Evaluations

• Test-retest reliability of a self-administered Alcohol, Smoking and Substance Involvement Screening Test (ASSIST) in primary care patients.

  McNeely J, Strauss SM, Wright S, Rotrosen J, Khan R, Lee JD, Gourevitch MN. Journal of Substance Abuse Treatment. 2014. 47(1): 93-101. PMCID: PMC4035183.

  Summary: The purpose of this study was to evaluate the test-retest reliability of the computer-administered Alcohol, Smoking and Substance Involvement Screening Test (ASSIST). Adult patients waiting for appointments in a primary care clinic were approached consecutively. Those who agreed to participate (N=146) completed the computer-administered ASSIST on the day of their appointment (Time 1), then again within 14 days (Time 2). The computer-administered ASSIST was feasible for the primary care population, as only 14% of participants had questions on using the computer or understanding ASSIST questions. Test-retest reliability was acceptable. Ninety-three percent of substance-related risk category classifications were consistent between Time 1 and Time 2. Kappa coefficients indicated almost perfect agreement in the results of the Time 1 and Time 2 assessments. Intraclass correlation coefficients (ICCs) for the ASSIST substance scales were high and ranged from 0.900 (tobacco) to 0.969 (overall drug score).

  Take Away: The computer-administered ASSIST has strong test-retest reliability and appears feasible for a primary care population.

• Feasibility and acceptability of an audio-computer self-interview version of the Alcohol, Smoking and Substance Involvement Screening Test (ASSIST) in primary care patients

  Spear SE, Shedlin M, Gilberti B, Fiellin M, McNeely J. Substance Abuse. 2016 Apr-Jun;37(2):299-305. PubMed PMID: 26158798; PubMed Central PMCID:PMC4962999.

  (Follow-Up of Previous Study.)

  Summary: Participants from the McNeely et al (2014) study were asked to participate in this qualitative study investigating the feasibility and acceptability of the computer-administered Alcohol, Smoking and Substance Involvement Screening Test (ASSIST). Researchers conducted semi-structured interviews and focus groups with 48 participants who had completed the computer-administered ASSIST twice. Interviews were transcribed and coded, then organized into themes. Results showed that participants found the computer-administered format acceptable. The vast majority of participants found the computer-administration easy to use, and completed the assessment quickly. Participants also believed that screening for substance use was important and could improve patient care. Despite this, participants believed that the inability to explain and elaborate responses was a negative aspect to the computer administration. Some stated that they didn’t disclose substance use because the computerized format made it impossible to explain specific situations or details. Several immigrant participants worried about the legal ramifications of disclosing substance use, especially when they were unsure who would have access to the results of the computer-administered survey.

  Take Away: Overall, the computer-administered ASSIST was feasible and acceptable to users. Issues of privacy and confidentiality may impact what information is disclosed during a computer-administered assessment.

• Screening for substance use disorder among incarcerated men with the Alcohol, Smoking, Substance Involvement Screening Test (ASSIST): Comparative analysis of computer-administered and interviewer-administered modalities.

  Wolff N, Shi J. Journal of Substance Abuse Treatment. 2015. 53: 22-32. PMCID: PMC4414742.

  Summary: This study compared the feasibility, reliability, and validity of the computer-administered Alcohol, Smoking and Substance Involvement Screening Test (ASSIST) with the interviewer-administered ASSIST for incarcerated men. Incarcerated men in one prison were recruited to participate (N=396) and completed the ASSIST two times. Participants were randomized to complete either the computer- or interview-administered version of the ASSIST initially. Within 14 days of completing the ASSIST, each participant was again randomized to either the computer- or interviewer-administered version and completed the ASSIST a second time. Participants completed the Structured Clinical Interview for DSM Disorders (SCID). The computer-administered ASSIST was feasible for participants. Less than 1% had difficulty using the computer or comprehending questions. Rates of lifetime and current substance use were almost equivalent between the computer- and interviewer-administered modalities and were similarly complete. Reliability was high for the computer-administered ASSIST. Intraclass correlation coefficients (ICCs) of global risk scales ranged from 0.877 to 0.926. Modality did not impact the test-retest reliability of the ASSIST. To examine the validity of the computer-administered ASSIST, the ASSIST scores and SCID abuse/dependence scores were compared. The computer-administered ASSIST and SCID scores were moderately correlated. For cocaine, amphetamine, and opioid abuse/dependence scores, the computer-administered ASSIST correlated more strongly with the SCID than the interviewer-administered ASSIST.

  Take Away: A computer-administered version of the ASSIST is feasible for use with incarcerated men and demonstrated good psychometric properties.