CT02 - MEPI-01

MEPI Subgroup Contributed Talks

Thursday, July 17 at 2:30pm

SMB2025 SMB2025 Follow
Share this

Zitao He

University of Waterloo
"Leveraging deep learning and social heterogeneity to detect early warning signals of disease outbreaks"
Identifying early warning signals (EWS) of shifts in vaccinating behaviors can be helpful in predicting disease outbreaks. Evolutionary game theory has been used to model individual vaccination decisions, while bifurcation theory has identified statistical EWS, such as increasing variance and lag-1 autocorrelation, near critical transitions. However, these conventional methods often struggle with noisy data. In this study, we improve coupled behavior-disease models by incorporating population heterogeneity, distinguishing between social media users and non-users, and examining the role of homophily in shaping disease dynamics. We develop deep learning classifiers, including Long Short-Term Memory (LSTM) and Residual Neural Networks (ResNet), trained on simulated data from a stochastic coupled model with Lévy noise that captures the heavy-tailed fluctuations characteristic of real-world systems. Our results show that these models outperform traditional statistical indicators in both sensitivity and specificity while offering clearer interpretability on empirical data. This approach provides a robust framework for detecting EWS and improving outbreak prediction, highlighting the power of deep learning in real-time public health monitoring.



Soyoung Kim

National Institute for Mathematical Sciences (NIMS)
"Optimizing Vaccine Efficacy Trials for Emerging Respiratory Epidemics: A Mathematical Modeling Approach"
Evaluating vaccine efficacy (VE) during emerging epidemics is challenging due to unpredictable transmission dynamics. An age-structured SEIAR compartmental model was developed using South Korea’s 2022 population and parameters from COVID-19 and the 2009 H1N1 pandemic to optimize RCT timing and sample size. Simulations varied trial initiation (±10%, ±20%, ±30% of the epidemic peak), follow-up (4–12 weeks), recruitment (2–12 weeks), and non-pharmaceutical interventions (10–20%). Results showed that VE remained stable, but sample size requirements fluctuated, decreasing post-peak before rising sharply. Starting trials 30% before the peak with extended recruitment minimized sample sizes without compromising power. NPIs expanded trial feasibility, and sample size estimates from simulated placebo cases maintained >85% power, avoiding under- or over-powering. This model provides a framework for designing adaptive and efficient vaccine trials in future respiratory epidemics.



Jonggul Lee

National Institute for Mathematical Sciences
"Quantifying Shifts in Social Contact Patterns: A Post-Covid Analysis in South Korea"
Social contact patterns are crucial for understanding infectious disease transmission, but detailed data has been scarce in South Korea. We conducted a two-week survey covering various periods, including school vacations and holidays. Participants provided information on their contacts, including location, duration, frequency, and characteristics of the contact person. Analysis of the data from 1,987 participants revealed 133,776 contacts, averaging 4.81 contacts per person daily. Contact numbers varied by age, household size, and time period. The highest number of contacts was observed in the 5-19 age group, lowest in the 20-29 group, and gradually increased up to the 70+ group. Larger households tended to have more contacts. Contact patterns differed significantly across time periods. Weekdays during the school semester showed the highest number of contacts, followed by weekdays during vacations, the Lunar New Year holidays, and weekends. During the Lunar New Year, there was an increase in contacts with extended family members, enhancing subnational social mixing. These findings provide valuable insights into social contact patterns in South Korea, which can improve our understanding of disease transmission and aid in developing region-specific epidemiological models.



Alexander Meyer

University of Notre Dame
"Estimating pathogen introduction rates from serological data to characterize past and future patterns of transmission"
The unpredictable timing of infectious disease outbreaks poses significant challenges for public health preparedness. For many pathogens, this unpredictability is due to uncertainty regarding introduction rates—the frequency with which the pathogen is introduced into at-risk populations. We present three model-driven advances toward quantifying pathogen introduction rates and their effects on outbreak timing and size. Our method relies on the assumption that pathogen introductions can only cause large outbreaks when population immunity is sufficiently low (i.e., the reproduction number R(t) > 1). First, we demonstrate that, for pathogens that cause lifelong immunity, introduction rates can be estimated from age-structured serological data. Second, we estimate annual rates of chikungunya virus (CHIKV, a mosquito-borne pathogen) introductions into 17 populations in Africa and Asia using serological data collected between 1973 and 2015. Our median estimates ranged from 1 to 70 CHIKV introductions per 10 million people per year. Finally, we used simulations to show how the introduction rate of a pathogen can shape its transmission patterns over time in affected populations. A lower introduction rate allows population immunity to wane between introductions, leading to large but infrequent outbreaks. In contrast, a higher introduction rate causes frequent low-level transmission, resulting in elevated population immunity that precludes large outbreaks. Together, these results illustrate how age-structured serology, a common type of epidemiological data, can be leveraged to better understand both historical and future transmission patterns in different populations.



Andrew Omame

York University Toronto, Canada
"Pre-exposure vaccination in the high-risk population is crucial in controlling mpox resurgence in Canada"
As mpox spread continues across several endemic and non-endemic countries around the world, vaccination has become an integral part of the global response to control the epidemic. Some vaccines have been recommended for use against mpox by the World Health organization (WHO). As the roll-out of mpox vaccines continue across the globe, it is imperative to develop mathematical models to support public health officials and governments agencies in optimizing vaccination strategies to curtail the resurgence of mpox. In this article, we develop a compartmental mathematical model to investigate the impact of vaccination in controlling a potential mpox resurgence in Canada. The model categorizes individuals into high- and low-risk groups and incorporates pre-exposure vaccination in the high-risk group and post-exposure vaccination in the high- and low-risk groups. The vaccine-free version of the model was calibrated to the daily reported cases of mpox in Canada from April to October 2022, from which we estimated key model parameters, including the sexual and non-sexual transmission rates. Furthermore, we calibrated the full model to the daily reported cases of mpox in Canada in 2024, to estimate the current mpox vaccination rates in Canada. Our results highlight the importance of pre-exposure vaccination in the high-risk group on controlling a potential resurgence of mpox in Canada, and the minimal effects of post-exposure vaccination in the high- and low-risk groups on the outbreak. In addition, our model predicts the possibility of mpox becoming endemic in Canada, in the absence of pre-exposure vaccination in the high-risk group. Overall, our modeling result suggests that pre-exposure vaccination in the high-risk group is crucial in controlling mpox outbreak in Canada. Stepping up this vaccination is sufficient to avert a potential mpox resurgence in Canada.



Rosemary Omoregie

University of Benin, Nigeria
"Mathematical Model For Dengue and its Co-Endemicity with Chikungunya virus"
A deterministic nonlinear mathematical model describing the population dynamics for Dengue and Chikungunya virus taken into consideration the effect of misdiagnosis due to the co-endemicity of the two viruses in the human population. It is necessary to understand the most important parameters involved in their dynamics that may help in developing strategies for prevention, control and joint treatments. The model is rigorously analyzed qualitatively and thresholds for eradication are established.



Binod Pant

Northeastern University
"Could malaria mosquitoes be controlled by periodic release of transgenic mosquitocidal Metarhizium pingshaense? A mathematical modeling approach"
Mosquito-borne diseases, such as malaria, remain a major global health challenge, necessitating the exploration of innovative vector control strategies. Naturally occurring entomopathogenic fungi have been shown to reduce mosquito lifespan, but their slow-acting nature has limited their practical application. Advances in biotechnology have led to the development of transgenic fungus strains (this study will focus on Metarhizium pingshaense strain) engineered to express insecticidal toxins, significantly increasing their efficacy against malaria vector mosquitoes. To our knowledge, this is the first deterministic model designed to assess the impact of fungal-based mosquito control. The proposed model accounts for multiple transmission pathways of the fungal infection, including mating-based transmission from infected males to females and indirect transmission via contact with infectious mosquito carcasses. The model is analyzed to determine equilibrium states, local stability conditions, and the reproduction number. Numerical simulations explore various release scenarios, evaluating the effectiveness of periodic versus continuous fungal release in different ecological settings. The results indicate that transgenic Metarhizium pingshaense has the potential to significantly reduce mosquito populations, particularly when release strategies are optimized.



Soyoung Park

University of Maryland
"Mathematical assessment of the roles of vaccination and Pap screening on the incidence of HPV and related cancers in South Korea"
Human Papillomavirus (HPV) is a major sexually-transmitted infection that causes various cancers and genital warts in humans. In addition to accounting for about 99% of cervical cancer cases, it significantly contributes to anal, penile, vaginal, and head and neck cancers. Although HPV is vaccine-preventable (and highly efficacious vaccines exist for preventing infection with some of the most oncogenic HPV subtypes in the targeted population), the disease continues to cause major public health burden globally (largely due to inequity in access to the main control resources (i.e., access to Pap smear and vaccination) and low vaccination coverage in most communities that implement routine HPV vaccination). This lecture is based on the use of a new mathematical model (for the natural history of HPV, together with the associated neoplasia) for assessing the combined population-level impacts of Pap cytology screening and vaccination against the spread of HPV in a heterogeneous (heterosexual and homosexual) population. The model, which takes the form of a deterministic system of nonlinear differential equations, will be calibrated and validated using HPV-related cancer data from South Korea. Theoretical and numerical simulation results will be presented. Conditions for achieving vaccine-derived herd-immunity threshold (for achieving HPV elimination in Korea) will be derived.



SMB2025
#SMB2025 Follow
Annual Meeting for the Society for Mathematical Biology, 2025.