Associations between social determinants of health and COVID-19 outcomes by variants of concern: a Bayesian spatiotemporal analysis.

BACKGROUND

Socially and economically disadvantaged populations were disproportionately affected by COVID-19, but it remains unclear whether and how variants of concern moderated the associations between social determinants of health and COVID-19 outcomes.

METHODS

We aggregated cases by week and health region (n=231) from 1 January 2021 to 27 February 2022, covering four major variants of concern waves-Alpha, Gamma, Delta and Omicron-in British Columbia, the third most populated province in Canada. We constructed a region-specific, continuous socioeconomic index (SI) in which a one-unit increase represents a 10% increase in socioeconomic status across the health region. We fitted Bayesian spatiotemporal models on four COVID-19 outcomes separately: cases, hospitalisations, intensive care unit admissions and deaths, adjusting for interaction between SI and variants of concern, percentage male and vaccine coverage. We estimated the marginal rate ratio (RR) between SI and the outcomes by variants of concern.

RESULTS

Overall, Alpha, Gamma and Delta heightened the associations between SI and all outcomes. The marginal RR between SI and cases was 0.96 (95% credible interval (CrI): 0.93, 1.00) in the absence of variants of concern, 0.93 (95% CrI: 0.88, 0.97) in the presence of Alpha, 0.93 (95% CrI: 0.88, 0.98) for Gamma and 0.90 (95% CrI: 0.87, 0.93) for Delta. Omicron reversed the direction of SI's association with cases (RR: 1.03, 95% CrI: 0.99, 1.07) and dampened the associations with other outcomes.

CONCLUSIONS

Variants of concern generally widened the socioeconomic disparities in COVID-19 outcomes. Tailoring and optimising pandemic preparedness and response measures to the specific needs of disadvantaged populations is vital for reducing the additional disease burden experienced by some communities in British Columbia and across Canada.