The ongoing COVID-19 pandemic necessitates a comprehensive understanding of populations at highest risk of death. A lack of high-quality micro-level data linking death records to socio-demographic information has hindered this understanding. Previous research, often based on aggregated data or limited samples, suggests that men, the elderly, racial and ethnic minorities, and those in lower socioeconomic positions are disproportionately affected. This study leverages unique access to complete Swedish data—linking all COVID-19 deaths up to May 7, 2020, with high-quality individual-level background information—to rigorously examine socio-demographic risk factors for COVID-19 mortality in Sweden. Sweden's less restrictive approach to containing COVID-19, compared to many other nations, provides a valuable case study with broader implications for other countries navigating the pandemic. The study examines how the risk of death varies across age, sex, civil status, income, region of residence, and country of birth, aiming to identify vulnerable populations and inform national strategies.

Existing literature, while limited by data availability, suggests correlations between socioeconomic status and severe COVID-19 outcomes. Studies indicate that men, older individuals, racial and ethnic minorities, and those in lower socioeconomic positions exhibit higher risks of severe illness or death from COVID-19. However, these findings often rely on aggregated data or smaller studies, lacking the comprehensive population-level analysis offered by this study. The initial findings contrast with the notion that COVID-19 does not discriminate, highlighting the need for more detailed investigations into sociodemographic inequalities in mortality.

The study utilized individual-level Swedish register data encompassing 7,775,064 individuals aged 20 and above on March 12, 2020. Data included socio-demographic information (sex, age, civil status, income, education, country of birth, and region of residence) linked to death records (COVID-19 and all-cause mortality) between March 13 and May 7, 2020. Cox proportional hazards regression models were employed to analyze the risk of death, stratified by sex and age group (working age vs. retirement age) to account for potential interactions. The analysis adjusted for various factors including age and residence in Stockholm (the initial epicenter of the outbreak in Sweden). Robustness checks, including stratified analyses and comparison to logistic regression models, were performed to validate the findings. The proportional hazards assumption was tested, and potential violations addressed through stratified analyses.

During the study period (March 13-May 7, 2020), 17,181 deaths occurred, with 3126 attributed to COVID-19. Key findings from the multivariate Cox proportional hazards models revealed significant associations between several socio-demographic factors and COVID-19 mortality. Being male, having low individual income, low education, being unmarried, and being an immigrant from a low- or middle-income country (LMIC) were independently associated with increased risk of COVID-19 death. Importantly, these associations were largely consistent with patterns observed for all-cause mortality, with the exception of country of birth and county of residence. The study found a stronger relationship between socioeconomic factors and COVID-19 mortality among working-age individuals compared to retirees. Among working-age individuals, low income was strongly associated with COVID-19 mortality, and there was a pronounced income gradient among men, but not women. Conversely, among retirees, civil status was a more significant predictor, with unmarried elderly individuals facing higher risk of death. Immigrants from LMICs exhibited consistently higher COVID-19 mortality risk across both age groups, despite adjusting for socioeconomic factors and region of residence.

The findings demonstrate that the impact of COVID-19 is not uniformly distributed across the Swedish population. Disadvantaged groups experienced a disproportionately higher risk of death, consistent with broader trends observed in mortality from other causes. However, the elevated mortality among immigrants from LMICs, independent of socioeconomic factors, warrants further investigation. This suggests that factors beyond socioeconomic circumstances might play a role. Potential explanations could include differences in healthcare access, occupation, living conditions, or susceptibility to severe illness. The stark contrast between COVID-19 mortality and all-cause mortality concerning country of origin suggests the presence of factors specific to COVID-19. The pronounced socioeconomic disparities in COVID-19 mortality among working-age individuals highlight the potential for interventions aimed at protecting these vulnerable populations. Similarly, the increased risk among unmarried elderly individuals underscores the importance of adequate social support systems for this demographic.

This study provides compelling evidence that socio-demographic factors significantly influence COVID-19 mortality in Sweden, with disproportionate impact on disadvantaged groups. The findings suggest that future interventions should focus on addressing socioeconomic inequalities and improving healthcare access and social support for vulnerable populations, including immigrants from LMICs and unmarried elderly individuals. Future research should investigate the underlying mechanisms driving the observed disparities and explore the specific vulnerabilities of migrants to severe COVID-19 outcomes. The insights gleaned from this study are relevant for informing public health strategies in Sweden and other countries facing similar challenges.

The study's relatively short follow-up period (approximately two months) may limit the ability to fully capture the long-term impact of COVID-19. The study also lacks information on emigration from Sweden during the study period, potentially underestimating mortality rates. While the Swedish data on COVID-19 deaths is considered accurate, some misclassification of COVID-19 deaths cannot be entirely ruled out. Finally, the use of all-cause mortality as a comparison may capture indirect effects of COVID-19, like reduced healthcare access due to pandemic-related restrictions.