Identifying Hazardous Alcohol Use in Primary Care: Concordance Between Phosphatidylethanol and Self-Reports.

Abstract (English)

PURPOSE: Screening for hazardous alcohol use is a challenging task in primary care. We examined concordance between the alcohol biomarker phosphatidylethanol (PEth) and 2 self-reported measures of alcohol use-score on a screening questionnaire and weekly drinks-in primary care. METHODS: We conducted a retrospective cohort study using regional primary care health records linked to national registries in Sweden. Analyses used 2 samples: (1) patients with a PEth test who also had an Alcohol Use Disorders Identification Test (AUDIT) score in the preceding or following 90 days (n = 3,679) and (2) patients with a PEth test who also had a documented self-reported weekly alcohol use in the preceding or following 90 days (n = 8,934). We defined hazardous alcohol use as a PEth level of greater than 210 ng/mL (0.30 μmol/L), an AUDIT score of 8 or higher, and a reported consumption of 10 or more drinks per week. Associations between PEth and self-reports were examined using correlation analyses and proportional-odds models adjusted for sociodemographic and clinical characteristics. RESULTS: The prevalence of hazardous alcohol use varied substantially by assessment method. The AUDIT questionnaire classified the largest proportion of patients as having hazardous use (26.5%), followed by PEth level (14.6% and 18.3% in the first and second samples, respectively) and by reported weekly drinks (8.5%). Each method identified patients not detected by the others. PEth levels correlated moderately with both AUDIT scores and weekly alcohol consumption, but relationships were nonlinear. At similar PEth levels, men and patients with higher education or income reported greater weekly alcohol use, while younger patients had higher AUDIT scores. CONCLUSIONS: An alcohol biomarker and self-reported measures capture overlapping but distinct aspects of alcohol use in primary care. For primary care clinicians, PEth should be interpreted alongside self-reported measures rather than used as a standalone test. Combining assessment methods may improve identification of alcohol-related risk and support more individualized, context-sensitive clinical conversations.