Psychiatric disorders among older U.S. women by socioeconomic and veteran status.

The Alcohol, Smoking and Substance Involvement Screening Test (ASSIST) yields substance-specific risk scores for nine substances, and can identify individuals in need of interventions for moderate levels of substance use. Humeniuk et al. established the construct, concurrent, and discriminant validity of the ASSIST 3.0 in a multi-national sample of 1,047 respondents from primary care and substance use treatment settings. The present study examined the factor structure of the tobacco, alcohol, and cannabis items in a sample of 2,599 adults awaiting a primary care appointment. Aims were twofold: (1) evaluate the adequacy of a three-factor oblique solution (ie, three correlated factors) in explaining the relationships among the ASSIST tobacco, alcohol, and cannabis items; and (2) assess the suitability of short forms for these three substances, composed of a subset of the respective items. Factor analysis has the ability to evaluate the measurement properties (eg, reliability) of a set of items and determine their ability to measure an underlying construct(s), and to determine a reduced set of items that display measurement properties equivalent or superior to a larger set of items. The present study draws from screening data from two randomized controlled trials of a brief intervention in primary care. The first was conducted at two community health centers in New Mexico, with the second conducted at one of the two centers. The larger of the studies is described elsewhere, and the second is recently completed. Both used the same screening protocol, which involved anonymous interviewer administration of the ASSIST 3.0 in English. Participants were not paid for their screening. The studies were approved by the Friends Research Institute and the Christus Health Institutional Review Boards.