Three ways of detecting hazardous drinking found three different groups of patients

In Swedish primary care records, a questionnaire flagged 26.5% of patients, a blood biomarker 14.6-18.3%, and self-reported weekly drinks 8.5%. Each method found people the other two missed.

Asking people how much they drink is the standard way of finding hazardous drinking, and it is known to be imperfect. Phosphatidylethanol, or PEth, is a blood marker that reflects alcohol consumption over the previous weeks and is sometimes treated as the objective answer. Researchers led by Viktor Mansson used Swedish regional primary care records linked to national registries to compare them directly, in two samples: 3,679 patients who had both a PEth test and an AUDIT questionnaire score within 90 days, and 8,934 who had a PEth test and a documented self-reported weekly intake. Hazardous use was defined in advance for each: PEth above 210 nanograms per millilitre, AUDIT of 8 or more, or 10 or more drinks a week. The three methods disagreed substantially. The questionnaire classified the most people as hazardous, 26.5%. PEth flagged 14.6% in the first sample and 18.3% in the second. Self-reported weekly drinks flagged the fewest, 8.5%. Crucially, each method identified patients the others did not. PEth correlated only moderately with both self-reports, and the relationship was nonlinear. There were systematic differences too: at similar PEth levels, men and patients with more education or higher income reported drinking more, while younger patients scored higher on AUDIT. That pattern matters, because it means the gap between blood and self-report is not random noise but varies with who the patient is. The authors' conclusion is the practical one: PEth should be read alongside what someone says, not used to overrule it.