Abbott’s 15-minute coronavirus test was hailed as a ‘game changer.’ Now, many Connecticut hospitals aren’t using it.

When Chicago-based Abbott Laboratories released a rapid coronavirus test — advertised to give positive results in 5 minutes and negative results in 13 — officials across the country hailed the new technology as a pivotal moment in the pandemic. But now, medical professionals across Connecticut are shelving the test, citing concerns about its accuracy.

In late March, Dr. Scott Gottlieb, the former commissioner of the U.S. Food and Drug Administration, tweeted that Abbott’s ID NOW COVID-19 test was a “game changer,” while President Donald Trump called the test “a whole new ballgame.”

In April, Gov. Ned Lamont and his administration repeatedly touted the governor’s work to convince Abbott executives to bring the ID NOW test to the state. Later that month, Lamont said that “to be able to get the testing results in less than 15 minutes is how we’re going to lick this virus.”

But then, numerous studies raised concerns that the ID NOW might be letting coronavirus cases slip through the cracks. Although Abbott said the situation was fixed by tweaking the test instructions, more studies are now pointing to the same problems under the new instructions. Because of that, some Connecticut hospitals say they don’t feel comfortable using the test.

Dr. Scott Sussman, who oversees COVID-19 testing in the Yale New Haven Health system, said the hospital system received an Abbott ID NOW machine as a donation from the Mohegan Tribe.

“We actually returned the machine because we did not feel it was accurate enough to use for any patient decision-making,” Sussman said. “As much as we wanted to use it, it just didn’t make sense.”

The Yale New Haven Health, Hartford HealthCare and Trinity Health of New England systems all said they have opted not to use the ID NOW over concerns of false negatives. Nuvance Health, which operates Danbury Hospital, the state public health lab and the New Haven CVS rapid-testing site are continuing to use the test to varying degrees.

Josh Geballe, the state’s chief operating officer, said Wednesday that the state is aware of the concerns about the test, but has not yet taken any action.

In a statement, Abbott spokesperson John Koval dismissed the concerns and said the test “is performing as expected” when used appropriately.

Studies bring skepticism

On March 27, the FDA granted “emergency use authorization” for the Abbott ID NOW. Emergency use authorization is an expedited process with far fewer requirements than full FDA approval. This type of authorization is only used during public health emergencies.

It wasn’t long before the concerns started rolling in.

Numerous studies found that the ID NOW test missed 15% of positive coronavirus cases — or more — when researchers used samples that were stored in viral transport media. That method of sample collection was one of two initially approved by the FDA.

Abbott responded to the concerns by working with the FDA to modify the usage instructions for the ID NOW. About a month after the test was initially authorized, the FDA amended the instructions and advised against using viral transport media, which Abbott said was diluting the samples. Instead, the FDA said, the ID NOW could only be used on samples delivered directly from a swab. Abbott now says the problem is solved.

“When the direct swab method is used, the test is performing as expected and we are confident in its performance,” Koval said in a statement.

But local hospitals aren’t all convinced — and there are more studies to back up their skepticism.

Yale New Haven Health’s Sussman said he and his colleagues were on edge about the ID NOW after reading a study from Loyola University Medical Center. That study, which used Abbott’s preferred method of direct swabs, found a 25% rate of false negatives — meaning the ID NOW would miss one out of every four positive cases of coronavirus, even under the updated instructions.

Koval said Abbott believes the Loyola study was flawed. Another Abbott spokesperson made a similar statement to MassDevice after a New York University study found an even more drastic rate of false negatives.

The lead author of the Loyola study, Amanda Harrington, said she and her team “stand by the findings.”

‘Concerns about the accuracy’

Dr. Syed Hussain, the regional chief clinical officer for Trinity Health of New England, said the hospitals in the New England region decided not to even pursue the ID NOW platform.

“It was launched with a lot of fanfare and there was excitement because we all want to move to a state where it’s instant,” Hussain said. But “we picked up on a similar thing, where it would have a higher false negative rate.”

Hartford HealthCare’s chief clinical officer, Dr. Ajay Kumar, said the hospital system has four ID NOW machines that were intended to run the rapid coronavirus test. But those machines aren’t being used.

“Our team did not think it was worth it,” Kumar said.

With the Loyola study in mind, a team at Yale New Haven Health’s Lawrence + Memorial Hospital set out to test the accuracy of their ID NOW machine.

The lab’s medical director, Dr. Victoria Reyes, said the team tested the positive control sample sent by Abbott — and the machine returned a negative result. After talking with an Abbott representative, Reyes said, her lab then tested the machine with an alternate positive control sample sent by another manufacturer — and the machine again returned a negative result.

“When we were doing the validations, it wasn’t meeting our expectations,” Reyes said. “We had concerns about the accuracy of the results that we were getting.”

Koval said Abbott believe the lab conducted the validation tests “inappropriately.”

“When ID NOW is used by health care professionals trained on the system, it performs as expected,” Koval said in a statement.

Reyes said there are a number of things that could’ve gone wrong with the validation of the ID NOW, including the possible incompatibility of the second sample her lab used. But she also said that many coronavirus testing platforms have been rushed in an attempt to meet the massive need.

“The situation is just that we need testing and everybody is scrambling to get their tests out there,” Reyes said.

‘An important tool’

Those who continue to use the ID NOW say the test’s speed makes it a useful diagnostic tool. Most other tests take at least 2 hours to process, compared to the ID NOW’s 15-minute turnaround.

It’s a point that Abbott emphasizes as well.

“ID NOW is an important tool that delivers information where it’s needed most — taking the fight to the front lines so that public health officials and health care providers can make critical decisions,” Koval said in a statement.

Nuvance Health, which includes Danbury and Norwalk hospitals, is using the test on every patient admitted to the hospital, according to chief quality officer Dr. Diane Kantaros.

But if a patient is displaying coronavirus symptoms, they’ll likely receive a more traditional coronavirus test if their ID NOW test comes back negative, Kantaros.

“If it’s negative, then we’ll go an extra two steps to ensure that,” she said. “If it’s positive, it’s very helpful and gives us a result quickly on admission.”

She added that any test has the potential to give a false negative, and the medical staff is trained to compare test results with the symptoms the patient is displaying.

The director of the state’s public health lab, Jafar Razeq, also said it’s important for clinicians to pay close attention to patient symptoms.

Razeq said the public health lab is using the ID NOW, too, on a mobile testing unit. With a second mobile testing unit slated to come online this week, the lab will be able to process 100 to 120 tests a day on the ID NOW, Razeq said.

“These units are really great for quick turnaround time,” Razeq said. “This is a service that so far has been getting nothing but positive feedback.”

But Razeq said the question of false negatives is a concern for all testing platforms that have been approved under the FDA’s relaxed regulations. There simply isn’t enough hard science to draw a conclusion about the test, Razeq said.

“We do not have enough and robust data on most of the [tests] in terms of their performance characteristics,” he said. “All that I can say is that no testing platform or testing methodology is perfect.”

Editor’s note: A previous version of this story misidentified the tribe that donated the testing platform to Lawrence + Memorial Hospital. It was the Mohegan Tribe.

Emily Brindley can be reached at ebrindley@courant.com.