The timeline followback (TLFB) takes more resources to collect than the Alcohol Use Disorder Identification Test (AUDIT-C). We assessed agreement of TLFB and AUDIT-C with the biomarker phosphatidylethanol (PEth) and compared changes in TLFB and PEth among persons with HIV (PWH) using secondary data from randomized trials. We calculated operating characteristics and agreement between TLFB (> 1 and > 2 average drinks/day), AUDIT-C ≥ 4 and PEth ≥ 20 among 275 men with HIV. Median age was 57 years, 80% were African-American; and 17% white. Sixty-eight percent had PEth ≥ 20, 46% reported > 2 average drinks/day on TLFB, 61% reported > 1 average drinks/day on TLFB, and 72% had an AUDIT-C ≥ 4. Relative to PEth, sensitivity for AUDIT-C ≥ 4 was 84% (kappa = 0.36), and for TLFB > 1 average drink/day was 76% (kappa = 0.44). Change in alcohol use appeared greater using TLFB measures than PEth. Strategies to robustly assess alcohol use in PWH may require both self-report and biomarkers.
Keywords: Alcohol consumption; HIV infection; Phosphatidylethanolamine.
© 2021. This is a U.S. Government work and not under copyright protection in the US; foreign copyright protection may apply.