After months of anticipation, the COVID-19 vaccine has been delivered to every state in the nation and inoculations are underway. But vaccinating more than 250 million adults throughout the country is a monumental task that requires careful planning and assessments of different approaches to distribution – without which, herd immunity can take longer to achieve.
At Los Alamos National Laboratory, we’re using mathematical models and computational simulations enabled by the laboratory’s supercomputing capabilities to understand how best to distribute the COVID-19 vaccine. And what we’ve learned is: While the vaccine is a critical weapon in fighting this virus, it’s not a magic bullet, at least not yet.
Our models look at individual communities based on government data. To understand the different outcomes based on how the vaccine will be distributed, we create various what-if scenarios that were developed in collaboration with local, state and federal governments to help them effectively plan for vaccine distribution and complementary mitigation strategies.
Our models can drill down to the county level by incorporating explicit demographics – age, gender, household size, etc. – and even different industries in which people work. This level of granularity, something unique to our models, gives us a clearer picture of the impact of the vaccine on a community and different populations within that community.
We ran multiple simulations based on various scenarios, including vaccine effectiveness, allocation and prioritized. We also simulated the percentage of people willing to get vaccinated, which will have a significant impact on the spread of the disease. Based on surveys of adults, 40% to 60% have said they are willing to get the vaccine, so we simulated the outcome based on that range. We also factored in variables such as school attendance, mobility data and public interactions in various businesses.
So when we did all this, what did we learn?
Consistently, these models illustrate that, for many months, the vaccine alone isn’t going to be enough to keep us safe. Due to the limited vaccine supply and the fact that immunity builds steadily for several weeks after vaccination, our models show that continuing to limit business activity will allow communities to flatten the curve and subsequently increase the potential impact of the vaccine. Furthermore, they show that opening schools at full capacity can increase the risk of COVID-19 spread, while the hybrid learning scenario – 40% of students go to school in person for two days and the other 40% go the other two days – in combination with limited business activity reduces risk, enables in-person education and increases the impact of the vaccine by flattening the curve.
Our models are not foolproof. Being able to account for uncertainties in people’s behaviors and the spread of a brand-new pathogen in a complex model is extremely challenging – and something we spend significant time trying to understand. But the models are still valuable in helping us to quantify the potential outcomes of different what-if scenarios.
And what they show us is that it’s critical for everyone to recognize the important role they play in slowing the disease’s spread. Because we don’t often see the immediate impact of our actions, it’s hard sometimes to understand that individual behaviors make a difference. But they do. By wearing masks, social distancing, and, when it’s available, getting the vaccine, we all can do a tremendous amount to protect ourselves and others and keep the virus at bay.
Sara Del Valle is a mathematical epidemiologist and leader of the COVID-19 modeling team at Los Alamos National Laboratory. Ben McMahon is also a mathematical epidemiologist who heads up Los Alamos’ part of the Department of Energy’s National Virtual Biotechnology Laboratory modeling effort to tackle COVID-19.