Researchers at the University of Waterloo in Ontario, Canada, have developed a new method of analyzing social media posts to help identify growing vaccine skepticism.
Vaccines have been under fire in recent years, fueled by misinformation on social media that has helped lead to increased rates of measles, once thought to be an illness that was all but eliminated in modern society. In response to such issues, researchers at the University of Waterloo began developing a novel approach that could give public health officials early warning of potential outbreak locations.
“In nature, we have contagious systems like diseases,” said Dr. Chris Bauch, professor of Applied Mathematics at Waterloo, in a statement. “We decided to look at social dynamics like an ecological system and studied how misinformation can spread contagiously from user to user through a social media network.”
Researchers used a machine learning model based on the mathematical concept of a tipping point. “It doesn’t matter if you’re looking at a person’s body having an epileptic seizure, or an ecological system like a lake getting overrun by algae, or the loss of herd immunity within a population,” Bauch said.
“Mathematically, there’s a common underlying mechanism.”
The team analyzed tens of thousands of public posts on X from California before the 2014 measles outbreak. Older methods, such as counting skeptical tweets, provided less warning before the outbreak. The “tipping-point” approach, by comparison, detected small patterns in the data that seemed to offer more lead time.
“The tipping point approach looks for patterns in the data that are not easily captured with conventional methods, like looking for trends in the number of skeptical posts over time,” Bauch told The Debrief in an email. “These patterns are shared across different systems, and manifest in similar ways whether you are talking about the onset of an epileptic seizure, a lake becoming covered with algae, or past climate shifts such as the end of previous Ice Ages.”
“The usual methods of predicting an outbreak by doing a statistical analysis of skeptical tweets don’t provide much lead time,” Bauch noted. “By using the mathematical theory of tipping points, we were able to get a much bigger lead time and detect patterns in the data much more effectively.”
According to Bauch, the model’s accuracy was further confirmed by comparing posting patterns in California with those in similar regions where no outbreaks had occurred at the same time.
While the model was first tested on X, Bau can also be used on other social media platforms such as TikTok or Instagram. However, analyzing images and videos requires more computing resources than X’s predominantly text-based content.
“We saw the same kinds of patterns in the tweets that we see in other systems like lakes or climate shifts,” says Bauch. “The pattern manifests as a kind of ‘wobble’, similar to how a top wobbles before it is about to fall over.”
“These wobbles are predicted by the mathematical theory of tipping points. And the AI algorithms we used are able to find the patterns with good,” he says.
“Ultimately, we would like to turn this into a tool for public health officials to monitor which populations are at the highest risk for a tipping point,” Bauch emphasized. “Applied mathematics can be a powerful quantitative tool aiding in predicting, monitoring, and addressing threats to public health.”
This research is part of Waterloo’s broader commitment to evidence-based decision-making and public trust in science. It aligns with the University’s Societal Futures network and its TRuST initiative, which brings together philosophers, computer scientists, communicators, and ethicists to explore why trust in science and how it can be rebuilt.
“Rumours spread through populations much like contagious diseases–from one person to the next. Some people might be ‘immune’ to rumours, while others are ‘susceptible’ to them, according to their worldview and opinions,” Bauch told The Debrief.
Fundamentally, this type of predictive work could help save lives and money in healthcare systems around the world. Given the growing number of capabilities organizations and government groups have at their disposal with AI, the ability to predict where and when a disease outbreak might occur is within our control more now than ever before.
The study, “Forecasting infectious disease outbreak risks from vaccine sentiments on social media: A Data-driven dynamical systems approach”, was recently published in Mathematical Biosciences and Engineering.
Chrissy Newton is a PR professional and the founder of VOCAB Communications. She currently appears on The Discovery Channel and Max and hosts the Rebelliously Curious podcast, which can be found on YouTube and on all audio podcast streaming platforms. Follow her on X: @ChrissyNewton, Instagram: @BeingChrissyNewton, and chrissynewton.com.