According to figures updated generated by the Centers for Disease Control and Prevention, the federal government likely overestimated the extent of Ebola’s spread.

In Septmember, the CDC predicted Sierra Leone, Guinea and Liberia would experience 1.4 million cases of Ebola cases by Jan. 20, 2015 without the implementation of additional interventions.

As of Tuesday, the total number of suspected or confirmed cases totaled 25,611 in the three countries — far lower than originally predicted by the CDC.

In a recent study published in the Proceedings of the Royal Society B, Aaron King, associate professor of ecology and evolutionary biology and mathematics at the University, and his colleagues, sought to answer what lead to this overestimation.

In the study, King and the other co-authors argue that faulty modeling was the cause for the errors in the initial predictions. In an e-mail to the Michigan Daily, King said no model is perfect, as they work through oversimplification.

“Like maps, (models) oversimplify reality,” he wrote. “Also like maps, that’s what makes them useful. The challenge in using models is to capture the key elements while ignoring the irrelevant details.”

King said one of the problems was that the model ignored the randomness in how Ebola is spread from person to person. An infected person can transmit the virus to either a few or many people, he said, so disease transmission is not entirely predictable.

“Models that ignore the random element in disease spread are simpler and easier to employ, but overstate the predictability of an outbreak,” he wrote.

Marisa Eisenberg, assistant professor of epidemiology and mathematics, wrote in an e-mail to the Daily that making predictions in the early phase of an epidemic is also challenging because researchers have to analyze with limited information.

“The lack of data and unreliability of what data there is can be very tricky to deal with,” Eisenberg wrote. “It’s a bit like weather prediction in some ways — making short-term predictions is more doable, but long-term predictions are more difficult and require more data.”

King also pointed out that the general modeling theory commonly used to predict outbreaks of diseases, including Ebola, H1N1, cholera and seasonal flu, assumes an exponential growth of the cases. Therefore, more Ebola cases lead to a greater number of predicted cases. Exponential growth is seen at the early phases of a disease outbreak.

Eden Wells, clinical associate professor of epidemiology, said changing conditions in the three West African countries included in the CDC predictions, such as additional foreign aid workers providing education about safe burial practices and providing medical help, could also have been why the disease spread much less than expected.

“Many of the assumptions that the modelers used were that the exponential growth would continue,” Wells said. “Because there were so many things rapidly evolving on the ground, the (exponential growth) didn’t happen. Models can overestimate because the assumptions change quickly.”

Wells added that although the model led to incorrect predictions of the outbreak, the overestimation was probably helpful in getting people to quickly respond to the disease.

“To me though, (overestimation) is not necessarily a horrible thing,” she said. “When we have to assume the worst possible scenario, it is a great way to get our government leaders and our healthcare providers around the world to move quickly.”

Leave a comment

Your email address will not be published.