Randomized controlled trials, the gold standard in evaluating treatments for alcohol use disorders (AUDs), use drinking-reduction endpoints as the main outcomes. However, considerable uncertainty exists about how best to define drinking-reduction outcomes. Although clinicians and the Food and Drug Administration (FDA) long considered total abstinence as the best indicator of treatment efficacy, this is now seen as an overly restrictive outcome. The FDA presently focuses on the proportion of subjects with no Heavy-Drinking Days (HDD ?5 drinks, or ?5 for men, ?4 for women) during a defined treatment period to define treatment efficacy. Another measure, Average Daily Volume (ADV; mean ETOH gm/day) developed by the World Health Organization (WHO) to define drinking levels for research purposes, has been adopted by the European Medicines Agency (EMA, the EU equivalent of the FDA), to define treatment success. However, because the evidence base for both HOD and ADV as clinical trial outcomes has many serious limitations, drinking outcome measures are considered a key methodological barrier to progress in developing more effective alcohol treatments. Empirical support for a drinking outcome's utility consists of its relationship to clinically relevant consequences, including interpersonal and occupational functioning, medical status, and alcohol use disorders. To examine empirical support for varying definitions of HOD, ADV and other drinking outcomes, the large, representative NESARC surveys (National Epidemiologic Survey on Alcohol and Related Conditions) offer important advantages, including richness and consistency of drinking and consequence measures. For this study, we will utilize data from NESARC Wave 1 (2001-2002; N=43,093) and Wave 2 (3-year follow-ups of Wave 1 participants, N=34,653) and NESARC-111 (2012-2013; N=36,318 new participants). We will use regression splines to determine the most informative functional relationships between HOD, ADV and consequences and identify key change-points in these relationships, to address the following key questions: (1) What levels of HOD and ADV are associated with higher risk of consequences? (2) What level of change in HOD and ADV over 3 years predicts change in consequences over the same period? (3) Are these associations modified by alcohol diagnoses and other alcohol characteristics (potential clinical trial eligibility, treatment history, early heavy drinking)? (4) Do these relationships differ between young and other adults (18-25 vs. 26+), by gender, race/ethnicity, or psychiatric comorbidity? We will also explore the following: Do alternative definitions of HOD and ADV influence findings? Does HOD or ADV offer advantages as outcomes? How do consequences relate to other common clinical trial outcomes: % days abstinent; mean drinks per drinking day? Study findings will help to define the relationship of different potential drinking outcome measures to consequences. The study is positioned to have a substantial impact on alcoholism treatment research by improving the ability to identify effective treatments for alcohol use disorders by defining the most clinically meaningful outcome measures.
The drinking outcome measures that are used to indicate efficacy of treatments for alcohol use disorders are crucial to the success of a randomized clinical trial, but the evidence base for commonly used outcomes is surprisingly sparse, with uncertainty over current drinking outcome measures limiting progress in identifying more effective treatments for alcohol use disorders. We therefore propose a careful, thorough and detailed analysis of the empirical relationship of drinking outcome measures to a range of clinically relevant functional and medical consequences in nationally representative data, in order to identify optimal treatment drinking outcome measures. Findings from this translational project have the potential to widely influence the methodology of randomized clinical trials of treatments for alcohol use disorders, thereby having a substantial impact on public health.
