Assessing substance use in multiproblem patients: reliability and validity of the Addiction Severity Index in a mental hospital population.

The Addiction Severity Index (ASI) is the most widely used measure of substance use in the field. Its reputation has been supported by reliability and validity studies. Despite its success, the psychometric properties of the ASI have not been examined in mental hospital populations. Our intent was to replicate prior studies and expand upon the validity of the ASI in a sample of 100 public psychiatric patients selected for a larger study. Findings revealed that a) reliability was acceptable, but there was only moderate agreement on the psychiatric scale severity score; b) the relationship between extent of rater training and reliability requires further study; c) despite some overlap, the scales were largely independent of each other; d) modification of the employment scale was necessary because of low correlations between the composite and severity score; e) raters are more responsive to client subjective ratings in psychiatric settings; f) ASI drug and alcohol scales correlate well with other substance use instruments and with DSM-III-R diagnoses; and g) the ASI can identify meaningful types of patient problems through cluster analysis. These findings, on the whole, support the use of the ASI drug and alcohol scales in public psychiatric hospitals.