Identifying adolescents with alcohol use disorder: Optimal screening using the National Institute on Alcohol Abuse and Alcoholism screening guide

Abstract

The National Institute on Alcohol Abuse and Alcoholism (NIAAA) screening guide (SG) uses a 2-question screening process to identify adolescents at risk for alcohol use. The aim of this study was to identify the optimal screening rule in terms of identifying adolescents with alcohol use disorder (AUD) using the NIAAA questions by examining whether the cutpoint should vary by gender, race/ethnicity, grade, and/or age. Youth aged 12 through 18 years (N = 1,573; 27% Black, 51% Hispanic) were screened using the NIAAA SG, and then completed a survey. We used receiver operating characteristic curve analyses to identify the optimal cutpoint for the NIAAA screener question on self-use with AUD as the outcome. We compared the resulting screening rule with the NIAAA SG rule. We found that the optimal cutpoint depended on age and grade of adolescents. The resulting screening rule was the same as the NIAAA SG, and thus independently validated the NIAAA SG, with the exception of screening for adolescents 18 years of age, for which a lower cutpoint was indicated. The performance of both screening rules was highly similar when applied to the study sample, with a sensitivity of 0.89 for the optimal screening rule and a sensitivity of 0.87 for the NIAAA SG. In settings in which the cost of a false positive is relatively low (depending on available resources and cost of the intervention), lower cutpoints for older adolescents should be considered, as this may increase sensitivity of identifying these individuals at risk for AUD. (PsycINFO Database Record

Figure 1

Empirical ROC curve among entire sample with associated point-wise confidence bands

Figure 2

Empirical ROC curves by grade, (a) 5^th–8^th grade and (b) 9^th grade and higher, with associated point-wise confidence bands