Health care workers screen a patient for COVID-19 at a drive-through coronavirus testing site on March 18, 2020 in Arlington, Virginia. Arlington County and Virginia Hospital Center have opened a temporary drive-through coronavirus testing site for Arlington residents and county employees with a letter from a licensed health-care provider. The results for patients tested are estimated to be available in 5-7 days. (Photo by Drew Angerer/Getty Images)

As health officials in Virginia develop new ways to track COVID-19, the disease caused by a new coronavirus, one thing becomes increasingly clear: the state’s public tally of laboratory-confirmed cases is likely a significant underestimate of the actual spread.

As of April 2, state epidemiologists were investigating another 132 suspected cases of COVID-19 — numbers that aren’t currently included on the agency’s public surveillance site

VDH only recently developed its own definition for “suspect” cases in the absence of federal criteria from the Centers for Disease Control and Prevention and the Council of State and Territorial Epidemiologists, said Deputy State Epidemiologist Carrie Holsinger. 

Normally, those organizations are responsible for drafting uniform definitions and passing them down to states “to increase the specificity of reporting and improve the comparability of diseases reported from different geographic areas,” according to the CDC. When epidemiologists look at cases of measles across the country, for instance, they want to make sure that states are defining and reporting the disease in the same way.

But when it comes to COVID-19, the CDC and CSTE still haven’t developed a suspect case definition. 

“We just don’t have that national standard yet,” Holsinger said. 

In an email on Friday, CDC spokeswoman Melissa Brower wrote that both agencies “are working in collaboration with state and local health public health departments to develop a standardized case definition for COVID-19, understanding the urgent and time-sensitive issues that must be addressed.” 

The agency did not comment on when it expected to finalize its definition for the disease.

VDH developed its own definition in recognition that there are still barriers to testing a meaningful percentage of the state’s population, Holsinger said. As of Friday, Virginia had 2,012 laboratory-confirmed COVID-19 cases with 312 hospitalizations and 46 deaths. A total of 19,005 people have been tested for the disease — less than 0.22 percent of Virginians, and far less than nearby states such as Tennessee, which has seen a similar number of cases.

At a press briefing on Wednesday, state laboratory director Denise Toney said Virginia had the capacity to test more than 2,000 residents and recently doubled the number of tests it could process in a day. The same week, VDH expanded its testing criteria to include people hospitalized with fever or symptoms of lower respiratory illness — potentially opening the door to a significant uptick in testing.

“Still, we know with the lab-confirmed cases, we’re only capturing a portion of the people with infections,” Holsinger said. Recent evidence suggests that up to 25 percent of people with COVID-19 may be asymptomatic, making them ineligible for testing under most criteria but still able to transmit the disease.

Tracking suspect, as well as lab-confirmed cases, makes it possible for Virginia health officials to get a better sense of how the disease is truly spreading. VDH is now encouraging providers to enter potential COVID-19 cases into a surveillance system for infectious diseases, Holsinger said. Once a case is entered in, state epidemiologists assess it to see if it meets the definition for a suspect case.

Right now, the criteria is tailored toward the most likely symptoms of the disease. Holsinger said the definition currently includes patients with an unexplained respiratory tract infection and a history of travel to areas with community transmission, patients with respiratory tract infections and close contact with a lab-confirmed case of COVID-19, or patients hospitalized for a severe respiratory tract infection that has no other explanation (confirmed through negative flu and respiratory panels).

If a patient meets one of those criteria, VDH will record it as a suspect case — a designation that allows health officials to investigate it in the same way they would a lab-confirmed case. Holsinger was quick to emphasize that the data is self-reported by providers, making it likely that the state is still missing a significant number of cases.

“We would love to say we’re catching every suspect case, but we expect that to be a significant underestimate,” she said. “We’re encouraging providers to report, but we acknowledge how busy they are right now.”

At the same time, VDH is conducting something called “syndromic surveillance,” using data from hospitals and urgent care centers to tally likely COVID-19 cases. The department — almost in real-time — is able to track intakes with symptoms that align with the disease, said Erin Austin, an enhanced surveillance coordinator at VDH.

Experts started with complaints of fever, fever with cough, fever with shortness of breath, or fever with trouble breathing, Austin added. Recently, the surveillance team started tracking cases of pneumonia — which could indicate more severe presentations of the disease — and any intakes with a specific mention of COVID-19.

“We’re really trying to determine where we’re seeing an increase of those symptoms and what the demand could be on the health care system,” Austin said, adding that state officials are seeing an uptick in emergency room patients with COVID-19 type symptoms.

VDH plans to start including both suspect cases and syndromic surveillance on the department’s public COVID-19 page in the next week or two, Holsinger said. State officials — including Gov. Ralph Northam and other public health leaders — have access to the data, though it’s unclear to what extent it’s currently informing policy decisions.

Health Commissioner Dr. Norman Oliver said Wednesday that the data on suspect cases and syndromic surveillance is still too preliminary to factor into decisions. But Northam’s spokeswoman, Alena Yarmosky, said the governor is considering all available data when it comes to major policy announcements, such as the stay-at-home order he issued last week.

“We’re aware that we’re limited in what we know in terms of testing and we’re now relying on a variety of information,” she said. “He’s literally consulting with VDH multiple times a day.”

Austin said the bigger goal was to post the data in an easily digestible public format for local officials, who could use it to better determine how COVID-19 is affecting their communities.

“We’re trying to use them to paint a bigger picture of, ‘Yes, the virus may be present in your area even though you’re not seeing a very high number of positive tests,’” she said. “We think it would be able to better inform them how the population is being affected, and how hospitals should be preparing.”