Mathematical modeling plays a growing role in infectious disease epidemiology, for studying the dynamics of pathogen invasion and persistence, understanding the determinants of disease disparities among populations, and predicting the impact of interventions. Deterministic compartmental models (based on ordinary differential equations) have been the traditional basis for this work during the past three decades. However, recent advances in statistical theory and methods have given rise to a new class of stochastic network models that provide an integrated framework for the statistical estimation of generative contact network parameters and the stochastic simulation of dynamic networks and transmission processes. Stochastic network models are more appropriate when infection is spread by a small number of highly structured contacts, as with HIV and other STIs, or for small scale assessments, where the effects of chance lead to wide variation in potential outcomes.
This course will start with a brief review of traditional compartmental (SI, SIR, SIS) models, and the methodology for classical descriptive network analysis. It will then provide a thorough introduction to the new statistical methods for network analysis: Exponential family Random Graph Models (ERGMs). The course will conclude with instruction on using these methods to develop an integrated framework for stochastic network models for epidemics, with a focus on empirical models of HIV transmission and control.
This will be a “hands-on” course, with integrated lectures, example-driven computer lab sessions, and extensive tutorial materials. The labs will develop programming skills in statnet and EpiModel, an integrated suite of R-based software packages that simplify statistical network analysis, simulation and visualization, and provide built-in functionality for both deterministic compartmental and stochastic network modeling of epidemics. Participants will learn to estimate and evaluate generative network models from empirical data, use the estimated models to simulate dynamic transmission networks consistent with the data, analyze the results (using all the functionality of R packages) and construct network movies that show the infection spreading on the network.
Target Audience: Researchers and PhD students from any field who have an interest in epidemic modeling
Apply for Course: https://catalyst.uw.edu/webq/survey/morrism/225354
Deadlines:
