Self-reported alcohol use versus phosphatidylethanol in behavioral trials: A study of people living with HIV in Tshwane, South Africa.

Background: Accurately quantifying alcohol use among persons with HIV (PWH) is important for validly assessing the efficacy of alcohol reduction interventions.

Methods: We used data from a randomized controlled trial of an intervention to reduce alcohol use among PWH who were receiving antiretroviral therapy in Tshwane, South Africa. We calculated agreement between self-reported hazardous alcohol use measured by the Alcohol Use Disorders Identification Test (AUDIT; score ≥8) and AUDIT-Consumption (AUDIT-C; score ≥3 for females and ≥4 for males), heavy episodic drinking (HED) in the past 30 days, and heavy drinking in the past 7 days with a gold standard biomarker--phosphatidylethanol (PEth) level (≥50 ng/mL)--among 309 participants. We used multiple logistic regression to assess whether underreporting of hazardous drinking (AUDIT-C vs. PEth) differed by sex, study arm, and assessment time point.

Results: Participants' mean age was 40.6 years, 43% were males, and 48% were in the intervention arm. At 6 months, 51% had PEth ≥50 ng/mL, 38% and 76% had scores indicative of hazardous drinking on the AUDIT and AUDIT-C, respectively, 11% reported past 30-day HED, and 13% reported past 7-day heavy drinking. At 6 months, there was low agreement between AUDIT-C scores and past 7-day heavy drinking relative to PEth ≥50 (sensitivities of 83% and 20% and negative predictive values of 62% and 51%, respectively). Underreporting of hazardous drinking at 6 months was associated with sex (OR = 3.504. 95% CI: 1.080 to 11.364), with odds of underreporting being greater for females.

Conclusions: Steps should be taken to decrease underreporting of alcohol use in clinical trials.