The Addiction Severity Index (ASI) is a widely used instrument that assesses problems in seven areas - medical, employment, alcohol, drug, legal, family/social, and psychiatric. Despite proven utility, evidence for validity of ASI summary indices is sparse and has not been well-established for the ASI-5 or ASI-5 Lite, the two versions used in VA. Much of the ASI research has used the two original ASI summary indices, the Interviewer Severity Ratings (ISRs) and the Composite Scores (CSs), comprised of lifetime and recent status items, similar to the ISRs; and 2) Evaluation Factors (Efs), consisting of only recent status items like the CSs. The new indices (CFs & Efs) are standardized and may provide better correspondence with external variables.
The research examined and compared the concurrent and predictive validity of the ISRs, CSs, CFs, and EFs of the standard interviewer-administered ASI-5 in a diverse sample of veteran patients. Also, the equivalence of an abbreviated baseline ASI-5 interview was compared to the standard ASI-5. The follow-up ASI-5 interview is already brief and there was no need to develop an abbreviated follow-up ASI-5 interview. Finally, the abbreviated patient self-administered baseline and follow-up versions of the ASI-5 were compared to standard ASI-5 interviews.
Participants were 216 patients with psychoactive substance use disorders entering treatment at the Philadelphia VA Medical Center. In one sample (n=146), 2 versions of the baseline ASI (i.e., ASI-5 and ASI-5 VA-Lite) were administered 2-3 days apart in counterbalanced order by different interviewers. An external validation battery was administered in the initial session following the first ASI. At 6 months post-admission, the follow up version of the ASI-5 was administered along with the external validation battery. For a second sample (n=70), 2 versions of the baseline ASI (i.e., ASI-5 and ASI-5 VA-Lite Self-Administered) were administered 2-3 days apart, and at 6 months post-admission, the follow up versions of both were administered. At each time, the validation battery was given to participants in the initial session. Outcome measures also included treatment completion and urine toxicology data.
Concurrent validity analyses found that overall the ASI summary indices (CSs, ISRs, Efs, CFs) were significantly correlated with the corresponding external validation measures. Additionally, the correlations among the four indices to the external validation measures were generally similar. Regarding predictive validity, multiple logistic regression analyses found that: 1) all four sets of ASI summary indices were similarly predictive of treatment completion; and 2) all four sets of indices were similar in predicting a drug-free 6 month follow up urine toxicology. The analyses comparing the equivalence of different versions of the ASI (standard ASI-5 vs. ASI-5 VA-Lite and standard ASI-5 vs. ASI-5 VA-Lite Self Administered) indicated high correspondence between the summary scores derived from each of the Lite versions with the summary scores derived from the standard ASI.
These results present little evidence to suggest the use of one set of indices over another. The validation of abbreviated and/or self-administered baseline and follow up versions of the SAI-5 provides treatment and research programs within and outside of VA with more efficient means of conducting patient assessments.
None at this time.
Substance Abuse and Addiction, Health Systems
Diagnosis, Technology Development and Assessment