Health Deterioration Precedes Alcohol Cessation and Biases the Nondrinker Reference Group: Longitudinal Evidence from Two Cohorts of the Korea Health Panel.

Former drinkers are often grouped with lifetime abstainers as the nondrinker reference group in alcohol-related health studies. This classification has the potential to bias estimates when alcohol cessation follows health deterioration. This study examined whether adverse health status and healthcare burden preceded alcohol cessation and whether the inclusion of former drinkers altered estimates for current drinkers. Using two independent Korea Health Panel cohorts from 2010-2018 and 2019-2023, adults were classified as lifetime abstainers, former drinkers, or current drinkers. Former drinkers were further classified by self-reported abstinence duration. Unweighted repeated person-period logistic regression with individual-level cluster-robust standard errors was used to assess health status by drinking category, next-year transition from current to former drinking, and differences between conventional and corrected nondrinker reference definitions. In both cohorts, former drinkers, especially recent former drinkers, had higher odds of physical limitation, poor self-rated health, hospitalization, and high out-of-pocket medical spending than lifetime abstainers. Among current drinkers, these indicators of health status and healthcare burden were associated with transition to former drinking in the following year, most consistently among adults aged 40 years or older. Including former drinkers with lifetime abstainers in the nondrinker reference group produced more inverse estimates for current drinking. Former drinkers in the Korea Health Panel appear to include a health-selected subgroup consistent with the sick-quitter mechanism. Separating former drinkers from lifetime abstainers and considering pre-cessation health status may improve the interpretation of alcohol-health associations.