What question did this study set out to answer?

This analysis aims to understand the relationship between social conditions, infection burden, and COVID-19 mortality across Scottish councils during 2020-2021.

June 12, 2026Open Access

A Bayesian spatial ecological analysis of social conditions, infection burden and COVID-19 mortality across mainland Scottish councils 2020-2021

Key Points

This analysis aims to understand the relationship between social conditions, infection burden, and COVID-19 mortality across Scottish councils during 2020-2021.
Examined COVID-19 mortality across 29 councils in Scotland using a two-year panel of data.
Utilized a Bayesian Poisson log-linear model to estimate area-level relative risks based on infection and social indicators.
Analyzed factors like unemployment, marriage, smoking cessation, and migration in relation to mortality.
Crude mortality was higher in the central belt compared to northern and rural councils, although spatial differences became more modest after adjustment.
Positive COVID-19 test burden was directly associated with higher mortality rates, indicating its role as an infection burden indicator.
Higher unemployment was correlated with increased mortality risk, while more marriages were linked to lower mortality.

Abstract

The COVID-19 pandemic has generated substantial spatial and social inequalities in mortality, yet within-country council-level variations remain incompletely understood. Scotland offers a useful case, with a severe epidemic but rich administrative data and marked socio-economic gradients. We examined how infection burden, hospital-related indicators and selected area-level social indicators were associated with COVID-19 mortality across mainland Scotland 2020-2021. We assembled a two-year panel for 29 mainland councils, using deaths involving COVID-19, population denominators, COVID-19 testing and hospital activity, council-level indicators of marriage, unemployment, smoking cessation and migration. A Bayesian Poisson log-linear model with conditional autore- gressive random effects was used to estimate area-level relative risks and covariate associations. Given the short panel, the analysis is interpreted primarily as a spatial ecological analysis conducted over two successive pandemic years. Crude mortality showed consistently higher COVID-19 mortality in the central belt than in many northern and rural councils. After adjustment, spatial differences remained but were modest: no council in 2021 had a posterior probability above 0.10 of exceeding the national mean mortality risk. Positive test burden was positively associated with mortality and is interpreted primarily as a proximal epidemiological indicator of infection burden. Social determinants were also important. Higher unemployment was associated with increased risk, whereas higher marriage counts were linked to lower mortality. Smoking quit rates showed a positive association with deaths, likely reflecting residual confounding by underlying deprivation and historical smoking prevalence. Hospital utilisation and migration indicators showed weaker and more uncertain effects. Together, these findings indicate that crude spatial disparities in COVID-19 mortality across mainland Scottish councils became more moderate after adjustment. Given the ecological design and the two-year panel, the findings should be interpreted as area-level associations rather than as evidence of specific local social mechanisms.

KI fragen

Bookmark

View Full Paper