Comparing the Performance of Count Regression Models to Assess the Impact of Climate on COVID-19 Incidence in Dhaka, Bangladesh

Abstract

COVID-19 transmission has been a significant public health issue in Bangladesh since March 8, 2020. Environmental factors such as temperature, humidity, wind speed, rainfall, and visibility are thought to influence the rapid spread of the virus. This study aims to compare various count regression models to explore the relationship between these environmental factors and COVID-19 incidence. We focused on the Negative Binomial, Discrete Lindley, and Discrete Weibull regression models due to the over-dispersed nature of the COVID-19 data. Our analysis indicated that the Discrete Weibull regression model provided the best fit, as determined by AIC and dispersion values. Diagnostic plots confirmed that this model met all necessary assumptions. Additionally, a simulation study with three different scenarios was conducted to validate our findings from the real COVID-19 data. Our analysis revealed that minimum temperature and visibility are positively associated with COVID-19 transmission, while maximum temperature and humidity show a negative correlation. These insights enhance our understanding of how environmental factors impact COVID-19 outbreaks in Dhaka, offering valuable guidance for developing effective strategies to mitigate transmission in Bangladesh.

Vol. 24(2)s (Special Issue), December, 2024, pp 149-162