Abstract
BACKGROUND: Associations between alcohol and mortality are well documented in epidemiological studies, but whether brief alcohol screening measures administered routinely in primary care similarly predict mortality is less established. This study examined whether alcohol consumption reported on the three-item Alcohol Use Disorder Identification Test, Consumption (AUDIT-C) during routine care was associated with all-cause mortality, and secondarily, whether age and sex modified associations. METHODS: This retrospective cohort study used electronic health record data from a large integrated health system and death data from the Washington Department of Health. Adult (≥18 years) primary care patients were included if they completed an AUDIT-C screen (March 1, 2015-December 31, 2021). AUDIT-C scores were categorized: no alcohol use (0), low-risk alcohol use (1-2 female/1-3 male), moderate-risk alcohol use (3-6 female/4-6 male), high-risk alcohol use (7-8), or very high-risk alcohol use (9-12). Time-to-event (death or censoring) was measured for up to 8 years (2015-2023). Cox proportional hazards models, adjusted for demographic and clinical covariates, estimated adjusted all-cause mortality across AUDIT-C score categories, with low-risk alcohol use as the reference. RESULTS: Among 531,851 primary care patients, a majority were aged 30-59 (52%), female (58%), non-Hispanic (94%), and White (70%). A total of 21,548 patients died (4.8%) during follow-up (median 1752 days). Compared to patients reporting low-risk alcohol use, those reporting no use or very high-risk use had higher mortality (hazard ratio [HR] = 1.40 [95% CI: 1.36-1.44] and HR = 1.58 [1.35-1.84], respectively); those with moderate-risk use had lower mortality (HR = 0.86 [0.83-0.90]). Associations varied by age (stronger among young adults) but not sex. CONCLUSIONS: Alcohol use reported on brief screens during routine primary care had an approximately convex association (J- or U-shaped depending on sub-population) with all-cause mortality as observed in prior epidemiological studies. Findings support the value of routine alcohol screening as a tool for identifying patients at risk for major adverse health outcomes like death.