Background: Item response theory (IRT) has been used to examine alcohol use disorder (AUD) symptoms and their psychometric properties but has not been previously applied to AUD symptoms from an American Indian sample.
Methods: Lifetime DSM-IV AUD symptoms and binge drinking (5+ drinks men/4+ drinks women) at ≥1, ≥4, ≥8, and ≥15 days per month during the period of heaviest lifetime drinking criteria were assessed in 530 American Indian participants. Exploratory factor analysis was used to examine the factor structure of the 10 AUD symptoms and each alcohol consumption criterion. Two-parameter IRT models generated marginal maximum likelihood estimates for discrimination (a) and threshold (b) parameters for 10 DSM-IV AUD symptoms and each consumption criterion. Differential item functioning (DIF) analysis was used to assess AUD symptom severity in groups defined by gender and age at interview.
Results: The AUD symptoms of “Withdrawal” and “Activities Given Up” were the most severe symptoms. “Tolerance” and “Social/Interpersonal Problems” were the least severe. All AUD symptoms fell on the moderate portion of the severity continuum, except “Withdrawal,” which fell at the lower end of the severe portion. The consumption criterion of 5+/4+ (male/female) at ≥8 times per month demarcated the portion of the severity continuum where AUD symptoms began to occur at a probability of 50%. DIF analysis showed significant gender and age at interview differences for “Hazardous Use,”“Tolerance,” and “Activities Given Up,” but not for the other AUD symptoms.
Conclusions: In this American Indian community sample, alcohol abuse and dependence did not represent distinct disorders. Only one AUD symptom was found outside the moderate portion of the underlying AUD severity continuum. Drinking 5+/4+ (male/female) drinks at a frequency of ≥8 times per month during the period of heaviest lifetime drinking was found to function well as both a risk and a diagnostic criterion for lifetime DSM-IV AUD. DSM-IV AUD symptom criteria, as currently assessed, may be limited in their ability to capture the full range of symptom severity of AUDs, at least in this high-risk population.