Item response theory analyses of DSM-IV and DSM-5 stimulant use disorder criteria in an American Indian community sample.

Abstract

Native Americans experience some of the highest rates of DSM-IV stimulant dependence (SD) of all U.S. ethnic groups. This report compares DSM-IV and DSM-5 stimulant use disorder (SUD) diagnostic criteria in an American Indian community sample. Demographic information, stimulant (methamphetamine or cocaine) use, and lifetime DSM-IV and DSM-5 diagnoses were assessed in 858 adult American Indians. Item Response Theory (IRT) analyses were used to assess SUD criteria in both DSM-IV and DSM-5 criteria sets along an underlying latent trait severity continuum and the effect of demographic variables on differential item functioning (DIF) in those criteria. The overall rate of DSM-IV SD was 33%, of DSM-IV SUD was 38%, and of DSM-5 SUD was 36% with no gender differences. All SUD symptoms in both the DSM-IV and DSM-5 datasets functioned on the moderate portion of the underlying severity continuum. "Craving" discriminated better than any other criterion at its level of severity in indicating the presence or absence of SUD. There was little DIF in groups defined by gender or any other demographic variable in either the DSM-IV or DSM-5 datasets. These findings indicate that in this American Indian sample, diagnostic criteria for DSM-IV and DSM-5 SUD function similarly in terms of severity and DIF and that the abolition of the DSM-IV distinction between stimulant abuse and dependence in DSM-5 is warranted.