Modeling the influence of settlement structure on the spread of influenza among communities.

Several factors play critical roles in the geographic spread of infectious diseases, including the movement of people between communities and the social and economic structure of groups of communities. A mathematical model has been developed to examine the individual and shared impact of these factors. This model was applied to the spread of the 1918-1919 influenza epidemic in three Aboriginal communities in central Canada. Previously published results from simulations of the model used parameter estimates for mobility patterns that were derived from historical documents from the study communities, especially Hudson's Bay Company post journals. We have termed this model the Frontier pattern. In this paper we extend that work to consider the consequences of three additional travel scenarios, which represent idealized trade and travel relationships between different communities. The three patterns include (a) the Central Marketplace scenario, an idealization of relations between a central community and its satellites in which a single central community satisfies all needs of the satellite communities; (b) the Sister Towns scenario, which allows travel among all communities and overemphasizes communication between satellites so that the ramifications of small town interactions are easier to observe; and (c) the Circuit scenario, which represents a hypothetical trade network that is a series of unidirectional dyad relationships linking all communities within a region. Results from simulations of the four mobility patterns are compared to assess the potential impact different social and economic relationships among communities may have had on the geographic spread of the 1918 influenza epidemic in central Canada and in other regions. These results reinforce conclusions that show that variation in patterns of mobility significantly influences the timing of epidemic peaks but only minimally alters the number of cases within a community. In addition, comparisons of the different models indicate that a central location in the social and political hierarchy of a region may be more important in influencing patterns of epidemic spread than just a central location with regard to travel patterns. Furthermore, who was actually travelling and where they were travelling had important consequences for epidemic spread. Am. J. Hum. Biol. 12:736-748, 2000. Copyright 2000 Wiley-Liss, Inc.