County-level temporal changes and correlates of six categories of alcohol-attributable mortality before and during the early phases of the COVID-19 pandemic.

Early during the COVID-19 pandemic, many U.S. adults changed their alcohol use, and mortality due to causes fully attributable to alcohol increased. We assessed county-level changes in and correlates of six types of fully and partially alcohol-attributable mortality causes before and during the early phases of the pandemic. We used Bayesian spatial regression models to study six categories of alcohol-attributable mortality in 2019 and 2020 in 3,107 counties in the contiguous U.S. by season. Outcomes included chronic fully alcohol-attributable deaths, poisonings, motor vehicle traffic accidents, suicides, homicides, and falls. Covariates included year, season, rurality, region, and socioeconomic and demographic county-level characteristics. Crude county-level rates and counts increased for chronic fully alcohol-attributable deaths, poisonings, falls, and homicides and decreased for motor vehicle accidents and suicides from the spring of 2019 to the spring of 2020. In adjusted multivariable models, compared to the spring of 2019, the spring of 2020 was associated with elevated relative rates of chronic fully alcohol-attributable deaths, poisonings, and homicides, and decreased relative rates of deaths caused by suicide, motor vehicle accidents, and falls. While some covariate relationships were comparable across outcomes (e.g., higher median household income was negatively associated with most outcomes), others varied by outcome (e.g., greater proportions of older populations were positively associated with falls but negatively associated with most other outcomes). Alcohol-attributable mortality trends from before to during the early phases of the COVID-19 pandemic varied by specific cause. Interventions taking county-specific conditions into account may be more effective in decreasing alcohol-attributable mortality than one-size-fits-all approaches.