Spatio-Temporal Patterns of the COVID-19 Pandemic in Bangladesh

Abstract

Coronavirus disease was first reported in Bangladesh in March, 2019. As of 31 July 2020, there were about 2.4 million confirmed cases and more than three thousand deaths in the country and this number is expected to increase rapidly. It is crucial to detect existing hotspots, their evolution over time and emerging high-risk areas of the COVID-19 pandemic for the country to optimize resource allocation, facilitate informed decision making as the pandemic prevails. Using daily confirmed case data at district level, present study detects statistically significant clusters of coronavirus cases across the country between 8 March-27 June, 2020. Considering the geographical adjacency, population data, physical infrastructure, healthcare facilities and local awareness; these clusters have been identified. Global and Local Moran’s I statistics measuring spatial autocorrelation are applied to unfold the spatial pattern and corresponding dynamics of COVID-19 spreading. Two nonparametric tests namely the Mann-Kendall (M-K) and Pettit tests are used to trace the temporal pattern and detect abrupt changes in the time series data. Effectiveness of applied government intervention is reviewed in conjunction with the spatio-temporal analysis. It is believed that these results can guide policy makers and local administrators to administer stricter interventions on public movement and social gathering, formulate socio-economic management strategies; and health officials to deploy resources for disease management. To the best of authors’ knowledge, this study is the first of its kind that analyses spatio-temporal patterns applying twelve important disease related spatial attributes to monitor COVID-19 in a developing country like Bangladesh.