COVID-19 simulation model creates scenarios
A glimpse of how behaviors might change the future sounds like science fiction, but it’s the reality of how Navid Ghaffarzadegan’s COVID-19 simulation model will create scenarios for planning purposes at Virginia Tech.
“Dr. Ghaffarzadegan’s model is one of several tools we can use to track a variety of indicators that impact the spread of COVID-19,” said Virginia Tech President Tim Sands. “A science-based approach to planning and decision-making helps us mitigate risk as we deliver on our education, research, and engagement missions.”
An associate professor in the Grado Department of Industrial and Systems Engineering, Ghaffarzadegan uses a combination of mathematical modeling, complex systems methods, and simulation techniques to gauge how proactive measures will impact the spread of the virus on the Blacksburg campus. Such factors as the continued adoption of face coverings, the frequency of public health messaging, and the availability of vaccines are used to predict the rate of infection and likelihood of completing the spring semester as planned.
“We are in a marathon,” said Ghaffarzadegan of his findings. “We’re doing great half of the way, people are cheering us on, but if we want the medal, we have to make it to the end of the line.”
Ghaffarzadegan’s work is now being integrated into the process of making proactive decisions about managing COVID-19 on campus and in communities by university leaders. A simplified, user-friendly version of his model is available for the public. People are encouraged to experiment with adjusting different behavioral factors and see how they might impact the virus’ spread.
Ghaffarzadegan’s model grew out of the diverse collaboration of Virginia Tech experts who came together during the summer to help the university safety navigate the pandemic. The group was made up of faculty, ranging from public health, industrial engineering, and biological science to physics, mathematics and statistics, as well as administrators well versed in campus operations.
“The diversity of expertise and viewpoints allowed the group to look at the problem from many perspectives and that generated very robust results,” said Ron Fricker, interim dean of the College of Science. “And as with our expertise, we also used a diversity of models in the process, which helped ensure that our conclusions were not specific to any one model and, thus, hopefully aligned with whatever would actually happen in the real world.”
Among the complementary models, Ghaffarzadegan’s work incorporated behavioral parameters and allowed real-time “what if” analysis. During the fall, he was able to fine tune those combinations to pinpoint which behaviors contributed to the completion of the fall semester as planned.
“The really exciting thing about this is that he could look at which factors are most important and show what might have happened if we had done things differently,” said Laura Hungerford, professor and head of the Virginia Tech Public Health Program in the Virginia-Maryland College of Veterinary Medicine. “The biggest message from that is how sensitive success is to the things our students have adopted – wearing masks, physically distancing, and avoiding crowds and poorly ventilated places."
In their simplest form, models are the weighing of factors to predict outcomes during a decision-making process. And their value not only comes in those predictions, but in helping people understand how factors interact and influence one another.
“Mathematical models help with much more complex predictions where factors interact in ways that aren’t intuitively obvious,” Hungerford said.
Ghaffarzadegan said such an occurrence was very clear when studying his model in collaboration with others. Not only did they see a greater positive impact when multiple public health policies were implemented together, but he found they actually have a multiplying positive effect.
“I was working closely with my Virginia Tech colleagues Uwe Tauber, Lauren Childs, and several others. Sometimes we took different approaches and had different assumptions, but many times our results were consistent. One common observation was that there is no silver-bullet, and measurable impacts come from implementing all policies. The reason is that there’s a synergic relationship between policies. For example, if more people wear masks, the spread of the virus slows down, and you can use your testing capacity to do more prevalence testing and find asymptomatic cases,” he said. “In simple words, mask wearing can make testing more effective too.”
Looking toward the spring semester, Ghaffarzadegan’s model provides a handful of critical insights for both the university and individuals:
- Frequent public health messaging is critical as both a reminder and method of positive reinforcement.
- We should continue to be patient; each person needs to commit to the same precautions they did in the fall semester; mask wearing and social distancing are as critical as before.
- Vaccination is needed but widespread positive impacts from vaccinations are months away.
“Most importantly, the model tells us that continued use of nonpharmaceutical interventions – mask wearing, physical distancing, maintaining your pod – will continue to be important throughout the spring term,” Fricker said. “It is very important we keep our guard up for the rest of the semester.”
And the science-based approach should also help people more clearly see the nearing of the end of our efforts.
“Being able to see these models not only highlights the best ways to manage the pandemic, but also help predict when we can get back to some pre-pandemic activities without putting our community at risk for sudden increases in spread,” Hungerford said.
— Written by Travis Williams