Validity of the SMAST in two American Indian tribal populations

The standardized evaluation of alcoholism and other psychopathologies in minority populations, particularly American Indians, has long been questioned. This study investigated the validity of one of the most commonly applied assessments for alcoholism--the Short Michigan Alcohol Screening Test (SMAST)--in two distinct American Indian tribal groups. We analyzed data collected from 1989 to 1995 from largely community representative samples of 456 Southwestern and 214 Plains Indians ages 21 or older. For comparison, alcohol dependence was diagnosed using Diagnostic and Statistical Manual of Mental Disorders, Third Edition-Revised (DSM-III-R) criteria from a detailed, modified version of the Schedule for Affective Disorders and Schizophrenia--Lifetime (SADS-L). Accuracy of the SMAST was quantified as sensitivity, specificity, likelihood ratios, and the area under the curve for receiver operating characteristics, using the DSM-III-R diagnosis as the reference. The standard SMAST cutoff score of > or = 3 had a demonstrated sensitivity 86% to 95%, but did not perform well in terms of specificity (23%-47%). Significantly higher cutoff scores (> or = 5 for both genders in the Southwestern tribe and 8 and > or = 6 for men and women in the Plains tribe) were required to demonstrate acceptable levels of specificity in both tribes. The findings suggest that the SMAST is not a valid tool to screen for alcohol misuse in these two tribal populations. The highly elevated and different thresholds required from one population to the next and from one gender to the next constitute a significant obstacle to the use of the instrument.