In light of the above evidence,12 here we investigate the impact of criteria heterogeneity within criterion count–based AUD diagnoses. As other research also suggests that severe AUD may be substantively different from mild or moderate AUD in terms of treatment needs,16 functional impairment,17 and other health sequelae,18 we sought to compare mild-to-moderate AUD (ie, endorsing 2-5 criteria, meeting criteria for either mild or moderate AUD) vs severe AUD while accounting for individual differences in criterion count among those with mild-to-moderate AUD. The current study aimed to (1) validate criterion count–based severity using DSM-5 categorizations of AUD, including comparing mild-to-moderate vs severe AUD, in a sample of 13 110 individuals; (2) use item response theory (IRT) modeling to identify criteria indicative of greater severity; and (3) evaluate whether the presence of certain high-risk criteria, identified through IRT modeling, indexes greater hazards of developing severe AUD in a related cohort of 2818 adolescents and young adults.
