Validation of the ASSIST for Detecting Unhealthy Alcohol Use and Alcohol Use Disorders in Urgent Care Patients

Background: Screening and brief intervention (SBI) is effective in reducing alcohol use, particularly among moderate risk patients. Results of SBI are inconsistent among patients with alcohol use disorders (AUDs). The Alcohol, Smoking, and Substance Involvement Screening Test (ASSIST) is used as a screening tool in many existing SBI programs. ASSIST validation studies have identified risk level cutoff scores using criteria for AUD and have not included a criterion measure for at-risk drinking (ARD), the group for whom SBI is most effective. This study examines the ability of the ASSIST to identify unhealthy alcohol use (ARD or AUD) and AUD in patients presenting to urgent care. Methods: Data were obtained from interviews with 442 adult drinkers presenting to 1 of 3 urgent care clinics. Subjects completed the ASSIST, a 90-day timeline follow-back interview to detect ARD, and a modified Diagnostic Interview Schedule to identify AUD. Validity measures compared the specificity and sensitivity of cutoff scores for the ASSIST in detecting unhealthy alcohol use and AUDs. Results: The optimal ASSIST score for detecting unhealthy alcohol use is 6+ for males (sensitivity and specificity 68 and 66%, respectively) and 5+ for females (62%/70%). Sensitivity, specificity, and receiver operating characteristic values were lower than those previously reported for the Alcohol Use Disorders Identification Test (AUDIT). For AUD, the optimal ASSIST cutoff scores are 10+ for males (63%/85%) and 9+ for females (63%/85%). While higher scores provided increased specificity, thereby reducing the percentage of false positives, sensitivity dropped sharply as scores increased. Conclusions: Optimal ASSIST cutoff scores for unhealthy alcohol use are lower than those commonly used in many SBI programs. Use of lower ASSIST cutoff scores may increase detection of unhealthy alcohol use and increase the numbers served by SBI programs.