B – 120 Detection Performance of Objective Alcohol Use Measures Relative to Retrospective Self-Report

Abstract Objective To assess breath alcohol concentration (BrAC) and transdermal alcohol concentration (TAC) alignment with Timeline Followback (TLFB)-reported alcohol use within a contingency management protocol. Method Adults (≥21 years, M=25.18±6.26) from the community were monitored for alcohol use across four weeks via TLFB, remote BrAC, and TAC. Inclusion criteria included: Clinical Institute Withdrawal Assessment for Alcohol Scale ≤10, Mini-Mental State Examination >25, and ≥2 past-month heavy drinking episodes (TLFB). Mania, hypomania, psychosis, and substance use disorder (except alcohol, cannabis, nicotine) were exclusionary. Financial incentives were contingent on four identity-verified, on-time, negative BrAC (≤0.02%) submissions/day. TAC was recorded every 30 minutes. TLFB was administered at study completion. Sensitivity, specificity, and odds ratios were calculated with TLFB as reference. Results Thirty-nine adults (16 male, 23 female) were included in analyses. Alcohol detection by BrAC (OR=2.83, 95% CI [1.87, 4.25], p< 0.001) and TAC (OR=3.50, 95% CI [2.30, 5.36], p< 0.001) were more likely on TLFB-reported drinking days than non-drinking days. BrAC identified 63% of TLFB-reported drinking days (95% CI [0.50, 0.74]) and TAC 59% (95% CI [0.34, 0.80]). Of non-drinking days, 66% aligned with BrAC (95% CI [0.53, 0.77]) and 82% with TAC (95% CI [0.72, 0.89]). Probability of agreement with self-report decreased from study start (TAC: 75.1%, 95% CI [62.0, 79.4]; BrAC: 82.4%, 95% CI [74.9, 88.0]) to end (TAC: 58.5%, 95% CI [48.0, 68.4]; BrAC: 75.1%, 95% CI [65.9, 82.5]). Conclusion Observed discrepancies between measures support the integration of continuous, objective monitoring to enhance the accuracy of behavioral alcohol use assessment. Further investigation is warranted.