Insights into the microscale spatial dynamics of dengue and chikungunya in Southeast Asia

Dengue;Chikungunya;Spatial statistics;Clustering;Disease dynamics;Epidemiology

Salje, HenrikNorris, Douglas ;

Johns Hopkins University

关键词: Dengue;    Chikungunya;    Spatial statistics;    Clustering;    Disease dynamics;    Epidemiology;   

Others  :  https://jscholarship.library.jhu.edu/bitstream/handle/1774.2/59411/SALJE-DISSERTATION-2014.pdf?sequence=1&isAllowed=y

瑞士|英语

来源: JOHNS HOPKINS DSpace Repository

【 摘 要 】

The spatial dynamics of many diseases are generally studied at the macroscale, including the spread of pathogens between countries and continents. Disease dispersal within communities is less well understood. This gap is partly due to a lack of statistical approaches that can accurately characterize spatial and temporal dependence of disease processes in the presence of underlying spatial heterogeneities that can hide any signal. Here we developed approaches that estimate (a) the mean distance between sequential cases in a transmission chain and (b) spatial dependence between cases over different time-frames (irrespective of who infected whom) from point pattern incidence data.Importantly, our approaches are valid where we only observe a tiny fraction of infections and there exist both multiple overlapping transmission chains and spatial heterogeneities in disease surveillance. We demonstrated the robustness of our approaches using simulation. We then applied them to geocoded dengue case data from Bangkok, Thailand, a disease that has been in endemic circulation in this city for decades. We estimated that the mean transmission distance for dengue in the city was 50m (varying between 44m and 64mbetween 1994 and 2006). Further, the aggregation of short range individual transmissions led to the presence of larger scale spatial temporal dependence, with clustering of all cases within any month observed at distances up to 1km.We also observed patterns of spatiotemporal dependence consistent with the expected impacts of homotypic immunity, heterotypic immunity and immune enhancement of disease at these distances. Our observations indicate that individual transmissions (which encompass both human and mosquito movements) tend to be not be much further than neighboring households, however, immunological memory of dengue serotypes occurs at the neighborhood level in this large urban setting. Infections between neighboring households driving disease spread was also supported for chikungunya, a pathogen transmitted by the same mosquitoes as dengue: we estimated a mean transmission distance of 60m (95% confidence interval: 50m - 70m) from an outbreak of the virus in a village in Bangladesh. The findings presented here have broad implications for understanding the mechanisms of dengue and chikungunya dispersal, the tailoring of intervention measures and the parametrization of mathematical models of disease spread. In addition, the methods presented have wide-ranging application across disease systems.