Epidemiological applications of contact and infection network data

This dissertation develops measures using epidemiologic contact and infection network information to analyze infection transmission systems. Simple models of infection transmission that allow control of mixing and basic population parameters are employed to generate infection transmission data from which contact and infection graphs are constructed and analyzed. Some of these infection network measures could potentially be estimated from nucleotide sequence data originating from sampled infectious agents circulating in a population.; A contact graph formulation is developed that accounts for the dynamic nature of contacts between individuals. We propose the use of "source counts" that summarize infection transmission potential from prior or concurrent relationships to a particular relationship or individual. Source counts can be used to assess changes in individual infection risk and to identify individuals that are essential for the maintenance of infection transmission.; We demonstrate conditions where established methods for estimating past population infection levels are problematic due to assumptions regarding population homogeneity. We conclude that convenience samples should be avoided when making these types of inferences from nucleotide sequence data.; Our most important results demonstrate that there is infection flow and population mixing information in the structure of the infection tree. We show that the distribution of coalescence events can be employed to identify geographic or social sources of infection when the source population is separated in the analysis. Among populations exhibiting constant prevalence, mixing between individuals determines the rate of coalescence to a single common infector as well as the proportion of short path lengths between pairs of individuals selected by risk status. These results provide evidence that nucleotide sequence data could elucidate information regarding infection transmission in populations.; We also present measures that cannot be observed or quantified in the field but may prove useful to investigators employing simulations to study transmission system dynamics. For example, the proportions of high and low risk common infectors for specific pairs of sampled nodes yields information regarding population mixing and infection flow. We believe that these results will facilitate the analysis of infection transmission through populations, especially in light of recent increases in the availability of nucleotide sequence data.