Development and use of open-source algorithms for space-time emerging hotspot analysis of routine dengue NVBDCP data in Punjab, India

Abstract

Background: Understanding spatiotemporal epidemiology using open-source and reproducible algorithms add value to routine health information systems. Objectives were to estimate spatial clustering, identify spatial clusters and space-time hotspots of dengue.Methods: Queen’s contiguity neighborhood matrix and row-standardized spatial weights were used. Spatial clustering was estimated using Moran’s I. Local Moran’s I with sensitivity analysis at 0.01, 0.05, and 0.1 significance levels were performed. The space-time cube model was developed. Gi* statistic and seasonal Mann Kendal test identified persistent and intensifying, persistent, persistent and diminishing, emerging, oscillating, new, historical, and sporadic hotspot sub-districts. Analysis was carried out using R version 4.1.0.Results: The expected Moran’s value was -0.00671. Significant spatial clustering was observed annually in 2016-2018 (p<0.01, <0.01, and 0.04, respectively) and was most common in August, followed by July and November. High-high, high-low, low-low, and low-high sub-district clusters were identified between Aug-Dec from 2015-19. Sensitivity analysis highlighted the core and spread of spatial clusters. Faridkot and Muktsar blocks/ sub-districts were persistent and intensifying hotspots.Conclusions: Spatial clusters were dynamic in space and time. The development of open-source algorithms provides a reproducible and scalable platform for future research and evidence for informed decision-making by public health managers.