Abstract
•Global spatial analysis does not accurately represent the key properties of the spatial distributions of parasitic diseases.•Estimates derived from global spatial models are inadequate for describing and depicting local relationships between climatic factors and parasitic diseases.•Our local spatial model is adequate for a proper investigation of the parasitic diseases.•Local estimates produce a map which has important implications for parasitic disease control programmes in the study region.
[Display omitted]
This paper describes spatial distribution of Visceral Leishmaniasis (VL) and determines its correlation with climatic factors in an endemic focus in northern and central Tunisia. Data on VL cases in children under five years of age were obtained by consulting medical reports from all Tunisian Pediatric Departments (TPD) during 2006–2016. Three key climatic factors, namely precipitation, continentality index and pluviometric coefficient of Emberger were used as predictor variables to model the VL geographical distribution. Data handling and statistical analysis were performed using R and Arcview GIS software systems. Bayesian local spatial model was employed to analyse the data. The results show a progressive increase in the VL incidence rates in regions with high levels of precipitation, but with low values of both continentality index and pluviometric coefficient of Emberger. A likely explanation of these findings arises from the opposite local effects of climatic factors which tend to cancel each other out in the calculation of the mean parameter estimate over the whole study area. We conclude that using non-local spatial analysis approach leads to misleading epidemiological interpretations, which in turn are of relevance for more efficient and cost-effective resource allocation for control and well manage the spread of VL in the study region and elsewhere in Tunisia.