Psychological distress and drinking patterns among working-age adults in Japan: a population-based analysis

The self-medication hypothesis proposes that psychological distress may contribute to increased alcohol use, yet population-level evidence in Japan remains limited. To examine population-level associations between psychological distress (K6) and drinking status among working-age adults in Japan, with comprehensive adjustment for sociodemographic and behavioral covariates.

We analyzed anonymized official statistical microdata from the 2019 Comprehensive Survey of Living Conditions (Japan). Respondents aged 20-64 years with complete data were included (n = 3,935). Psychological distress was measured using K6 and categorized as low (0-4), moderate (5-12), and high (≥ 13). Drinking status was derived from frequency and amount per drinking day: heavy drinking was defined as estimated ethanol intake ≥ 44 g/day (≥ 2 gou/day); non-drinkers included "rarely," "quit," or "do not drink"; remaining participants were moderate drinkers (reference). Multinomial logistic regression estimated adjusted odds ratios (aORs) for heavy vs. moderate and non-drinkers vs. moderate, adjusting for age, gender, household composition, education, employment, income quartile, smoking, and psychiatric outpatient care.

Of 3,935 participants, 20.6% were heavy drinkers (n = 810), 38.8% moderate drinkers (n = 1,527), and 40.6% non-drinkers (n = 1,598). High distress accounted for 4.5% (n = 177) and moderate distress for 25.3% (n = 995). Psychological distress was not associated with drinking status in cross-tabulations (p = 0.578). In adjusted models, distress was not significantly associated with heavy drinking or non-drinking. Moderate distress showed a borderline tendency toward heavy drinkers (aOR 1.21, p = 0.073), while high distress was null (aOR 1.02, p = 0.927). Covariates showed robust associations, including lower odds of heavy drinkers and higher odds of non-drinkers among women, higher odds of both heavy drinkers and non-drinkers among adults aged 20-39 (vs. 60-64), and strong positive association between smokers and heavy drinkers.

Psychological distress was not an independent predictor of drinking status after adjustment, suggesting limited support for a simple population-level self-medication pathway in Japan. Longitudinal studies and subgroup analyses are needed.