A flu epidemic in Portland? There’s a model for that.

Engineering Prof. Siqian Shen and collaborators from Sandia National Laboratories, a contractor for the U.S. Department of Energy, have created an optimization model that could help public health officials make decisions about which places to close in epidemic situations.

The project uses data from Portland, Oregons social networks and censuses to model how the flu would spread when cities close facilities to contain it. It builds on previous simulations by taking into account people’s continued movement after certain facilities are closed, known as compensatory behaviors.

The model gives the option of closing five different facilities and then simulates how the population would move given those compensatory behaviors. Shen said the optimal choice would be the combination of five different facilities that spreads the population out the most.

“After you think about it, you kind of see the intuition behind it,” Shen said.

Shen added that when it comes to modeling infectious diseases, there are two main questions: how much to invest in preventative measures, and how much to invest in intervention after the fact.

Prevention, measured in vaccination rates, is straightforward — the more people get vaccinated, the better a city is protected in the event of an outbreak. Intervention on the other hand, is harder to quantify. Policymakers might close facilities or work with security agencies to restrict travel out of certain areas, among other options.

“At the beginning we had several very complicated models involving more different variables than this model,” Shen said.

She said this led her team to a number of “failed models,” but eventually they were able to narrow it down to the “most sensitive part of the model” — how people move around the city when certain places are closed.

Simplification of the model also made it a more powerful tool, Shen said. The simulation takes data from 100 people moving between 195 locations. To use this model, a city would have to collect travel data from only 100 people. This includes information about the disease and census data to determine how susceptible each person would be. It could then help in showing what closings would disperse, and thereby protect, the population the most.

The team chose Portland because the data happened to already be aggregated. Shen said she’s hoping to gather more data from other cities to expand on the model as the project moves forward.

As for a potential epidemic at the University, Shen said she’d have to collect the data about where students study first.

“I have no idea,” she said. “I guess it also varies among the students who are studying on North Campus versus Central Campus.”

Leave a comment

Your email address will not be published.