Use of spatial analytics and geographic information systems in childhood diarrhoea in rural areas of India: A protocol paper

Abstract

Introduction and Aim: The most recent advancement in epidemiology is using Geographic Information Systems (GIS) to comprehend illness based on clustering and spatial analysis. Using spatial analysis has shown positive changes in sanitation, reducing diarrhoea's burden. This study put forwards the methodology used to analyze spatiotemporal clustering and hotspot areas of childhood diarrhea in Karkala taluk, Udupi during 2015 to 2018. Materials and Methods: This study describes the methods used to analyze the space-time clustering and local indicators of spatial association (LISA) of childhood diarrhoea using the GeoDa and SaTScan programmes. At the Karkala taluk, data on under-five diarrhoea was gathered throughout time and space to develop the geospatial models. While the timing of the onset of the diarrhoea impacts the temporal characteristics, the child's position at the time of the event determines the spatial characteristics. The child's location and time of diarrheal episode was considered for the geospatial analysis. Results: The Kulldorff scan statistic (SaTScan v10.1) was employed for geographical analysis, and GeoDa software used for LISA and hotspot analysis. The randomness distribution of childhood diarrhoea over time and space was determined using the Kulldorff scan statistic software. The SaTScan locates and evaluates the statistical significance of geographical clusters based on the number of cases inside and outside the cylinder window. The scanning window in the software uses a cylinder with an elliptical or circular base to study intervals for time, space, and space-time. The different combinations of High and Low categories of the spatial association are covered by the hotspot analysis utilizing the GeoDa software. Conclusion: To build data-driven public health decision-making, it is required to link spatial and temporal patterns of diarrheal incidence among under-five children with predictive variables derived from routine data.