Coronavirus testing is often scattershot. That could improve in North Dakota when a new study launches

FARGO — Public health officials struggle to get an accurate picture of how the coronavirus is spreading through the population — a job made all the more difficult because up to 40% of those infected show no symptoms.

Scarce testing supplies are often reserved for those who do have symptoms, although North Dakota has been more aggressive than many states in surveillance testing to determine the infection rate for the population at large.

Still, it’s been more scattershot than public officials would like.

But North Dakota could have a much better handle on the coronavirus pandemic under an initiative to scientifically select a representative sampling of households around the state and test them over time.

Researchers at North Dakota State University and the University of North Dakota, in partnership with the North Dakota Department of Health, are designing the study that the group hopes to launch this fall.

“I’m concerned about the fall,” said Dr. Paul Carson, a physician who specializes in infectious diseases at NDSU and one of the study leaders.

Fall is a strategic period to start the testing study, since students will be returning to classrooms, cooler weather will mean people will spend more time indoors and influenza season will start — factors that could drive an upswing in the virus.

The study will track viral tests — using the now-familiar nasal swabs — as well as antibody tests and T-cell immune response tests, a combination Carson said will provide a much more “granular” understanding of the virus around the state.

“We’re very excited to get that going,” he said. “We’re just in our heavy-duty planning now.”

The T-cell test measures how a person’s immune system responds to the virus. Carson compared it to the immune system’s “memory.”

Thus far in the pandemic, T-cell response testing has not been conducted at population scale.

Researchers are determining how large the sample size must be and how frequently they will be able to conduct the serial testing. The project, which will be partly funded by a grant from the Centers for Disease Control and Prevention, likely would span a year, he said.

They are selecting a representative study sample — with variables including geography, age, gender and household size — to come up with an acceptable margin of error, similar to political polling.

Earlier this year, Indiana conducted statewide sampling for prevalence of the coronavirus using a sample size of 3,600 for a state with a population almost nine times larger than North Dakota, which Carson said will require a much smaller sample size.

“We’ve got to do a lot of math in the near future,” he said, to determine sample size and testing frequency.

So far, North Dakota has tested more than one-fifth of the population, more than 165,700 unique individuals.

A large portion of testing involves repeated testing of health care workers, with a focus on employees of nursing homes and other long-term care centers.

“Although they are not a perfect representative sample of the North Dakota population, they are still a good ‘snapshot’ of the susceptible population out and about in the community,” Carson said, adding they reflect a variety of roles, including nurses, nursing aides, janitors and clerks. “And they are a group that everyone identifies as being especially important for regular testing.”

Critics of North Dakota’s method of calculating the positivity rate of the virus, a closely watched indicator, say the way the state lumps in serial tests with individual testing skews the results, giving a better reflection of the prevalence of the virus in nursing homes than among the population at large.

In monitoring influenza and respiratory syncytial virus, a common respiratory virus, the North Dakota Department of Health determines the positivity rate by dividing the number of positive cases by the total number of tests.

That method does not involve the serial testing critics contend muddies the coronavirus positivity picture and results in a rate that is deceptively low, as previously reported by The Forum .

For example, the daily positivity rate for results the state released Friday, Aug. 7, 2.6%, is far lower than the 7.4% weekly moving average positivity rate indicated Friday by John Hopkins University , widely followed for its COVID-19 trend reporting.

The Centers for Disease Control and Prevention provides yet another number, showing North Dakota’s positivity rate ranging up to 5%.

The positivity rate isn’t a good gauge of virus prevalence, Carson said. He relies much more on deaths and hospitalizations. There is no consistency in the way states report their numbers, he said, and the CDC hasn’t required standardized reporting.

Also, some states, such as North Dakota, do far more surveillance testing than others, making accurate comparisons difficult, Carson said.

“The variability of state-to-state testing capabilities and their strategies are all over the board,” he said, making state-by-state positivity comparisons “a fruitless effort.”

It’s not valid to compare North Dakota’s method of determining positivity for the flu to its method for the coronavirus, which involves significant repeat testing of vulnerable populations, a spokeswoman for the North Dakota Department of Health said.

“Because this is a global pandemic, it’s not accurate to compare COVID-19 to influenza or RSV,” said Nicole Peske, chief of communications for the department. “Individuals do not normally get repeat or ongoing testing for flu or RSV. Typically during flu season, people experience symptoms and go in to their health care provider to be tested. It wouldn’t be practical to be tested repeatedly.”

Repeat testing is important for vulnerable populations, Carson said.

“The fact that they may test negative last week does not mean they are no longer susceptible and no longer worthy of surveying,” Carson said of the serial testing group. “Their exposure risk continues on. It is still extremely valuable information to see what has happened to them this week, and the next, and the next.”

Ideally, he said, he’d like to see the state track both those who are tested because they have symptoms and those tested in surveillance testing and to report them as separate categories.

Once North Dakota launches its three-pronged statewide testing initiative based on a scientific sample of subjects, the state will have much more sophisticated information for tracking the virus, Carson said.