Reliability and Validity of a Two-Question Version of the World Health Organization's Alcohol, Smoking and Substance...

The Alcohol, Smoking and Substance Involvement Screening Test (ASSIST), developed for the World Health Organization (WHO), screens for risks associated with the use of tobacco, alcohol, and seven categories of drugs. Although the ASSIST has acceptable psychometric properties, it is relatively long for a screening test. This study was designed to identify a subset of questions from the full ASSIST instrument having comparable psychometric properties for the classification of low-, moderate-, and high-risk substance use.

The study used three data sets from prior studies using the WHO ASSIST. Samples 1 and 3 were obtained from WHO multisite studies conducted in seven countries. Sample 2 included patient data from a U.S.-based screening and brief intervention program that incorporated the ASSIST into its clinical protocol. Samples 1 and 2 were used to conduct psychometric analyses for combinations of ASSIST items. Sample 3 was used to estimate sensitivity, specificity, and positive and negative predictive value for a two-item ASSIST.

Based on correlation statistics, reliability metrics, and validation analyses, a new, two-item version is proposed. The ASSIST-FC contains one question about the frequency (F) of current use and a second question about current or past concern (C) expressed by others. The ASSIST-FC demonstrates no substantial loss in reliability, validity, and predictive ability when statistically compared with the full-length ASSIST.

The ASSIST-FC has advantages for clinical applications in settings where a brief, efficient, reliable screening test is needed to identify patients with hazardous and harmful substance use who would benefit from a brief intervention. It can also be used to identify patients who are manifesting symptoms of substance dependence that would require further diagnostic evaluation.