Wendy A. Lutz CDHS- Buffalo State College

Alcohol Screening Instruments Fall 2001

Alcohol Screening Instruments for Women: Recommendations

In making recommendations of alcohol screening instruments for use in a Social Service setting it is necessary to consider the following criteria:

· The tool must be a screening tool and not diagnostic. A caseworker or other Social Service professional needs to “screen” for potentially harmful or risky drinking in order to make an appropriate referral for further assessment.

· The tool should be practical for a Social Service professional to utilize the instrument without lengthy training in its administration.

· The tool should have proven reliability and validity.

· The tool should have proven sensitivity (the probability that an individual who is a risk drinker will test positive) and specificity (the probability that an individual who is not a risk drinker will test negative).

· The instrument should be relatively brief allowing for oral administration in an interview setting.

· The instrument should be easily scored.

· It is preferable that the tool is appropriate and valid with women.

Given these criteria, five instruments were considered: the Michigan Alcohol Screening Test (MAST), the Brief Michigan Alcohol Screening Test (BMAST), CAGE, T-ACE, TWEAK and the Alcohol Use Disorders Identification Test (AUDIT) (see previous documents for citations and full instruments with scoring.)

All of these instruments have proven validity and reliability and are all potentially useful in a Social Service setting. However, the MAST was dismissed due to its length (22 questions in the revised edition) and difficulty in scoring. The MAST, the BMAST and the CAGE have not been recommended due to inferior sensitivity and specificity with women (Shafer & Cherpitel, 1999; Cherpitel, 1999; Svikis et al. 1996) and exclusive use of questions that directly ask about quantity, frequency and consequences of alcohol use which can trigger denial and minimization in an individual who is not motivated to recognize a problem and lead to a false negative screen (Piazza et al 2000). The TWEAK and T-ACE, two recommended instruments also have the liability of direct indicators, but this is outweighed by the high sensitivity and specificity to women. Additionally, the CAGE includes a question about guilt (G) that has been shown to lead to false positive screens in women, yet does not include a question about tolerance to alcohol which has been shown to be a reliable indicator in women (Russell, 1996).

The recommended instruments are the AUDIT, TWEAK and T-ACE. These instruments also have liabilities, but the advantages outweigh them.

The Alcohol Use Disorders Identification Test (AUDIT )

The AUDIT was developed under the auspices of the World Health Organization to serve as a multicultural screening instrument (Babor & Grant, 1989). Unlike the other instruments, AUDIT was designed to specifically screen for alcohol problems at earlier stages (Piazza, 2000). An additional advantage is that AUDIT is comprised of both direct and indirect (obvious and non-obvious) indicators; there are direct questions about quantity, frequency and consequences which could trigger denial and minimization but additional questions about remorse, the concern of others which are indirect are also included.

The AUDIT is a paper-and-pencil or orally administered screen consisting of ten items relating to three areas of harmful drinking: 1) amount and frequency of alcohol consumption, 2) dependency symptoms and 3) harmful effects (Russell, 1996). The questions are specific to drinking behavior within the last 12 months.

The AUDIT has the most complex scoring of the recommended instruments. Each item is scored on a scale of 0-4 with a maximum of 40 points. A score of 8 or higher is indicative of harmful drinking.

In a study of 1,330 ER patients who reported being current drinkers, the AUDIT had an overall sensitivity of 85%, but was more sensitive for men (93%) than for women (72%) and was equally sensitive for African-Americans and Caucasians (Cherpitel, 1997). Despite its lower sensitivity for women in this study, Russell (1994, p. 60) suggests that further study is necessary since it is possible that the non-obvious indicators may make the instrument more sensitive to populations (in her example, obstetrics patients) that may have high levels of denial.

The AUDIT’s multi-cultural basis, the inclusion of obvious and non-obvious indicators, and it’s design for detection of earlier stage problems recommend it as a companion instrument to the TWEAK and the T-ACE.

TWEAK

The TWEAK combines questions from the MAST and CAGE that showed particular validity with women and adds a tolerance question. The questions were determined in a study of obstetrics patients in Michigan (Russell and Skinner, 1988.). The questions culled from the MAST addressed blackouts, feeling the need to cut down on drinking and having close friends or family worry or complain about the persons drinking. The tolerance question from the T-ACE was added as well as the question regarding morning drinking from the CAGE.

The TWEAK’s scoring is on a seven point scale with a score of two indicating a positive screen for at-risk drinking. The tolerance question scores two points if a woman reports drinking five or more drinks without falling asleep or passing out. The Worry question scores two points and each of the other questions score one point.

The TWEAK has a lifetime timeframe that may affect its overall performance. Additionally, due to the direct form of the questions, the TWEAK may not detect earlier stages of alcohol problems and could trigger denial.

In a study of 2,717 African-American obstetrics patients in Michigan, the TWEAK had a sensitivity of 91% and a specificity of 77% (Russell, 1996. p. 1438).

The way in which the tolerance question is asked in both the TWEAK and the T-ACE deserves a note. Initially, tolerance was determined by asking “How many drinks does it take to get high?” In initial testing the sensitivity of the TWEAK was 83%. When the question was changed to “How many drinks can you hold before passing out or falling asleep?” sensitivity improved to 91%. The same phenomenon occurred with the T-ACE with sensitivity improving from 70% to 89% (Sokol et al., 1989)

This instrument is recommended because of its tested sensitivity and specificity in a female population, limited or no training necessary for administration, as well as it’s brevity and ease of scoring.

T-ACE

The T-ACE is similar to the CAGE, except that it substitutes a question on tolerance (the same question as appears in the TWEAK) for the CAGE question on guilt. This substitution is based on studies that indicated that a tolerance question and all items on the CAGE with the exception of the guilt question, contributed significantly to the prediction of risk drinking (Russell, 1994. p. 58).

The T-ACE scoring is on a five-point scale with a score of 2 or higher indicating risk for harmful drinking. Like the TWEAK, the tolerance question scores two point if a woman reports drinking five or more drinks before passing out or falling asleep. One point is scored for a positive response on the remaining items.

Like the TWEAK, the T-ACE has a lifetime timeframe and may not be sensitive to earlier stages of alcohol problems.

In the same study referenced in the TWEAK, the T-ACE had a sensitivity of 88% and a specificity of 79%.

The T-ACE is also recommended for its consistent sensitivity and specificity in a female population, limited or no training necessary for administration, as well as its brevity and ease of scoring.

All of these instruments can be orally administered although the brevity of the T-ACE and TWEAK make them more conducive to an interview setting. They also have significant limitations. Screening instruments are obviously self-reporting mechanisms that are only as accurate as the self-report. In no way can a screening instrument replace an in-depth assessment of an alcohol problem. The relationship between the interviewer and the client will also affect the outcome. There are no accurate bio-medical screenings available for alcohol problems so self-report questionnaires like the ones discussed here are the best alternative to determine the need for further assessment of alcohol problems. A further limitation of these instruments is their specificity to alcohol. These instruments do not screen for other drug problems.

It is estimated that substance abuse contributes to at least 50% of all child welfare services cases and in some parts of the country, the prevalence may be as high as 90% (National Center on Addictions and Substance Abuse [NCASA], 1999). Good screening tools and service integration with local treatment facilities that can provide assessment of alcohol and other drug problems has become even more vital after the enactment of the Adoption and Safe Families Act of 1997 which compels Social Service workers to determine whether a child will return to the home in a much shorter time period. In addition to use of tools for screening potential alcohol problems in women, Social Service workers can become more familiar with alcohol and other drug problems and how they affect women and their families. Screening women for alcohol problems is a start to more effective casework.


References

Cherpitel, C. J. (1997). Brief screening instruments for alcoholism. Alcohol Health & Research World, 21, 348-351.

Russel, M. & Skinner, J. B. (1988). Early measures of maternal alcohol misuse as predictors of adverse pregnancy outcomes. Alcoholism: Clinical and Experimental Research, 12, 824-830.

Russel, M. (1994). New Assessment Tools for Risk Drinking During Pregnancy
T-ACE, TWEAK, and Others. Alcohol Health & Research World, 18, 55-61.

Russel, M., Martier, S. S., Sokol, R. J., Mudar, P., Jacobson, S., & Jacobson, J. (1996). Detecting risk drinking during pregnancy: A comparison of
four screening questionnaires. American Journal of Public Health, 86, 1435-1439.

Sokol, R. J., Martier, S. S., & Ager, J. W. (1989). The T-ACE questions: Practical prenatal detection of risk drinking. American Journal of Obstetrics and Gynecology, 160, 863-870.

Pg. 1 of 8