Predicting whether a pathogen will have an impact on a few people or an entire population would be a huge achievement in global health security. Public health leaders would be able to determine the most effective response, whether it is expending resources on vaccination, or quarantining people in their homes, or just letting a disease run its course if it isn’t life threatening.
Researchers have turned to information technology to develop mathematical models that may predict the next infectious disease outbreak, but the models so far rely on data from past events to predict the future. Thus, the results can be mixed since pathogens mutate and don’t always affect people in the same way.
More recently, the promise of advances in computing power has enabled real-time collections of data from the Internet and the natural environment. Scientists hope so-called “big data” may improve the accuracy of prediction models.
Scientist and health journalism student Prajakta Dhapte became fascinated with this predictive process and decided to delve into the modeling arena for a recent story in Georgia Health News: “Predicting Pandemics: It’s not easy but researchers are trying.”
“Viruses are highly elusive in nature and you never know when or where the recipe for the perfect pandemic is brewing,” Dhapte said in a new “How I Did It” piece for AHCJ. “Viruses are way ahead of their time and this really fascinates me, their ability to evolve is incredible.”
A virologist studying for her master’s degree in journalism from the University of Georgia, Dhapte decided to pursue mathematical modeling to understand whether it is “a reliable method to predict infectious disease outbreaks and if any progress has been made in this area.” What she found is that the quality and quantity of real-time data about infectious diseases is lacking and that the search engine statistics and existing algorithms can’t yet capture the complexity of how pathogens may impact human populations.
“The newest prediction models remain ‘way too simple’ to forecast transmission,” Jorge A. Alfaro-Murillo, an associate research scientist at Yale School of Public Health’s Center for Infectious Disease Modeling, told Dhapte for her story.
But scientists keep trying, with the support of the Centers for Disease Control and Prevention, the National Institutes of Health, the U.S. State Department, private organizations like the EcoHealth Alliance, health systems and other organizations.
The CDC funds the Epidemic Prediction Initiative, which is monitoring dengue, influenza, and mosquito activity as part of efforts to enhance the predictability of whether there will be a flare up of disease. The NIH support the Models of Infectious Disease Agent Study (MIDAS), which has supported projects ranging from the development simulation models of measles spread to Ebola outbreaks.
The CDC’s predictive flu project, called FluSight, has been developing different mathematical algorithms matched with data collection, often using data gathered through Google searches or a model developed with data on flu-related hospitalizations. No one model has emerged as of yet that is better than another.
Still, there is evidence that existing mathematical prediction models deployed by the U.S. Department of Health and Human Services can predict severity of the flu better than just using historical baseline trends, according to a February 2019 study published in the Proceedings of the National Academies of Sciences.
Even so, the authors of the study said the sample sizes of the models studied were too small to draw definite conclusions about which models work best.