News: CJ Exclusives

COVID-19 data often incomplete or unavailable, researchers say

Decisions about N.C. schools, jobs, and businesses made behind fog of unclear information

N.C. Department of Health and Human Services Secretary Dr. Mandy Cohen at a May 2020 COVID-19 briefing. (Pool photo)
N.C. Department of Health and Human Services Secretary Dr. Mandy Cohen at a May 2020 COVID-19 briefing. (Pool photo)

As COVID-19 cases and hospitalizations rise across North Carolina, key parts of the state’s science and data remain inaccessible to the public. 

“North Carolina is relying on the data and the science to lift restrictions responsibly, and right now our increasing numbers show we need to hit the pause button while we work to stabilize our trends,” Gov. Roy Cooper said June 24 when he announced the modified Phase Two restrictions. Phase Two — which allows limited indoor restaurant dining while keeping bars, gyms, fitness centers, and bowling alleys, among other places closed — won’t be lifted before Aug. 7.

The lives and livelihoods of millions of North Carolinians depend on that science and data. But researchers say critical information is incomplete, unclear, or unavailable. 

North Carolina had 91,266 confirmed COVID-19 cases, 1,142 hospitalizations, and 1,568 deaths on Wednesday, July 15. 

But the true COVID-19 case count remains uncertain with renewed test shortages and without random testing. Infections could be 10 times higher than the state’s tally. Deaths and hospitalizations lag weeks behind cases, and not all hospitals report COVID-19 patients. Publicly available data on hospitalizations and hospital capacity remain limited.

North Carolina offers interactive data when it comes to COVID-19 cases and deaths. But the state doesn’t require hospitals to report COVID-19 patients, and it doesn’t break down hospitalizations for the public. 

Residents can’t track hospitalizations by county, age, race, ethnicity, or gender. Information about the severity of hospitalizations is also limited, as the state’s figures fail to separate intensive care unit beds from general hospital beds. 

Transparent hospitalization data is crucial to fighting the virus. Without this data, reopening the economy and preventing unnecessary deaths becomes that much harder, researchers say. 

“We can’t understand it properly because of the lack of this data,” said Pinar Karaca-Mandic, University of Minnesota professor and project lead of the national Hospitalization Tracking Project.

Here’s a look at the state’s science and data, what the public knows, and what it doesn’t:

How reliable is the daily COVID-19 case count?

Not very, researchers say. 

North Carolina confirmed 91,266 cases of COVID-19 as of July 15, but, based on the number of reported recoveries, the actual number of active cases is closer to 24,142. Still, actual infections could be up to 10 times greater than the state’s count, according to estimates by the Centers for Disease Control and Prevention.

North Carolina has completed almost 1.3 million coronavirus tests. Each case represents one person, even if that person ordered multiple tests, according to the dashboard of N.C. Department of Health and Human Services. 

The state grappled with testing shortages for months, and it limited testing to patients with symptoms. Researchers believe the bottleneck hid the true number of coronavirus cases. 

The state ramped up testing in early June for patients who didn’t show symptoms. A month later, test shortages returned, skewing the state’s case counts. As the wait time for test results climbed toward seven days, during a July news briefing state health department secretary Dr. Mandy Cohen told reporters she was concerned about “ballooning turnaround times.” 

“The testing delays are a huge problem,” Karaca-Mandic said. “If today I’m seeing the data that was collected seven days ago, it doesn’t help me.”

The state graphs cases by the date of reporting and by the date of testing. Longer turnaround times affect all case counts during July, the state’s dashboard says. 

“I don’t trust case counts. I hardly look at them, except to see if they are moving in a particular direction,” said Julie Swann, head of the Department of Industrial and Systems Engineering at N.C. State University. She was a science adviser for the CDC’s H1N1 pandemic response.

How many people have COVID-19? 

The state says 8% of COVID-19 tests are coming back positive. But don’t rely on that estimate, researchers say. 

Labs weren’t required to report negative COVID-19 tests until July 7. To avoid inflating the rate of positive cases, the state includes only labs that electronically report both positive and negative tests. Roughly 70% of total tests over the past two weeks included positive and negative results. 

The percentage of positives reported is based from a sample size that’s smaller than the total tests reported. The data behind the 8% figure isn’t publicly available. 

Without randomized testing, it’s difficult to know how many people have the virus. Current testing is skewed toward those showing symptoms and the sickest, says Karaca-Mandic.

The positivity rate is also influenced by targeted testing. The state aims testing at people who are most vulnerable to get COVID-19, including those in nursing homes and minority communities. But targeted testing could obscure how far the virus has spread in the community, Karaca-Mandic said. 

“That rate by itself doesn’t tell much about the true rate of infection,” Karaca-Mandic said. 

Does the state know many people are hospitalized with COVID-19? 

Mostly. Hospitals choose whether or not to report COVID-19 hospitalizations. Not all hospitals report this information to the state, and reporting fluctuates each day. The number of patients hospitalized is subject to the percentage of hospitals reporting. July 15, 89% of hospitals reported — that’s 89% of the state’s hospitals, not 89% of the hospital beds.

Do hospitals have capacity?

Yes. As of July 15, 24% of inpatient hospital beds were empty, along with 19% of intensive care unit beds. There were 2,436 ventilators available. 

North Carolina had 1,142 COVID-19 hospitalizations on Wednesday, July 15, up 27% from 823 hospitalizations on June 13. Hospital reporting increased as well, but seven-day averages show a similar jump. 

Hospital admissions lag behind infections, as the virus can take weeks to make people sick enough to go to a hospital. The impact of today’s case count on hospitalizations will manifest itself in two weeks or more. 

Patients with COVID-19 occupied 5.4% of the state’s 21,114 staffed inpatient and ICU beds reported on July 15. The state doesn’t tell the public how many of those patients are in general hospital beds or intensive care units. It also doesn’t identify the location of patients in hospitals with the virus. 

Nor does it publicly break down hospital capacity to show where the virus could overwhelm the health care system. The capacity of rural and urban areas exists under the same data umbrella. 

Hospitals usually operate at fairly full capacity, while adjusting for seasonal illnesses or major events. They’re now monitoring virus spread and can add capacity if needed, said Tatyana Kelly, a N.C. Healthcare Association vice president.

Are hospitalizations spiking or staying flat in my county? 

Unclear. North Carolina doesn’t publish this information. Florida began breaking down hospitalizations by county, but North Carolina doesn’t. 

How many people have recovered?

The state says 67,124 patients were presumed to be recovered on Monday, July 13. The state publishes its estimate of the number of people recovered from COVID-19 symptoms every Monday. 

Does the 91,266 COVID-19 case count include recoveries?

Yes. If 67,124 patients have recovered from COVID-19 symptoms, then roughly 24,142 known active cases are left. 

That active case count reflects how the virus is spreading within North Carolina, says Rachel Graham, assistant epidemiology professor in the Gillings School of Global Public Health at the University of North Carolina at Chapel Hill.

“The active case number is really the number that should be looked at,” Graham said. “The active case count is going to give you a good idea of what it looks like right now. Total case count is cumulative, and it’s been cumulative since tests started.”

The day-to-day change in cases gives the most accurate picture of community spread, said Graham. North Carolina reported a 1,782 increase in cases since Tuesday, but test shortages complicate even that number. 

“Total cases is just one metric, and one should look at the different metrics together — percent positive, hospitalizations, etc. — and not only look at one metric,” said Kelly Conner, spokeswoman for NCDHHS. “You need all of them to see the whole picture.”

What do officials mean when they say a patient has recovered?

Unsure. The state’s definition of recovered says that patients no longer have symptoms, but the data behind the estimates isn’t publicly available. NCDHHS did not respond to requests for comment. 

Is the state breaking COVID-19 data down by race or ethnic group?

Yes and no. Deaths and cases get reported by race, ethnicity, age, gender, and county. Hospitalizations don’t. 

African Americans accounted for roughly 24% of cases and 33% of North Carolina’s 1,568 deaths. By ethnicity, Hispanics made up 44% of cases and 10% of deaths. But researchers can’t see how ethnicity, race, or gender affects patients’ odds of being hospitalized.

What about age?

Yes — for infections and deaths, but not hospitalizations. 

Demographics have shifted since the beginning of the outbreak. More younger people are getting the virus, with 69% of confirmed cases in people younger than 50. 

Patients 75 or older accounted for 58% of deaths. Those between the ages of 50 and 64 made up 16% of deaths. Patients 49 and younger totaled 6% of deaths.

How long is average hospitalization over time? 

Unclear. The state doesn’t release this data. But it estimates the median recovery time for hospitalized COVID-19 patients as 28 days from the date of COVID-19 testing. 

Can I track COVID-19 cases by ZIP code or county? 

Yes. The state began publishing COVID-19 cases by county and ZIP code in May. That data shows only lab-confirmed cases. 

Are patients diagnosed with COVID-19 but hospitalized for something else getting counted toward the state’s COVID-19 hospitalizations? 

Yes. Hospitalizations include patients admitted to the hospitals for other reasons who tested positive for COVID-19 and who were placed on infection prevention precautions.

“When you think about the role of elective procedures, a lot of our health systems are doing pre-screening. They’ll screen you as an outpatient then depending on the results they make a decision on whether this is the time to do an elective procedure or not,” Dr. Betsey Tilson, the state’s chief medical officer, said during a June press briefing. “So [the classifications] may be playing a small role. But I don’t think it’s playing the majority of the role.”

What about other deaths that aren’t reported as COVID-19? 

Unclear. All-cause deaths provide a window into unreported COVID-19 deaths and how delays in care affect mortality. North Carolina relies on a paper system of reporting death certificates, and it’s one of three states without an electronic registry. 

In the CDC’s data charts on excess deaths associated with COVID-19, North Carolina’s section warns “data in recent weeks are incomplete. Only 60% of death records are submitted to NCHS within 10 days of the date of death, and completeness varies by jurisdiction. Weights may be too low to account for underreporting, numbers of deaths are likely underestimated.”

Can I see if the virus is spreading in schools or child care settings?

Yes. North Carolina has linked one cluster with six cases to schools, and another 11 clusters with 83 cases to child care. Residents can also find where the clusters are by county. The state updates this data Tuesday and Friday. 

Is the state tracking the virus’ spread in nursing homes?

Somewhat. Nursing home residents accounted for roughly half of recorded deaths in North Carolina. The state tracks outbreaks in nursing homes — which the state defines as two or more cases in the same facility. But lawmakers blasted the state for failing to regularly test residents. In June, the state released a plan to test roughly 70,000 nursing home residents and staff by the end of August.

There were 136 outbreaks in nursing homes and another 92 in residential care facilities. Together, they tallied 6,672 cases and 845 deaths. 

What about prisons?

Yes. There are 1,920 cases and 31 deaths in correctional facilities. The state launched a plan to test all 31,000 offenders for COVID-19 over 60 days, starting in June. 

What about meatpacking plants? Or agricultural workers?

Yes and no. The state publishes a weekly report of cases within congregate living settings, including in homeless shelters and migrant worker housing. The numbers don’t include farmworkers who live in private residences. The report doesn’t name the farms with outbreaks. 

The most recent report says “there are no cases or evidence of COVID-19 transmission through food handling, packaging or processing.” The state previously released a list of outbreaks and case counts at meat processing plants by county without naming the specific plants. 

What else is the state doing to track COVID-19?

The state is tracking symptoms of COVID-19 using systems developed for the seasonal flu. Called symptomatic surveillance, the network relies on reporting from emergency rooms. It only flags those who seek care. The state also employs contact tracers, who look for people who have been in contact with people infected with the virus, but delays in test turnaround times will hinder their work.