The Limits of Sexual Network Data: Implications for Mathematical Modelling of STI

Tuesday, November 18, 2003 - 9:30am - 10:20am
Keller 3-180
Marie-Claude Boily (Laval University)
Empirical and theoretical studies have highlighted the importance of the local (egocentric) and the global (sociocentric) network structure on the individual risk of infection and the spread of diseases in populations. Transmission dynamics models of STI (sexually transmitted infection) and HIV/AIDS have been instrumental in highlighting the importance of quantifying sexual behaviour such as the average and variance in sexual activity, the mixing pattern, concurrency and others in order to understand epidemiological trends. Detailed individual based models of partnership formation and dissolution (micro-simulation or network models) are increasingly being used in order to capture the full complexity of sexual networks and more realistically simulate the course of STI and HIV/AIDS. However, at the moment, the complexity of models has outstripped the level of behavioural data currently available for most populations.

The objective of the talk is to review and discuss the limits of currently available sexual network data and the implication for mathematical modelling of STI. It will be shown how the limited network data available, compounded by our incomplete understanding of individual behaviour, has a number of consequences for the formulation and validation of network models, the interpretation of model results, and the formulation of research questions and data collection. In the case of a lethal disease, like HIV/AIDS, not only is it important to understand the impact of the network structure on the spread of disease, but is it also important to understand and assess the impact of the spread of disease on the network structure. It will be shown that this is particularly relevant to understand the different impacts of antiretroviral therapy or vaccination in heterogeneous populations largely afflicted by the epidemic.