(PS6-D85) World Health Organization Risk Drinking Levels Differentiate Alcohol Use and Functioning Among Non-abstinent Mandated College Students

Purpose: Quantifying meaningful change in college students’ alcohol use has proven difficult, given that most interventions target moderate drinking rather than abstinence. World Health Organization (WHO) risk drinking levels (i.e., low, moderate, high, or very high risk) have been used to quantify individual change in alcohol use disorder clinical trials, as a drinking reduction endpoint. We therefore sought to validate the utility of WHO risk drinking levels for differentiating alcohol use and functioning among college students. To capture risk of acute harm, risk categories were derived using typical drinks per drinking day. A sensitivity analysis compared results when categorizing based on typical drinks per day.   

Method: Data were drawn from the baseline survey of three intervention trials for college students who were mandated to intervention and who were not abstinent (N=1437). Alcohol consumption on each day of the week was assessed. WHO risk levels were created using the published sex-specific cutoffs and dummy coded, with low risk as the reference. Alcohol outcomes included prevalence of positive AUDIT scores (score 8+), peak drinks in one day, and consequences. Grade point average over 3.5 (i.e., Dean’s list) and risk for clinical depression served as markers of functioning. Logistic or over-dispersed Poisson regression models were estimated for the binary versus count outcomes, controlling for sex, ethnicity, and year in school.

Results: With drinking risk level defined by typical drinks per drinking day, 15%, 21%, 38%, and 26% of participants were classified as low, moderate, high, and very high risk, respectively. Compared to low risk students (35%), those in the moderate (45%), high (78%), and very high (94%) risk levels were significantly more likely to be positive on the AUDIT (ps< .001). Moderate to very high risk levels also differed from low risk in peak drinking and consequences (ps< .001). Prevalence of Dean’s list students only differed between very high and low risk drinkers (p=.02), and prevalence of risk for depression was significantly lower among high relative to low risk drinkers (p=.02). Using typical drinks per day, 68%, 23%, 7%, and 2% were classified as low, moderate, high, or very high risk. Results largely replicated. However, very high risk (100%) did not differ from low risk (57%) in AUDIT positive prevalence (p=1.00). There were no differences among the risk levels in risk for depression (ps >.11).  

Discussion: Whether defined by drinks per drinking day or drinks per day, WHO risk drinking levels meaningfully differentiate hazardous alcohol use and consequences among mandated college students. Defining risk using drinks per drinking day captures a broad range of individuals at risk for consequences, whereas the typical drinks per day definition may be useful for identifying those requiring intensive intervention. Given their utility for differentiating consumption, consequences, and academic performance, WHO risk levels hold promise as a means for quantifying individual-level, post-intervention reductions in drinking among college students.