An urban-to-rural continuum of malaria risk: new analytic approaches characterize patterns in Malawi

• Published in: Malaria Journal
• Main Authors: Peter S. Larson, Joseph N. S. Eisenberg, Veronica J. Berrocal, Don P. Mathanga, Mark L. Wilson
• Format: Article in Journal/Newspaper
• Language: English
• Published: BMC 2021
• Subjects: Urbanicity; Environmental risk; Malaria prevention; Remote sensing; Spatial analysis; Arctic medicine. Tropical medicine; RC955-962; Infectious and parasitic diseases; RC109-216; Arctic
• Online Access: https://doi.org/10.1186/s12936-021-03950-5; https://doaj.org/article/9796675924a94a9499604697735faea2

Abstract Background The urban–rural designation has been an important risk factor in infectious disease epidemiology. Many studies rely on a politically determined dichotomization of rural versus urban spaces, which fails to capture the complex mosaic of infrastructural, social and environmental factors driving risk. Such evaluation is especially important for Plasmodium transmission and malaria disease. To improve targeting of anti-malarial interventions, a continuous composite measure of urbanicity using spatially-referenced data was developed to evaluate household-level malaria risk from a house-to-house survey of children in Malawi. Methods Children from 7564 households from eight districts throughout Malawi were tested for presence of Plasmodium parasites through finger-prick blood sampling and slide microscopy. A survey questionnaire was administered and latitude and longitude coordinates were recorded for each household. Distances from households to features associated with high and low levels of development (health facilities, roads, rivers, lakes) and population density were used to produce a principal component analysis (PCA)-based composite measure for all centroid locations of a fine geo-spatial grid covering Malawi. Regression methods were used to test associations of the urbanicity measure against Plasmodium infection status and to predict parasitaemia risk for all locations in Malawi. Results Infection probability declined with increasing urbanicity. The new urbanicity metric was more predictive than either a governmentally defined rural/urban dichotomous variable or a population density variable. One reason for this was that 23% of cells within politically defined rural areas exhibited lower risk, more like those normally associated with “urban” locations. Conclusions In addition to increasing predictive power, the new continuous urbanicity metric provided a clearer mechanistic understanding than the dichotomous urban/rural designations. Such designations often ignore urban-like, low-risk pockets ...