Alcohol Use Disorders Identification Test (AUDIT)

The AUDIT was developed by the World Health Organization to identify hazardous and harmful alcohol use in adults and to serve as a brief (10-item) screening instrument for excessive drinking. It was intended to identify a broad spectrum of problem drinkers and distinguish them from hospitalized alcoholics and normal drinkers and to be used in primary care health settings cross-culturally. It was validated on primary care patients in 6 countries (Norway, Australia, United States of America, Kenya, Mexico, Bulgaria). Items were selected from the pool of items based on their correlation with daily alcohol intake, frequency of drinking 6 or more drinks per episode, and their ability to discriminate between hazardous and harmful drinkers. A cut-off score of 8 produced sensitivity in the mid-0.90s for this group. (Ref: 2,6-8)

The AUDIT contains a Core section of 10 items that make up three subscales (3 items on dependence symptoms, items on problems caused by alcohol or harmful alcohol use; and 3 items on hazardous alcohol use or alcohol consumption) (Ref: 8). An optional Clinical Screening Procedure (8-item) consists of 2 items about traumatic injury, 5 items on clinical exam, and a blood test. The Core items focus on the recent past rather than leaving alcohol consumption to ‘ever.’ The Clinical items are reflective of physical effects and do not refer directly to drinking. (Ref: 5) It can be used in multiple clinical and research areas, and provides a framework for intervention by helping practitioners identify those who would benefit from reduction or cessation of drinking. The AUDIT has been standardized cross-nationally. (Ref: 2,6-8)

Free; the training module costs $75.

2-4 minutes for the AUDIT Core; 10 minutes for the AUDIT Clinical

1982

Flesch-Kincaid Grade Level of 7.7; a person reading at a 7th grade level could comprehend the AUDIT. Items are usually less than 20 words each and do not contain technical terms.

Each response has a range of 0-2 (0=no, 1=yes) or 0-4 (from 0=never to 4=daily). (Ref: 2)

Self-administered paper and pencil or computerized; Interview. The interview form and computerized form provide 2 opportunities for skipping patterns, which will shorten the administration time (skipping patterns are considered too complex to follow in the paper administration). (Ref: 8)

Scored by hand (no computerized scoring available). Score range is 0-40 on the Core scale and 0-24 on the Clinical scale. (Ref: 5) The original study suggested a cut-off of 10/11. (Ref: 3) Scores of >=8 are indicative of hazardous and harmful alcohol use and possible alcohol dependence. For men and women over age 65, the cut-off should be reduced to 7 in order to increase sensitivity. (Ref: 8) Interpretation for scores is as follows: 1) patients scoring 8-15 should receive simple advice focused on reduction of drinking, 2) patients scoring 16-29 should receive brief counseling and follow-up monitoring, 3) patients scoring 20 or more need further diagnostic evaluation for alcohol dependence. (Ref: 8)

Multiple languages available, including: Turkish, Hindi, Greek, German, Dutch, Polish, French, Japanese, Portuguese, Spanish, Danish, Chinese, and Italian.

World Health Organization (WHO)

PDF Including Self-report and Interviewer-administered versions.

Avenue Appia 20
1211 Geneva 27
Switzerland

publications@who.int

www.who.int

top

Factor analysis of the original 150 items in the interview schedule resulted in the extraction of 7 factors. The dominant first factor encompassed most of the items on the alcohol-specific scales drinking behavior (dependence), adverse psychological reactions, and alcohol-related problems. The measures of alcohol intake loaded on a separate factor. (Ref: 4) No factor analytic information was found for the 10 items comprising the final AUDIT instrument.

The AUDIT was validated on 913 heavy drinkers and 297 alcoholics (Ref: 3,7), who were primary care patients from 6 countries. (Ref: 8)

In a group of non-hazardous drinkers, cocaine abusers and alcoholics, test-retest r=0.86. (Ref: 8)

No information found.

One study reports reliability based on the four scales (conceptual domains) in the original 150-item screening instrument (Ref: 4) for both drinkers and alcoholics. Among drinkers and alcoholics, the ‘alcohol dependence/drinking behavior’ scale had reported Cronbach’s alpha ranging from 0.80 (drinkers) to 0.98 (drinkers) across the 6 countries in the validation study. For the ‘adverse psychological reactions’ scale, Cronbach’s alpha range=0.45 (alcoholics) to 0.90 (drinkers). For the ‘alcohol problems in the previous year’ scale, Cronbach’s alpha range=0.25 (alcoholics) to 0.80 (drinkers). Among drinkers only, the ‘alcohol problems ever’ scale had a reported Cronbach’s alpha ranging from 0.40 to 0.83 (and lower among alcoholics, though these values are not reported. Another article summarized mean intrascale reliability values for the conceptual domains (Ref 2) for the same data, only among drinking patients. The mean value (across the 6 countries) of Cronbach’s alpha for the ‘drinking behavior’ scale (all 13 items) was 0.93. The mean value for ‘adverse psychological reactions’ scale (4 items) was 0.81. These two scales showed little variation across the 6 countries. The ‘alcohol problems in the previous year’ scale (number of items not available) and the ‘alcohol problems ever’ scale (5 items) had mean Cronbach’s alpha values of 0.69 and 0.65, respectively, with more variation across the 6 countries.

For the 10-item AUDIT, a literature review reports that coefficient alphas range from 0.75 to 0.94. (Ref: 6)

No information found.

According to self-report measures, the AUDIT showed construct validity with risk factors, drinking consequences, and drinking attitudes. The range of correlations follow (Ref: 5):

AUDIT scores correlated with DSM-III diagnoses (p<0.001) and with self-reports of drinking frequency (p<0.001). (Ref: 3) Concurrent validity was high for the AUDIT Core and Clinical with both men and women, when assessed with the Michigan Alcoholism Screening Test and MacAndrew Alcoholism Scale (Ref: 5):

No information found.

No information found, however, the AUDIT was designed to be a screening instrument rather than an instrument to evaluate change over time. Responsiveness evidence is not particularly pertinent for establishing the validity of a screening instrument.

No information found.

Yes. (Ref: 3,5 and references cited within Ref: 6-8)