News →

Models Predict Uptick in Cases as Virginia Reopens

Medial professional inserts nasal swab into wincing National Guard member in Richmond
A member of the National Guard gets tested for COVID-19 in a training exercise earlier this week (Crixell Matthews/VPM News)

COVID-19 deaths and infections could skyrocket in Virginia in the coming months if left unchecked, according to updated modeling from the University of Virginia.

But researchers at UVA’s Biocomplexity Institute said the most dire predictions will almost certainly be avoided as people and governments change behavior to cope with new cases.

Their latest update on May 19 models a lifting of public health restrictions on May 15. It projects a statewide peak of between 101,000 and 145,000 weekly cases in July, depending on the amount of social distancing practiced.

In the Richmond metro area, the most extreme scenario in UVA’s model shows cases rising to around 23,500 the week ending July 5 -- a total of around 2% of the region's population infected in a single week.

Those outcomes would cause a dramatic surge in hospitalizations, overwhelming facilities in Northern and eastern Virginia, according to a May 13 briefing sent by UVA to VDH and obtained by VPM through a Freedom of Information Act request.

The UVA model isn’t the only projection showing a rise in cases. The Centers for Disease Control and Prevention tracks 10 different projections. All of them show COVID-19 deaths will continue to rise for the next month, though the most optimistic show a leveling off. One so-called ensemble model from researchers at the University of Massachusetts, which attempts to synthesize other projections, shows Virginia accumulating nearly 500 new COVID-19 deaths by mid-June.

Chris Barrett, the Biocomplexity Institute’s executive director, said he’s confident the worst of UVA's projections will be avoided.

“Those peaks are the effect of the model if you don’t do anything based on where you are now,” Barrett said. “You would expect to never see those peaks. Because you would expect people wouldn’t just walk straight into that, they would do something.”

Skeptics of the models have pointed to those discrepancies as proof of their unreliability. While some early models predicted hospitals would be overwhelmed, those predictions haven’t been borne out.

Madhav Marathe, a professor in Biocomplexity at UVA, said that’s the point.

“I think that it's actually a success for models to have warned folks and not let it happen,” Marathe said. “And I think that's what will happen in this case. Now, when does society react to that fact -- it's a societal decision.”

Several studies of cell phone mobility data have shown an uptick in travel in recent weeks. But Marathe said it was unclear how factors like widespread use of masks and keeping six feet apart might influence the spread of COVID-19.

In April, the institute signed a contract with the Virginia Department of Health to provide Virginia-specific weekly updates to its model. The UVA model is one of several reviewed by Gov. Ralph Northam.

Researchers agree that social distancing is still the best tool for avoiding the worst outcomes.

“We think that the social distancing has indeed continued to work and has bought us time,” Marathe said. The institute estimates that distancing allowed Virginia to avoid over 80,000 cases.

Commissioner of Health Norman Oliver argued Virginia was operating under a different scenario than the one outlined in the UVA model, which tracks what happens if Virginians’ social interactions start to return by 25% or 50% increments to what they were before the pandemic.

“We’re talking about people still staying at home as often as they can -- not really going out unless necessary or to enjoy some of the reopened restaurants,” Oliver said. “But even there, it’s at limited occupancy and so-on.”

If cases rise sharply, Oliver said he would return to an old message: Maintain social distancing. And he said the state is hiring 1,000 contact tracers and expanding testing to help find and isolate the infected.

Many public health experts expect some version of a rolling epidemic that consists of small peaks and valleys rather than a dramatic spike. Models are most helpful and accurate a few weeks out, when policymakers can use them to forecast a surge in hospital beds or to try to predict the effects of different policies, according to Bryan Lewis, research associate professor at UVA.

“That’s the point of the modeling -- to play out different scenarios that could be horrific to see in the real world,” Lewis said.

UVA is currently working on a model that would show the effect contact tracing would have in infections, Lewis said, and is in ongoing discussions with VDH on that work.

Correction: An earlier version of this article misstated the percentage of Richmond that the UVA model projects would be infected. The story has been updated to reflect the fact that the numbers were calculated for the Richmond region, not the city alone.