New method can calculate transmission weeks in advance so public can better prepare for outbreaks
Using a novel forecasting approach that combines mathematics with artificial intelligence (AI), a group of Canadian and U.S. researchers have devised a way to predict the rate of transmission of infectious diseases, such as the seasonal flu and COVID-19.
According to the team’s calculations, new confirmed flu cases in U.S. laboratories are expected to exceed 1,600 per day by the end of November – almost double compared to the same time last year, which saw 955 new cases per day.
While the team – based at the Interdisciplinary Lab for Mathematical Ecology & Epidemiology (ILMEE) at the University of Alberta – found that transmission rates are traditionally highest in December, this year’s forecast anticipates an earlier flu season surge.
“The initial increase in cases this year will be slightly earlier than 2023, which peaked at the end of December with 3,299 new confirmed daily cases reported in U.S. laboratories,” said Hao Wang, ILMEE director and professor and Tier 1 Canada Research Chair in Mathematical Biosciences at the University of Alberta. “It’s important to know transmission patterns because if we can predict spikes in advance, public health officials can take preventative action to contain the spread of diseases.”
Hao Wang’s team includes Xiunan Wang, assistant professor in the Department of Mathematics at the University of Tennessee at Chattanooga. The duo co-authored a paper about the new forecasting approach in SIAM Journal on Applied Mathematics, a publication of Society for Industrial and Applied Mathematics (SIAM).
Hao Wang
Xiunan Wang
Using a discrete inverse method – which analyzes data from previous years to determine future outcomes – the researchers showed that the properties of both transmissibility and number of infections can be quite different.
“There can be significant delays between the peaks in transmission rates and the peaks in the number of infections,” said Xiunan Wang. She explained that the transmission rate refers to how quickly and efficiently a disease spreads within a population, representing the likelihood that an individual becomes infected after contact with an infected person. This rate captures the potential for rapid spread but does not directly reflect the visible number of infections, which may take longer to peak due to incubation periods, delayed symptom onset, or other external factors.
Infection data alone isn’t enough
“Public health officials typically provide infection and death data, which are important for tracking an epidemic’s impact,” Xiunan Wang said. “However, this data alone doesn’t offer a complete picture, especially when it comes to understanding the underlying transmission dynamics. To uncover more accurate transmission patterns, it’s crucial to analyze the time-varying transmission rates.”
Hao Wang explained that combining the team’s transmission data with other data-driven technologies – such as machine learning – can forecast future disease dynamics based on factors like weather conditions, policy decisions, or human mobility trends (data collected from cellphones), thus providing guidance to public health authorities about the implementation of more effective control strategies.
In this case, his team incorporated data from late 2015 to September 2024 in their mathematical model to generate a real-time transmission forecast of the number of projected infections and fatalities.
“Despite observing a significantly higher peak of flu infections in late 2022 in reports from the U.S. Centers for Disease Control and Prevention (CDC), our team incorporated historical U.S. hospitals flu data into our model and estimated that while transmission and fatality rates increased during the COVID-19 pandemic, they have remained relatively stable when comparing pre- and post-pandemic periods,” Hao Wang said.
“The transmission rate among individuals under 18 years old did not exhibit significant changes before and after the COVID-19 pandemic, which can be attributed to the fact that this age group primarily comprises students whose contact patterns remained largely unchanged,” he added. “In comparison, in looking at flu-caused fatality, the long COVID impact on seniors 65+ was most severe during 2021 to 2023, although the fatality rate returned to the pre-COVID level during 2023-2024.”
More information about the team’s forecasting method – which can also apply to diseases with non-seasonal cycles, such as measles – is available in the full article, which is temporarily free to read in SIAM’s publications library.
About Society for Industrial and Applied Mathematics (www.siam.org)
Society for Industrial and Applied Mathematics (SIAM), headquartered in Philadelphia, Pennsylvania, is an international society of more than 14,000 individual, academic and corporate members from 85 countries. SIAM helps build cooperation between mathematics and the worlds of science and technology to solve real-world problems through publications, conferences, and communities like chapters, sections, and activity groups. Learn more at siam.org.