Carnegie Mellon University
August 26, 2020

Mathematical Sciences Professor Publishes and Critiques COVID-19 Modeling

By Ben Panko

New research from Carnegie Mellon University and the University of Pittsburgh published in the journal PLOS One models how age-targeted strategies to mitigate the transmission of COVID-19 could lead to better final outcomes in mortality and hospitalization from the novel coronavirus.

"COVID-19 has a striking age-dependent mortality rate," said Maria Chikina, an assistant professor at the University of Pittsburgh Medical School. "An infected 75-year-old is 1,000 times more likely to die than an infected 15-year-old, and an infected 55-year-old is 20 times more likely to die than an infected 25-year-old."

Chikina co-authored this study with Associate Professor of Mathematical Sciences Wesley Pegden to dig into this "unusual characteristic" of the disease. Using a standard "Susceptible, Infected, Removed" epidemiological model, the pair sought to model the dynamics of how often people in various age groups have contact with each other, how that contact affects the transmission of COVID-19 and how adjusting those contacts would affect the impact of the pandemic.

The researchers concluded that heterogenous strategies focusing on reducing transmission of COVID-19 to and among the highest-risk age groups are most effective at reducing mortality and intensive care unit usage. These strategies were more effective compared to homogeneous strategies targeting all age groups equally, in situations where population immunity plays a role in controlling the epidemic. In particular, their modeling shows that age-targeted mitigations can allow immunity to contribute to epidemic control even while significantly reducing the burden of infection among at-risk groups.

"We view our modeling as demonstrating a qualitative point: strict age-targeted mitigations can have a powerful effect on mortality and ICU utilization, even if relative transmission rates among age groups will eventually normalize," the authors concluded in the study. "We expect that public policy motivated by this kind of finding would have to be responsive; for example, by relaxing restrictions on larger and larger groups conservatively, while monitoring the progress of the epidemic."

This is not Chikina and Pegden's only work related to COVID-19; in March, they published an analysis attacking prominent modeling of the pandemic for seeming to use selective time ranges to cover up key information. These models, including one from the University of Washington and one shared widely by the New York Times, showed that strict- and medium-term social distancing would drastically reduce the number of infections and deaths from COVID-19.

However, these models only depicted a time span of a few months, and each failed to show the sharp increase in infections and deaths from COVID-19 that one would expect after strict social distancing ended. In fact, Pegden and Chikina were able to hack the Javascript code from the New York Times model and run it farther into the future to show that hundreds of millions of expected infections were being hidden off to the right of the chart.

"Two months of mitigations have not improved the outcome of the epidemic in this model; it has just delayed its terrible effects," the pair wrote. While they stress that delaying a sharp increase in infections can be helpful if it allows for the opportunity to develop treatments and improve hospital capacity, these models did not make this clear and instead implied that a few months of social distancing would, on its own, lead to fewer total infections and deaths.

"The duration of containment efforts does not matter if transmission rates return to normal when they end, and mortality rates have not improved," Chikina and Pegden wrote. "This is simply because as long as a large majority of the population remains uninfected, lifting containment measures will lead to an epidemic almost as large as would happen without having mitigations in place at all."

While criticizing hypothetical modeling can seem abstract in the scheme of pandemic-related problems, Chikina and Pegden note that dishonesty and poor communication can damage the credibility public health experts and officials need to convince people to follow their advice.

"The public should not be misled by presenting false stories of hope to motivate behavior in the short-term," Chikina and Pegden wrote. "Public health depends on public trust."

Even more so, Pegden notes, people should understand that pandemic modeling should not be viewed as precise predictions of the future.

 "On the other hand," Pegden said, "using models to explore crucial qualitative questions is an important tool we have to understand the possible impacts of the various options we have before us."