Ohio State University
Infectious disease epidemiology uses statistical methods borrowed largely from chronic disease epidemiology. These methods treat infections as independent events, ignoring the defining feature of infectious disease. Our research attempts to link stochastic epidemic models to the analysis of infectious disease data, including study design and causal inference. Throughout this project, we intend to apply the methods we develop to previous and ongoing research at the University of Washington and the International Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR.B). We must learn what is important to measure and then learn how to measure it. By developing methods for the analysis of stochastic epidemic models, we hope to distinguish between robust and parameter-sensitive behaviors of infectious diseases. We will then adapt methods from survival analysis to likelihoods derived from transmission models to develop methods for point estimation and hypothesis testing that account for the transmission of disease. Through simulation, we can investigate the behavior of these tests and estimators in situations common in real-world infectious disease data, such as unobserved infection times, asymptomatic infections, and seasonal or geographic variation. Since disease transmission can lead to complex time-dependent confounding, we must learn to identify situations where non-standard techniques of controlling confounding are necessary and adapt them specifically for infectious disease epidemiology. Finally, we will adapt cohort, case-control, and case-crossover study designs to obtain valid and efficient estimates of the effects of exposures on infectiousness, susceptibility, and other aspects of disease transmission, and we will explore new designs and randomization procedures for vaccine and intervention trials. Once adequately developed, all of these methods need to be implemented in computer programs that are accessible to epidemiologists, biostatisticians, and other public health practitioners. Despite the optimism of the mid-twentieth century, infectious diseases remain a tremendous burden and threat to public health worldwide. By developing methods for study design and statistical analysis that account for the transmission of disease, we hope to provide future epidemiologists and biostatisticians with tools for a much more detailed and accurate analysis of the mechanisms by which exposures and interventions affect the spread of infections. Detailed, place-specific knowledge of these mechanisms will allow more timely and effective local interventions, ultimately protecting global public health.