New Rochelle, a city of 80,000 people in upstate New York, was one of the earliest COVID-19 hot spots in the country. In early March last year, it experienced a serious outbreak even before nearby New York City became the initial epicenter of the virus.
Because it’s a midsize city, lessons drawn from New Rochelle are easily translatable to other similar-size communities across America. That’s why a team of engineers at New York University used it as the basis of a new predictive model for assessing COVID-19 transmission rates, and for finding out whether vaccination priority systems are effective. The study’s results led them to question the need to prioritize certain high-risk groups, especially when available doses are so low, since that vaccination strategy didn’t appear to lower infection rates as much as lockdowns and other measures to stop the spread.
The team ran simulations of the city during the pandemic. To do this, they used what’s called agent-based modeling, meaning the simulation accounts for the actions of each individual agent—each person in that town. They recreated the entire town in “granular” detail using maps and Census data. Then, they superimposed a representation of the pandemic onto it, replicating the behavior, mobility patterns, and the probabilities of people getting infected and infecting others. They ran a series of simulations with changing pandemic variables, such as different vaccination priorities, testing availabilities, and lockdowns.
“These models are very good, not to do predictions, but to understand what can be the effect of alternative approaches,” says Maurizio Porfiri, an NYU engineering professor who led the team. For instance, they drew a conclusion from one of their “what-if scenarios,” that having more drive-through testing locations, as opposed to hospital testing sites, could help lower transmission.
But, the main focus was on vaccinations, and they found that prioritizing certain high-risk groups for vaccinations only has a “marginal effect” on curbing COVID-19 cases versus just giving it people randomly. “That’s largely because there isn’t yet enough supply. “Prioritizing starts to make sense when you have a large enough number of vaccine doses,” says Alessandro Rizzo, another member of the team and a visiting professor from Turin Polytechnic. “If you have a small amount, it doesn’t make a great difference.”
The shortage of doses is part of the reason why the CDC has recommended a phasing strategy, implemented by most state and local governments, that pushes populations like older people and essential workers to the front of the line. But, the researchers found that vaccinating a significant number of people of any category could be more effective in lowering cases than picking and choosing populations. Volume and speed are the real priorities, they say, and recommend that the government works quickly to obtain more doses. “Don’t spend your time and resources on figuring out how to distribute 100 doses,” Rizzo says. Then: “When you have tens of thousands of those doses, then you figure out how to prioritize.”
So, are governments’ phasing strategies wrong? “Absolutely not,” both Porfiri and Rizzo flatly emphasized. Vaccinating elderly people first is necessary and right for reducing deaths. “For example, if you give 2,000 vaccines to people who are older than 80 years old, you may end up saving their lives,” Porfiri says. “But, if you ask the question whether you stop the chain of transmission, 2,000 won’t do the trick. The disease is still going to spread.” They added that continuing to prioritize by age could be more effective in reducing the death count than putting frontline workers at the front of the queue. That’s been a decision based more on economic outcomes, Rizzo says, than purely on infection rates, because the government wants to keep society running as much as possible.
While the government strives to get more doses, they said lockdowns should be in place. The model found that it’s not until about 25% of the population gets vaccinated that lockdowns should get lifted. That’s not a threshold, but “simply a rough indication of when a vaccine can produce an effect that could be at least comparable” to lockdowns. “The lockdown was extremely successful,” Porfiri says. “It created a lot of fatigue, [and] there are consequences from a mental health point of view. But from the transmission point of view, it proved to be excellent.”
The platform the team created is open source, meaning governments around the nation could use the model for their own towns of similar sizes, or even expand it to state levels—although the mapping would be incredibly labor intensive. But, while authorities can use this as a tool to help make decisions, the lessons garnered for the public are just as useful, they say. Whether or not lockdowns happen, public health precautions like wearing masks and social distancing remain imperative, even as vaccines roll out.
“What we can do with our lifestyle can be as impactful, or more impactful, than the vaccine itself,” Porfiri says. “It’s on us to contain the disease, until we reach a good amount [of vaccines] and things can look different. Don’t expect this vaccine to do a miracle. There is no miracle.”