Spatial Dynamics and Multiscale Regression Modelling of Population Level Indicators for COVID-19 Spread in Malaysia.

ABSTRACT

As COVID-19 dispersion occurs at different levels of gradients across geographies, the application of spatiotemporal science via computational methods can provide valuable insights to direct available resources and targeted interventions for transmission control. This ecological-correlation study evaluates the spatial dispersion of COVID-19 and its temporal relationships with crucial demographic and socioeconomic determinants in Malaysia, utilizing secondary data sources from public domains. By aggregating 51,476 real-time active COVID-19 case-data between 22 January 2021 and 4 February 2021 to district-level administrative units, the incidence, global and local Moran indexes were calculated. Spatial autoregressive models (SAR) complemented with geographical weighted regression (GWR) analyses were executed to determine potential demographic and socioeconomic indicators for COVID-19 spread in Malaysia. Highest active case counts were based in the Central, Southern and parts of East Malaysia regions of Malaysia. Countrywide global Moran index was 0.431 (p = 0.001), indicated a positive spatial autocorrelation of high standards within districts. The local Moran index identified spatial clusters of the main high-high patterns in the Central and Southern regions, and the main low-low clusters in the East Coast and East Malaysia regions. The GWR model, the best fit model, affirmed that COVID-19 spread in Malaysia was likely to be caused by population density (ß coefficient weights = 0.269), followed by average household income per capita (ß coefficient weights = 0.254) and GINI coefficient (ß coefficient weights = 0.207). The current study concluded that the spread of COVID-19 was concentrated mostly in the Central and Southern regions of Malaysia. Population's average household income per capita, GINI coefficient and population density were important indicators likely to cause the spread amongst communities.