike anything it had done before, in order to process TestUtah’s tests.

Those are some of the violations identified in a new report by regulators from the Centers for Medicare and Medicaid Services who inspected the lab at Timpanogos Regional Hospital after they received a complaint.

Nevertheless, state officials have renewed the contract with Nomi Health, the software company that operates TestUtah — and specifically agreed that the federal violations will only scuttle the deal if the lab ends up losing its certification altogether.

Nomi, which has billed the state more than $7.6 million for testing and web services through May, will be paid at least another $2.6 million to run TestUtah through mid-July. More payments are possible, depending on how many test sites it operates and how many tests it conducts. Nomi and its partner Utah tech companies have landed more than $60 million in no-bid contracts to manage testing systems in three states.

State officials have requested proposals from other mobile testing vendors but extended TestUtah’s contract before the scheduled date to open bids.

“We worked to get the [request] out as soon as possible once the governor announced in early May that we were returning to the usual purchasing process,” said Tom Hudachko, spokesman for the Utah Department of Health. “Given the timeline, it would not have been possible to get the RFP [request for proposals] out and have a vendor selected prior to Nomi’s contract expiring.”

Questions about test accuracy have dogged TestUtah since April, when data showed the statewide testing system was producing an unusually low rate of positive results.

But MountainStar, the company that owns Timpanogos Regional Hospital, characterized the violations regulators found at the hospital’s lab as procedural, focusing on the lab’s written policies.

“Many of the concerns noted in the ... audit had to do with clearly delineated policies around lab procedures. While all safety and efficiency procedures have been in place from the very beginning of the pandemic, they are now more clearly written down in an official policy,” hospital spokesman Mike Graul wrote in a statement.

What inspectors found

The federal report does not shed light on whether or how many patients’ results may have been inaccurate. But investigators did find “multiple COVID-19 tests” with irregular outcomes — and no records “showing what, if any, action was taken prior to reporting patient results,” according to the report.

Inspectors found that the hospital lab “modified” how it was processing coronavirus tests on April 10, apparently by introducing a machine to automate one of the steps. But the lab either did not perform or document required maintenance on more than 10% of the days inspectors checked, according to the report — though Graul said the user manual calls for maintenance only “weekly” or “annually.”

The lab also ran no experiment to verify that the machine produced accurate test results — something lab directors are required to do, inspectors wrote. In fact, they noted, Timpanogos Regional’s lab director was unavailable for almost the entire month that TestUtah was getting off the ground.

“The named physician medical director for the lab is in the high-risk category for COVID-19 and has been following guidelines to isolate at home. Delegation for all validation and in-lab oversight throughout the COVID-19 testing process was given to an interim medical director,” Graul wrote.

But a lot changed under the interim director’s watch. Before April 1, the lab did not test specimens collected off-site, from other providers, and it performed about 17,000 tests per month, inspectors wrote.

Since TestUtah began sending its COVID-19 tests there, it has been running about 52,000 tests per month.

But the lab did not implement a number of protocols and guidelines that are generally required for medical testing. “The only COVID-19 data the laboratory is monitoring as part of their [quality assessment] plan is turnaround time,” inspectors wrote.

For instance, the lab’s process for documenting complaints doesn’t account for complaints that originate outside of the hospital — and although staff could recall at least one complaint about its coronavirus testing, inspectors could find no record of it.

The lab also had no system to document and communicate problems with test collection that occurred off-site, inspectors wrote. When the lab received leaky samples — a problem noticed by other lab technicians who have handled TestUtah’s tests — “collection sites responsible for leaking specimens were not identified nor contacted.”

Moreover, the lab continued to test leaky specimens despite its own rules forbidding it if the leaks could expose the lab employees, the report states.

MountainStar said there are “multiple safeguards in place to ensure vials are securely sealed and transported, including securely sealing the vials with a cap and then sealing the vial in a bag."

”There is minimal risk that a leaky specimen could be contaminated,” Graul said.

More questions, from storage to data collection

Leaks weren’t the only problems with the samples, according to inspectors. The manufacturer’s instructions for the tests require samples to be refrigerated after collection at 36 to 46 degrees and then shipped on an ice pack.

"For all referral specimens sent to the laboratory for COVID-19 testing, the laboratory is unaware of what temperature the samples were stored at prior to shipment to the laboratory," inspectors wrote. TestUtah samples were not sent with ice packs as instructed, inspectors wrote.

And once they arrived at the lab, they were stored either at room temperature or refrigerated — not frozen at -22 degrees, as required by the test instructions, inspectors wrote.

The federal guideline “does not state that the lab’s transportation, storage and preservation policies and procedures must conform to certain criteria,” Graul countered.

“It merely requires that such policies must be established, in writing. The lab has such written policies which meet this standard. Further, the lab is working to implement processes to go above and beyond this written policy including on-site refrigeration and stringent processes around transportation of these kits and specimens.”

Test results also were incorrectly named for TestUtah patients, inspectors said. When samples from the hospital’s own patients produced an “invalid” result, the samples “were recollected” and retested. But for patients from the drive-through TestUtah sites, an “invalid” test meant they simply received a result of “inconclusive.”

“The lab and Test Utah initiative is now using the word ‘invalid’ instead of ‘inconclusive,’” Graul wrote.

There also were problems with how the test requests and results were labeled, inspectors wrote. Federal rules require specimens to come with a request form with an array of details: who’s requesting the test, the sex and age of the patient, the date of collection, the test desired, and other information.

But that’s not what TestUtah samples arrive with, inspectors wrote. “Specimens sent to the laboratory for COVID-19 testing are not accompanied by any type of requisition,” they wrote. Meanwhile, three of the result reports they reviewed also didn’t have the patient identification number — “the only patient information the lab receives.”

That was done by design, Graul said. “In order to ensure patient privacy, the Utah Department of Health dictated that the lab and the Test Utah initiative would not know the names of the patients tested through the program,” Graul wrote.

“The state worked with its Silicon Slopes partners to implement an advanced barcode system for identifying each patient. ... A system for multiple verification has now been added to the system by our Silicon Slopes partners while continuing to put the upmost importance on protecting patient privacy.”

The inspectors also said the lab’s director generally “failed to include a system for the COVID-19 testing that addresses the identification of problems, resolution, and a detailed plan of the actions that should be taken by the laboratory to prevent future recurrence of any problems.”

The lab’s supervisors, meanwhile, “failed to provide day-to-day supervision of testing personnel and reporting of test results for COVID-19,” the report said.

State: Deficiencies are not disqualifying

The inspectors’ findings apparently led to a dispute among health officials over whether the state’s public health lab should send tests to Timpanogos Regional Hospital for processing, in order to relieve a testing backlog from growing coronavirus outbreaks around the state.

The director of the Utah Public Health Lab, Dr. Robyn Atkinson-Dunn, said she was stripped of her title after she objected to sending patients’ samples “to a laboratory that has clear documentation from CLIA that they are not up to par.” CLIA refers to Clinical Laboratory Improvement Amendments, which govern medical labs nationwide.

The backlog has since been addressed, and the state lab is processing its tests in-house, said Tom Hudachko, spokesman for the Utah Department of Health.

“We make the decision on where to divert samples by which lab has the capacity to take our excess on any given day. Earlier this week that happened to be [Timpanogos Regional],” Hudachko wrote in an email responding to The Tribune’s questions.

He acknowledged federal inspectors found “some areas of non-compliance” but “determined those findings do not rise to the level that would close the lab." The hospital’s lab "has indicated they will address the areas of non-compliance,” Hudachko said.

Nomi Health added in a statement Friday: “We, like the State of Utah — which has continued to send samples to the Timpanogos Lab following the report — trust our partners are delivering quality care in a compliant way. Our partners have shown significant transparency around the validation of their processes and have worked closely with state and federal regulators to share data and results.”

According to a May 29 letter from federal regulators, the Timpanogos Regional lab had 10 days to respond to the violations. In order to keep its certification, the hospital lab has to show, among other things: how it identified patients who may have been affected by each “deficient practice”; how it corrected the problem for patients who were affected; what measures it took to make sure the problem doesn’t happen again; and how it will monitor that solution’s success.

The hospital’s response was not included in the report CMS released this week. If the corrections pass regulators’ muster, inspectors may return to the lab for a follow-up visit. If not, the lab could face fines of more than $6,000 per day or lose certification and Medicare payments.

The report does not address results of a now-scrapped accuracy check that state health officials were reviewing with Timpanogos Regional at the time of the inspection.

After TestUtah refused to join Utah’s other labs in a collaborative “proficiency challenge” to check the sensitivity of one another’s coronavirus tests, Timpanogos Regional instead agreed to a more rudimentary, “compromise” experiment: Exchange 90 samples with the state lab and compare results.

The state lab and Timpanogos Regional tested the samples in early May, and state health officials and corporate executives over the hospital reviewed the results for two weeks. But after health officials reported that a final draft had gone to UDOH director Jeff Burton for approval, Burton said he would not release them. There were problems with one batch of the samples the public health lab provided, he said.

Correspondence from MountainStar shows the public health lab sent over a new batch of samples; it’s not clear why that was inadequate to release the results. But, Graul said, “specific details of the many measures of validation conducted by the lab are clearly presented in our response to [federal inspectors].”

Burton said he has confidence that TestUtah’s results are reliable, but state health officials will devise a new test to assess them.

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. Infectious disease experts questioned whether the test the tech companies were using was sensitive enough to be reliably accurate, and one doctor on a state task force warned against having tests “routed to a small community hospital lab inexperienced with highly complex molecular testing.”

While the federal audit of Timpanogos Regional Hospital’s lab was underway, TestUtah’s samples were briefly rerouted to ARUP Laboratories in Salt Lake City. There, technicians found that more than 10% of the samples could not be tested because they arrived with too little fluid in the tube; patients were not notified of the problem for more than a week after that.

Burton has said the tests were rerouted to ARUP for about three days in mid-May to allow the hospital lab to set up equipment that would automate part of its coronavirus test analysis, and to confirm the machines produced accurate results. It’s not clear what machines the hospital obtained in May; according the federal audit, the lab obtained a machine to automate the same step in early April and did not verify its accuracy.

Reporters Bethany Rodgers and Sean Means contributed to this story.