Abstract
BACKGROUND: The variability of youth alcohol use is thought to necessitate longer Timeline Followback (TLFB) reporting windows to accurately capture alcohol use patterns. However, longer TLFB windows can increase burden for participants and study staff. This study assessed youth alcohol use patterns across two different 30-day periods (TLFB days 1-30 and 31-60) to inform future research protocols. METHODS: Participants (N = 481 with 504 visits, ages: 15-25) completed a shared intake to determine eligibility for numerous substance-related studies. Alcohol consumption over the past 60 days was assessed using a modified TLFB that recorded total number of drinks, average drinks per day, average drinks per drinking day (DPDD), total number of drinking days, maximum number of drinks in a single day, and total number of binge drinking days (4+ drinks for females, 5+ drinks for males). TLFB data were divided into two 30-day periods: the most recent (Days 1-30) and the more distal (Days 31-60). Analyses were conducted overall and within subgroups by sex, age (under 21 or 21+), presence of alcohol use disorder (AUD), and other substance use. Alcohol use patterns were assessed using Spearman correlations, Wilcoxon signed rank tests, and intraclass correlations (ICC) for test-retest reliability. RESULTS: All alcohol outcomes were significantly correlated between reporting windows (p's < 0.001). Alcohol use was significantly higher during the more recent reporting window (TLFB Days 1-30; p's < 0.05), except for DPDD (p = 0.58). Overall, the differences between reporting windows were small (on average, less than three total drinks and <1 drinking day or binge drinking day) and suggested clinically similar use patterns. The ICC values suggest moderate-to-good stability between reporting windows. Similar patterns were noted in each subgroup analysis. CONCLUSIONS: A 30-day TLFB window appears sufficient for youth alcohol use assessment, providing comparable and agreeable self-reported alcohol use patterns, with reduced burden.