Spatio-temporal analysis of covid-19 pandemic in Bangladesh

Abstract

The world has been experiencing the shock of Novel Coronavirus (COVID-19) with multiple waves. Bangladesh, like other countries, is witnessing pandemic waves as of March 8, 2020. Within three successive waves in Bangladesh, approximately two million cases of COVID-19 infection and thirty thousand deaths have been reported.Various demographic, climatic, and hospital factors have already been proven as linked to the infection scenario in the literature. Implemented aspatial countermeasures, according to some studies, could also have an impact on the pandemic situation. The regional and temporal growth of pandemic waves could be regulated based on these features. Although many studies have been conducted around the world, the majority of them have focused on the initial COVID-19 outbreak, projections of spatial diffusion and temporal trends during this time, mapping of infected cases related to public movement, and public perception of the pandemic, and advancements in healthcare technology. To the best author’s knowledge, there are still few studies comparing the spatio-temporal dynamics of waves and emphasizing both the temporal and spatial aspects. In this context, this study aims to look at how spatial factors affect the COVID-19 pandemic's temporal variation in Bangladesh. Under this aim, the temporal patterns of the three successive waves of COVID-19 are compared from both spatial and aspatial perspectives. The effects of spatial features on waves are investigated later. In this research, the incident rate is considered to represent the spread.Sixty-four districts of Bangladesh comprise the study area for this research.The data on district-wise confirmed cases and the daily number of tests of three waves have been collected from the Directorate General of Health Services (DGHS).Additionally, information on demographic characteristics, healthcare facilities, infrastructure facilities, geographic features, meteorological attributes, and economic attributes have been obtained from other secondary sources. Sen’s slope estimation, the Pettit test, and the Mann-Kendall (M-K) test have been used to examine the temporal pattern. Global and local spatial autocorrelation methods (Moran’s I) have been used to figure out the spatial pattern. Scan statistics have been employed in conjunction with a discrete Poisson probability approach to comprehend the spatiotemporal growth of the outbreak. Finally, the Ordinary Least Squares (OLS) model and spatial regression models have been used to identify the effects of spatial components. Bangladesh as a whole exhibited no statistically significant trend throughout the first and second waves, but the deployment of long-term spatial containment measures caused an abrupt drop in infection during the first wavein almost all districts. The Bangladesh government implemented many of these non-pharmaceutical measures during the initial wave, including the closure of educational institutions and social gathering places, lockdown, social distancing measures, case isolation, demarcating of the most infected regions, and movement restrictions.The third wave had a noticeable decline in incident rates and lasted for a shorter period. In all waves, Dhaka and the neighboring areas were the first hit and were infected more heavily. Due to the arrival of new virus strains in the surrounding nations, the Western area of the country suffered a cluster of greater COVID-19 occurrences during the second and third waves.In terms of infection, theEastern portion, including the districts of Sylhet and Mymensingh division, were always in a better state due to their local geographic features. The country’s testing facilities were unevenly distributed, which had an impact on the number of confirmed cases in areas. May to July is identified as the period of time when the infection was most likely to spread. Urban population density, poverty rate, the number of ICU beds in each district, and the population over 65 are all identified as contributing spatial features during the first wave. On the other hand, factors that are determined to be significant in the second wave include rainfall, the number of bus stops, and the distance from the most affected area. The third wave highlights important contributions of urban population density, hospitals per thousand people, temperature, and rainfall. To conclude, it is anticipated that the findings of this study could help to formulate an appropriate pandemic management framework that takes into account seasonal and regional factors to combat future outbreaks.