When TestUtah sites returned an unusually low number of positive results for the coronavirus this spring, state officials devised two unsuccessful attempts to discover the reason — then promised a third try.
The new review never got underway. But decisions by state officials have forced TestUtah to stop working with the two testing partners it had chosen — Salt Lake City company Co-Diagnostics Inc., which had supplied its testing kits, and Orem’s Timpanogos Regional Hospital, which had processed them.
Meanwhile, the details of an agreement that accompanies the state’s contract with Nomi Health, the Orem tech startup behind TestUtah, remain under negotiation, said Tom Hudachko, spokesman for the Utah Department of Health. A contract to pay Nomi for its sample collection services was signed in July: it’s unclear why the service agreement, which defines its operations, hasn’t been finalized as it continues to run 11 testing sites across Utah.
Co-Diagnostics had been providing its COVID-19 test to TestUtah — under a no-bid contract with Nomi — since Utah’s tech sector launched the state-funded testing program at the beginning of April. Swabs shoved up noses from Logan to St. George were analyzed with the company’s test.
But when officials updated testing contracts in August, they required Nomi to switch labs and said the new lab would decide which test it would process.
Fulgent Therapeutics, based in Temple City, Calif., has signed on with the state to process TestUtah’s tests until July 2025. Fulgent developed its own test, which is now being used at TestUtah sites.
Nomi Health had submitted a proposal that included TestUtah continuing to use the hospital as its testing lab, said Chris Hughes, director of the state Division of Purchasing and General Services. But that proposal did not meet the score threshold required by the state to be considered a finalist, Hughes said.
Asked whether shortcomings found in a recent federal inspection of the hospital played a role in declining that proposal, Hughes said the state “does not provide debriefings on why a vendor’s proposal received a certain score.” However, the scorecard for Nomi Health’s proposal showed it received barely half the points available for its level of expertise, the category where its score had the biggest gap compared to Fulgent.
Nomi said in a statement, “While our lab [proposal] was ultimately not accepted, we respect the state’s process.”
The company noted the state did accept its sample collection proposal and that it has successfully transitioned to Fulgent as its lab partner in connection with that work. “We are proud of the tremendous work we have done in partnership with TestUtah," it said, "to provide the state with a comprehensive community testing solution that delivers affordable, quality, and efficient testing to get Utahns back to work.”
And Nomi CEO Mark Newman said in a statement that Nomi will continue to use Co-Diagnostics' tests at its testing sites in Iowa and Nebraska.
“Co-Diagnostics is a terrific Utah company,” Newman said, "who has been, and continues to be, a strong and valued partner to Nomi in its national testing efforts.
“They were among the first in the world to deliver [Food and Drug Administration]-authorized COVID-19 testing kits at an affordable cost with abundant supplies,” he said. “Despite those who refuse to acknowledge it, the fact remains: Their tests have been rigorously validated multiple times by third-party experts.”
Andrew Benson, head of corporate communications for Co-Diagnostics, did not comment specifically on the state’s decision that led to TestUtah no longer using its test. But in a statement, he said: “Co-Diagnostics remains committed to assisting in the battle against the coronavirus pandemic across the world and in our own community by providing high-quality molecular diagnostic solutions.”
A dropped investigation
TestUtah received months of scrutiny in Utah, starting with data showing its COVID-19 tests early on produced positive results at less than half the rate of the state’s other test providers.
Utah health officials proposed a full accuracy check, a “proficiency challenge” with other labs processing coronavirus tests in the state. But TestUtah and Timpanogos Regional Hospital declined to participate. In a second attempt, described as a less-sophisticated “compromise” experiment by state epidemiologist Dr. Angela Dunn, the TestUtah companies agreed to compare results with the state lab.
But after weeks of back-and-forth with the hospital, the state refused to release the results. Jeff Burton, then the director of the Utah Department of Health, said the test was flawed because there were “questions about the validity” of some of the samples used. He said he had confidence that TestUtah’s results were reliable but added that state health officials planned to devise a new test to assess them.
“We’re always concerned about accuracy,” Burton said.
Hudachko confirmed no new test was ever pursued; state health officials were waiting, he said, for a federal review of Timpanogos Regional Hospital’s lab to be completed.
At about the same time that the state was attempting to assess TestUtah’s performance, federal inspectors were examining the hospital’s lab. They found it in violation of its certification guidelines, due to problems with how the lab was processing its COVID-19 tests. The hospital has responded to the alleged violations, and federal regulators have not yet disclosed the outcome of the case.
But the inspectors' report didn’t establish whether TestUtah’s results were accurate — or, if they weren’t, whether the error was tied to sample collection, the lab process or the Co-Diagnostics test.
State data had showed that even for patients with symptoms, TestUtah’s positive results were below those reported at other test sites in the state. Co-Diagnostics attributes the difference to the fact that it was using a more lenient definition of who qualified to be tested, based on a broader list of symptoms than health care providers were using.
Now it’s been a month since TestUtah used the Timpanogos Regional lab, Hudachko said, and he doubts there would be any remaining samples to compare.
The health departments in Utah’s two largest counties say they have no record of false results from someone who was swabbed at a TestUtah site.
According to emails The Salt Lake Tribune obtained from Iowa health officials, early attempts there to confirm the test’s accuracy showed it was identifying the virus in only 49% of positive samples — but after making adjustments to the testing procedure, that state’s lab director said the tests were found to be at least 95% accurate.
“At the start ... the lab did run into problems but worked closely with Co-Diagnostics and resolved the issues that it encountered,” Dr. Mike Pentella, director of the Iowa State Hygienic Lab, wrote in a statement to The Tribune.
Co-Diagnostics coronavirus tests are also being used in other countries, where they were first marketed, and there’s nothing in the public record indicating concerns about the tests there.
A new test of the test
There was a separate attempt this summer to resolve the question of the accuracy of Co-Diagnostics' tests, but it also ended without a definitive answer.
The Utah Journalism Foundation funded the new test with a $60,000 grant, said Fred Esplin, a member of the board of directors of the nonprofit, which supports The Tribune and gives grants to other news projects in the state.
Esplin said the foundation agreed to finance the test at the request of Paul Huntsman, who sits on the foundation’s board and is chairman of the board of the separate nonprofit that oversees The Tribune.
“The thinking was,” Esplin said, “given questions raised about the efficacy of the test, it would be good to remove that question or to resolve that question.”
The test was performed both for and with companies with financial ties to Co-Diagnostics. The foundation awarded the grant to TNG Dx, a New York-based company that describes itself online as a wholesale distributor created in response to the pandemic. It is a distributor of Co-Diagnostics' tests, said Benson, with Co-Diagnostics.
The test was conducted in part by data scientist Dr. Meghan Lockard, who was contracted to do the work by TNG Dx, where she is a scientific adviser. TNG Dx said in a statement to The Tribune that it was vetting Co-Diagnostics' COVID-19 test to determine whether it wanted to distribute the test.
While performing the test, Lockard collaborated with Arches Research Inc., according to her report. Arches Research, a subsidiary of Utah company PolarityTE Inc., is a customer and partner of Co-Diagnostics. It uses Co-Diagnostics' tests in the coronavirus testing services it offers, and recently signed an agreement to process tests for other Co-Diagnostics customers.
“The collaboration with Arches,” Lockard explained in an email, “was a more rigorous replication of the validation experiment I performed myself.”
Lockard obtained free tests from Co-Diagnostics, her report said. To evaluate them, she did not use samples from patients. She instead began with inactive samples of the SARS-CoV-2 virus, provided by a medical supply company.
Using genetic material from the virus, she created 93 “replicates,” or samples that mimic the viral load that might be found in patients who are experiencing symptoms.
Her report, completed in early September, found the Co-Diagnostics tests performed even better than advertised.
“We found the test to be 100% sensitive when performed on 93 COVID-19 RNA replicates in total,” Lockard wrote, “whereas the Co-Diagnostics reports a sensitivity of 99.52% on 631 replicates.”
Lockhard was referring to data about past testing on replicates that is included on a fact sheet included in Co-Diagnostics' test kits.
James Westgard, an emeritus professor in the Department of Pathology and Laboratory Medicine at the University of Wisconsin Medical School, read Lockard’s report at the request of The Tribune.
He noted her study did not test the kits on real samples from patients, and she did not compare the results to those from other tests, hallmarks of the more rigorous approval process usually required by the FDA.
During the pandemic, Co-Diagnostics and other companies have been awarded emergency use authorizations that allow COVID-19 tests to enter the marketplace based on a lower standard of evidence, such as “allowing the use of ‘contrived’ samples where materials were spiked with the virus, rather than actual clinical specimens” in validation tests, Westgard explained.
Lockhard’s study “again only makes use of artificial samples and analyzes 93 replicates of a negative sample and 93 replicates of a positive sample, not 93 patient samples without COVID-19 and not 93 patient samples with documented COVID-19,” he wrote in an email to The Tribune. “Thus, this later validation report is of little value.”
Lockard agrees that she considered the validation using the replicates to be only a first phase of testing. If Co-Diagnostics' tests passed, she explained, she intended to conduct a second phase using patient samples. In a news release issued by TNG Dx about her test results, she said: “We will immediately begin working with outside labs to validate our findings for large-scale asymptomatic sampling.”
She explained in an interview: “Patient samples are very expensive, and so the Utah Journalism [Foundation], who was funding the study, wanted to do it in two phases so they would have the first right of refusal should the test not pan out in Phase I.”
But Lockard learned later in September — from a Tribune reporter — that there would be no financing of a second stage from the foundation. Esplin, with the foundation, said it will not finance a second phase with patient samples nor did he ever hear any discussion of the nonprofit financing a second phase.
Lockhard said she does not plan to proceed with more testing unless someone steps forward to pay for it. TNG Dx did not return calls seeking comment.
Benson, with Co-Diagnostics, also did not comment specifically on the TNG Dx test. But he said in his statement that the company followed FDA guidelines to verify the accuracy of its tests.
Court fights loom
Two shareholder lawsuits allege Co-Diagnostics and its leaders have falsely claimed its tests are 100% accurate, allowing them to benefit from inflated share prices by cashing in on stock options and selling stock.
Both suits, filed in federal court in Salt Lake City, point to drops in the share price after the TestUtah partners declined to join the state’s first accuracy challenge, and after the state hygienic lab in Iowa reported Co-Diagnostics tests found COVID-19 in 95% of positive cases.
The second suit was filed in September, but Co-Diagnostics has answered the earlier suit, filed in June. Its lawyers contend executives and company literature accurately described the results of specific tests and have asked for the suit to be dismissed. They point out that the suit doesn’t include any evidence of specific stock sales and doesn’t claim any sales were suspicious or unusual.
Early Co-Diagnostics investors have seen a boon during the pandemic. The stock was trading on the Nasdaq for 90 cents a share at the close of 2019. It climbed as high as $30.80 on Aug. 3. Co-Diagnostics stock has proved to be volatile, closing as low as $8.24 on Sept. 4 and at $14.20 on Monday.
Follow @natecarlisle