Each morning, hundreds of thousands of Virginians click on a state website to get their daily dose of COVID-19 infection numbers.
Usually at 9 a.m., the Virginia Department of Health’s dashboard rolls out the daily numbers, a morose lottery that reports how many unlucky Virginians now have the disease, have been admitted to a hospital or have died. On Saturday, for instance, the department reported that 12,366 Virginians have tested positive for the virus out of 72,178 tests, that 436 people have died and that 1,942 have been hospitalized.
These are cumulative numbers, and unless you keep a daily log, you can’t know with a simple glance that these daily totals represent 772 more people knowing they have the illness, 26 more Virginians dying from the virus and 105 more admitted to hospitals than the day before.
Nor does Virginia tell you how many people have been discharged from hospitals or how many are believed to have recovered.
People are interested in the numbers. The health department’s website has logged 30 million more hits this spring than it did last year, and the landing page for the coronavirus has had 20 million views since it went up in March.
But what are these numbers actually telling people about the spread of the disease, the risk to individuals and the potential for overwhelming hospitals’ ability to care for the very ill?
How concerned should we be when on April 15, the daily report had 41 more people dying than the previous day, a marked spike over earlier days when single digits or numbers barely above 10 had been recorded? What does it mean when positive cases rise 700 or drop 300 in a day? Do the increases mean to hunker down, it’s getting worse? Do declines signal we’ve passed the peak and it’s OK to relax?
Dr. Lilian Peake, the state’s epidemiologist, said the daily reports reflect data entered into the health department’s system during the previous 24 hours. The numbers do not necessarily represent how the disease moves through a state or a community, and they include test results that were taken from sick people as soon as the day before or as far back as two weeks or more. They log deaths that might have happened weeks ago.
“We’re looking at it from different perspectives. Here you have a reporter looking to see what changed that day and looking for the story behind that. But really the purpose of epidemiology is to talk about the trends,” she said during an April 17 phone interview.
It’s less important to know the total of cases or how many are reported daily, she said, than it is to know how many people actually become ill each day.
All about the curve
To view the key disease tracking tool, known as an epidemiological curve, you need to click on the department’s weekly report and then on the “case timing” tab.
This is where the health department plots the dates of when people first began to feel ill. Local health department investigators determine this by interviewing each person and their doctors. If someone is on a ventilator or has died, case investigators make a best guess as to when symptoms first appeared.
The first two Virginia cases were reported to the health department on March 7.
“As we continued to get more cases and we went and interviewed them, the earliest known symptomatic cases occurred Feb. 28. We know of three cases, two in the Northern region and one in the Eastern region, that actually got sick on Feb. 28,” Peake said. “So that’s when we know it began to circulate in Virginia.”
Virginia’s epidemiological curve looks like the left side of a rising arc.
“For all disease outbreaks, you start with a susceptible population, the proportion of the people who can get sick. With a novel virus like this, it’s all of us,” she said. “At some point enough people are infected so it slows down the rate. There aren’t as many people who can get it quickly. It gets to a peak and it starts to decline.”
Researchers are modeling how they think the virus will act by making assumptions gleaned from its spread elsewhere.
The state model built by University of Virginia’s Biocomplexity Institute looks at what has occurred in each locality, and how people usually travel and interact, and then it predicts how the virus will affect people if they keep to the same patterns or stay home. The model pushes Virginia’s peak in infections out to the summer.
Peake can look at the curve and see that social distancing practices put into place in mid-March slowed the pace at which the disease was spreading.
Not overwhelming hospitals
From a public health perspective, slowing the spread of the disease doesn’t necessarily mean keeping any one particular person from falling ill. The goal is to keep many individuals from becoming seriously ill at the same time.
“It is important to us to explain how fast this is affecting people and how many people are getting seriously ill. That can help us to understand the impact on our health systems,” she said. “We have a good sense of what our hospitals can manage, and so we want to be looking at what is the potential impact of people that are expected to be hospitalized. That has been the key factor of this disease worldwide, is the number of people who can be hospitalized at one time is so great that it threatens our health care system. That doesn’t just affect people who need to be treated for COVID-19, it also affects other people who would be coming in with a heart attack or other types of severe illness.”
Virginia’s hospitals have more than 6,000 beds and nearly 2,900 ventilators, according to the Virginia Hospital and Healthcare Association’s COVID-19 dashboard. The association reports that the hospitals care for about 1,300 patients each day who have COVID-19, with about 400 of them in intensive care units and about 250 on ventilators.
The hospitals on Thursday asked the governor to allow them to restart elective surgeries and other procedures that were put on hold to prepare for the pandemic. They said they have enough capacity to handle the virus patients and to begin to treat other patients. Gov. Ralph Northam instead extended the ban until May 1.
Not all of the hospital patients are spread evenly across the state, and the UVa modelers have predicted that different regions will see peak patients at different times, starting in Northern Virginia and progressing to Southwest.
Northam has said he will rely on science and data to guide his decisions on how and when to reopen commerce. He often talks about his phone conversations with Maryland’s governor and the District of Columbia’s mayor, and about how Virginia will need to coordinate with its neighbors.
For example, if Virginia allowed people to dine in its restaurants before D.C. did, then the neighbors might come for dinner, and who knows what viruses might hitchhike across the Potomac with them.
While data from Washington and from Maryland’s nearby counties show similar trends in infection rates to Northern Virginia, the data tell different stories in other parts of Virginia.
Fairfax, Fairfax County, Arlington, Alexandria and Prince William and Loudon counties on Wednesday accounted for nearly half the COVID-19 cases in the state, 45% of the hospitalizations and 37% of the deaths.
In the Roanoke and New River valleys, the virus has killed four people, three of them Botetourt County residents, and has caused hospitalizations for 27 people out of the statewide total of 1,942. The cases are occurring at an even slower rate in counties west of the New River, with the exception of Washington and Smyth counties, where the Mount Rogers Health District has reported four outbreaks in long-term care and health facilities.
Northam’s executive orders so far have been applied statewide. He said Friday that Virginia is one commonwealth and that he would not use the data to consider easing restrictions in parts of the state that are less affected by the virus.
Who is affected
“One way we can use the data is to use the large numbers and look at who’s being affected, and that gives us opportunities to come up with strategies and let those people know of ways to protect themselves,” Peake said.
“What’s more important than the exact number of people who are hospitalized is looking at the information about those people. We absolutely know the people who are older are getting sicker, if you will,” she said.
The data show that children 9 and under, who represent 12% of Virginia’s population, account for fewer than 1% of the cases.
And it’s much the same for 10- to 19-year-olds, who make up 12.6% of the state’s population and about 2% of the COVID-19 cases.
No children have died in Virginia, and very few have been hospitalized.
“If you compare that to the older population who are 80 and older: That’s 3.5% of the population of Virginia, but it accounts for 42.6% of the deaths,” Peake said. “That’s what these data are really for. How to put the data together to see who is at risk and what actions can be put in place to protect them.”
That’s why it is also important to collect data on race.
“You can determine whether the population is disproportionately affected and have a different type of intervention,” she said. “If you look at data and see African Americans are disproportionately being hospitalized and disproportionately dying, then you need to understand why. What is going into that?”
It could be there’s a need for more testing to identify cases earlier and make sure they get the care they need, she said.
Initially, race was not collected since it isn’t filed with the electronic lab reports, but the information can be collected through health district investigations and hospital data and vital records.
Dr. Norman Oliver, the state’s health commissioner, said Monday that African Americans, who make up about 20% of Virginia’s population according to the U.S. Census, represent 28% of the COVID-19 cases for which race is known, and 34% of the deaths. It is not known what percentage of all the people who have been tested are African American.
What isn’t known
Virginia reports on how many people have been tested.
Virginia Secretary of Health and Human Services Daniel Carey has acknowledged that the state is at the bottom at rankings in per capita testing, and that the goal is to do thousands more tests each day.
He said the state has not been able to secure all of the supplies and equipment it needs amid shortages and competitions among the states.
Peake said some states had a boost because a commercial lab was headquartered there, such as LabCorp in North Carolina.
Also, Virginia did not have a federal testing site, so it missed out on those supplies.
Everyone has called for more testing, but how much is enough?
Tests have been reserved for top priorities: people in hospitals and those who care for them, emergency service workers who may have come in contact with an infected person, and those who live or work in nursing homes and other long-term care facilities.
“It’s helpful to understand how we know whether we are testing enough. The World Health Organization did talk about this recently, and they said countries that have extensive testing, typically fewer than 12% of their tests are positive,” Peake said. “So in Virginia right now, it’s about 15% of our tests are coming back positive.
“I do think the testing has been targeted at places with a lot of transmission. And if we were having a large percentage of our tests coming back positive, then there would be more concern,” she said. “That being said, I want to see the tests increase.”
The health department has not analyzed testing by localities or health districts.
“We want to understand if there are any differences in testing rates so we can focus testing there, and to correlate to cases and also to look at percent positive tests over time,” she said.
The department publishes the number of tests by district, and the number of positive cases. Calculations on Wednesday showed that in Roanoke, 1.5% of the 1,855 tests administered were positive. In the New River Health District, 5% of the 1,309 tests given were positive. In Richmond, 15.7% of the 1,647 tests given were positive. And in Fairfax, 23.6% of 9,759 tests were positive.
The state recently added hospitalizations and deaths by locality to its online dashboard.
But what it doesn’t report is how many of the local cases are linked to outbreaks in nursing homes, jails, homeless shelters or other places where people live or gather in close proximity.
It publishes information on outbreaks by category type and by health district, but with health districts often encompassing several localities it is not known exactly where they are occurring. The federal government last week directed long-term care facilities to tell families of positive cases for staff and residents and required them to report cases directly to the Centers for Disease Control.
Oliver, Virginia’s health commissioner, has said state code does not permit the department to release information on facilities.
On Wednesday, he said the homes can share information on positive cases with each other. This is important because nurses, certified nursing assistants and other providers often work at more than one home.
But the information still won’t be public.
The state posts information that includes the number of outbreaks, how many people are affected overall and how many have died. But that information lags media reports.
The site on Saturday still showed Roanoke as one of the few health districts without an outbreak, even though one of the city’s nursing homes had reported that a resident and three staff members were infected.
The total of deaths arising from long-term care facilities also was pegged at 78 on Thursday, which is most likely an undercount given the high deaths reported from homes in Henrico County, Harrisonburg and Northern Virginia.
It often takes time before deaths are entered into the system. Sometimes case investigators don’t know the person has died, or they haven’t had time to enter the data, or the case is not known until the death certificate is filed with vital records.
Concerned about COVID-19?
Sign up now to get the most recent coronavirus headlines and other important local and national news sent to your email inbox daily.