Spatial patterns of COVID-19 and non-COVID-19 mortality across waves of infection in England, Wales, and Scotland.

Recent studies have established the key individual-level risk factors of COVID-19 mortality such as age, gender, ethnicity, and socio-economic status. However, the spread of infectious diseases is a spatial and temporal process implying that COVID-19 mortality and its determinants may vary sub-nationally and over time. We investigate the spatial patterns of age-standardised death rates due to COVID-19 and their correlates across local authority districts in England, Wales, and Scotland across three waves of infection. Using a Spatial Durbin model, we explore within- and between-country variation and account for spatial dependency. Areas with a higher share of ethnic minorities and higher levels of deprivation had higher rates of COVID-19 mortality. However, the share of ethnic minorities and population density in an area were more important predictors of COVID-19 mortality in earlier waves of the pandemic than in later waves, whereas area-level deprivation has become a more important predictor over time. Second, during the first wave of the pandemic, population density had a significant spillover effect on COVID-19 mortality, indicating that the pandemic spread from big cities to neighbouring areas. Third, after accounting for differences in ethnic composition, deprivation, and population density, initial cross-country differences in COVID-19 mortality almost disappeared. COVID-19 mortality remained higher in Scotland than in England and Wales in the third wave when COVID-19 mortality was relatively low in all three countries. Interpreting these results in the context of higher overall (long-term) non-COVID-19 mortality in Scotland suggests that Scotland may have performed better than expected during the first two waves. Our study highlights that accounting for both spatial and temporal factors is essential for understanding social and demographic risk factors of mortality during pandemics.

Infection rates from Covid-19 in Great Britain by geographical units: A model-based estimation from mortality data.

This study estimates cumulative infection rates from Covid-19 in Great Britain by local authority districts (LADs) and council areas (CAs) and investigates spatial patterns in infection rates. We propose a model-based approach to calculate cumulative infection rates from data on observed and expected deaths from Covid-19. Our analysis of mortality data shows that 7% of people in Great Britain were infected by Covid-19 by the last third of June 2020. It is unlikely that the infection rate was lower than 4% or higher than 15%. Secondly, England had higher infection rates than Scotland and especially Wales, although the differences between countries were not large. Thirdly, we observed a substantial variation in virus infection rates in Great Britain by geographical units. Estimated infection rates were highest in the capital city of London where between 11 and 12% of the population might have been infected and also in other major urban regions, while the lowest were in small towns and rural areas. Finally, spatial regression analysis showed that the virus infection rates increased with the increasing population density of the area and the level of deprivation. The results suggest that people from lower socioeconomic groups in urban areas (including those with minority backgrounds) were most affected by the spread of coronavirus from March to June.