Does subtle screening for substance abuse work? A review of the Substance Abuse Subtle Screening Inventory (SASSI) Sarah W. Feldstein & William R. Miller University of New Mexico Center on Alcoholism, Substance Abuse, and Addictions (UNM CASAA), Albuquerque, NM, USA ABSTRACT Aim Through a complex combination of direct (face-valid) and indirect (subtle) subscales, the Substance Abuse Subtle Screening Inventory (SASSI) is purported to detect substance use disorders with a high degree of validity regardless of respondent honesty or motivation. This review evaluates empirical evidence regarding the reliability and validity of this widely used screening instrument. Methods Source documents were 36 peer-reviewed reports yielding data regarding the SASSI’s internal consistency, test–retest reliability, psychometric structure, convergent and diver- gent validity and criterion (predictive) validity. Results The total N of the studies reviewed equaled 22 110. Internal consistency is high for the overall SASSI and for its direct but not its indirect (subtle) subscales, suggesting that the instrument taps a single face-valid construct. SASSI classifications converged with those from other direct screening instruments, and were also correlated with ethnicity, general distress and social deviance. Studies found test–retest reliability lower than that reported in the test manuals. Sensitivity was found to be similar to that for public domain screening instruments, but on specificity the SASSI appears to yield a high rate of false positives. Conclusion No empirical evidence was found for the SASSI’s claimed unique advantage in detecting substance use disorders through its indirect (subtle) scales to circumvent respondent denial or dishonesty. Recommendations for screening and for future research with the SASSI are offered. Keywords Adolescent, adult, SASSI, screening, substance abuse. Correspondence to: Sarah W. Feldstein, Center on Alcoholism, Substance Abuse and Addictions (CASAA), University of New Mexico, MSC116280, Albuquerque, NM 87106, USA. E-mail: swfeld@unm.edu Submitted 10 February 2006; initial review completed 10 May 2006; final version accepted 2 October 2006 INTRODUCTION Direct and indirect scales The quest for accurate and efficient instruments to screen for substance use disorders (SUDs) has been ongoing for at least 60 years [1]. The usual purpose of screening is to identify cases that warrant more careful evaluation to confirm or disconfirm an implied diagno- sis. There has been particular interest in instruments to circumvent denial and accurately detect SUDs regard- less of the respondent’s honesty or awareness. Miller [1] described these as ‘indirect scales’, differentiating them from direct scales that ask obvious questions about substance use. A concern with direct scales is that their items are face-valid, and thus easier to falsify. Early attempts to create indirect scales derived alco- holism subscales from the Minnesota Multiphasic Person- ality Inventory (MMPI [2]). Five such scales were developed, comprised of items with little or no obvious relationship to alcoholism (e.g. ‘I think Lincoln was a greater President than Washington’). Miller [1] con- cluded that these indirect scales had no higher sensitivity than direct scales, and were prone to poorer specificity (more false positives). Most MMPI-based indirect scales have subsequently fallen into disuse. THE SUBSTANCE ABUSE SUBTLE SCREENING INVENTORY (SASSI) The SASSI [3,4] is a more recent screening instru- ment, widely used in the United States, and marketed METHODS AND TECHNIQUES doi:10.1111/j.1360-0443.2006.01634.x © 2007 The Authors. Journal compilation © 2007 Society for the Study of Addiction Addiction, 102, 41–50