Immune response, transmission model, and contact network determinants of disease spread

Research Project 3 (RP3) proposes to develop a multiscale framework for modeling infectious disease transmission on contact networks that incorporates three critical sources of heterogeneity often overlooked in existing epidemic models - susceptibility, transmissibility, and contact patterns. First, individuals vary markedly in their susceptibility to infection based on their previous immune system exposures, nutritional status, and other variables. Second, the immune response influences viral replication rate, resulting in some individuals serving as superspreaders. Third, symptoms arising from immune activation after infection, such as lethargy, can induce individuals to change their behavior in the context of an outbreak in ways that influence the contact network. Hence, incorporating these factors will result in more realistic models of disease transmission to guide outbreak control. The overall goal of RP3 is to develop a generalizable framework for immune structured epidemic models on dynamic contact networks. In Aim 1, we propose to develop ordinary differential equation (ODE) models for immune control of viral replication. Model parameters determining the rate of viral replication and the efficacy of immune response will be estimated using viral load measurements and immune assays collected over the course of the disease progression. Outputs from this individual-scale model will affect susceptibility (prior vaccination reduces peak viral load and disease duration) and contact patterns (symptomatic individuals have fewer dynamic contacts). In Aim 2, we will use observed social mixing patterns and social determinants of health (SDOH) data to create synthetic populations and networks, model sampling constraints in partially observed populations, and build agent-based models (ABM) of epidemic spread. In Aim 3, we will develop a novel continuous time Markov chain (CTMC) model with individual heterogeneity that bridges model outputs from Aims 1 and 2 to predict disease outbreak dynamics. The proposed CTMC model overcomes challenges of working with discretely observed data and uses a latent variable framework to convert otherwise intractable equations into a feasible Bayesian sampling problem. This approach enables the estimation of model parameters and the validation of uncertainty estimates using simulation studies and standard perturbation analyses. Overall, RP3 proposes to develop a comprehensive approach for modeling disease spread that accounts for susceptibility, transmissibility, and contact patterns, and delivers a flexible and modular framework for incorporating existing knowledge and data sources on disease outbreaks. Individual, population, and bridging models for disease transmission will be built, and applied to case studies for HIV-1 (sexual transmission) and SARS-CoV-2 and inlfuenza (airborne transmission). By estimating parameters and providing uncertainty quantification within a multiscale mechanistic framework, our method leads to interpretable conclusions about outbreak size and duration in immune-heterogeneous populations that can be used to evaluate 'what-if' scenarios for policy generation, including for potential new disease outbreaks.