Urban environments exhibit pronounced heterogeneity in environmental conditions, population density and wealth, whose effects on infectious disease dynamics remain poorly understood. A better understanding of the population dynamics of climate sensitive diseases such as malaria requires the consideration of heterogeneities in socio-economic and associated demographic conditions and their interaction with environmental variation in space and time. In concert these two dimensions, environmental and socio-economic, would define the relevant spatial scales at which to address transmission dynamics. Here we address the role of both climate and economic drivers and how their effects vary at different spatial resolutions by , taking advantage of a spatially and temporally resolved dataset of Plasmodium falciparum counts for the city of Surat, India. We combine statistical analyses and a Bayesian hierarchical mixed model framework to analyze the spatiotemporal pattern of urban malaria risk, and its association with socioeconomic indicators, population density, and the environmental parameters of temperature and humidity.
Results/Conclusions
Spatiotemporal malaria risk is largely stationary in time despite interannual variation, a finding demonstrating the importance of spatial drivers. Results of the generalized linear mixed model show the importance of population density, climate and socioeconomic factors, and spatially structured and unstructured random effects. Socioeconomic and demographic spatial variation can be summarized into three main axes of variation corresponding respectively to poverty and access to water, work environment and human mobility, and population density. Previously undescribed influences on urban malaria include positive effects of humidity and human density on per-capita disease risk, and a negative effect of high temperature. Humidity and temperature show contrasting spatial scales at which they most clearly influence malaria. Average humidity becomes most significant at aggregated scales influencing the interannual and seasonal variation of the disease, whereas temperature emerges as most important for spatial variation together with economic level as we disaggregate the system. An understanding of these effects provides a basis for more targeted intervention, such as vector control, based on transmission ‘hotspots’.