The Alcohol Use Disorders Identification Test was not affected by ethnic and sex bias
ACP J Club. 1999 Mar-April;130:43. doi:10.7326/ACPJC-1999-130-2-043
Steinbauer JR, Cantor SB, Holzer CE III, Volk RJ. Ethnic and sex bias in primary care screening tests for alcohol use disorders. Ann Intern Med. 1998 Sep 1;129:353-62.
Are there ethnic and sex biases in 3 self-report screening tests for alcohol use disorders in a primary care population?
Blinded comparison of the results of 3 self-report screening tests with diagnosis according to Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV ) criteria.
University-based, family practice center in Galveston, Texas, USA.
1333 adults (mean age 43 y) randomly selected from appointment lists stratified by sex and ethnicity (white, African American, and Mexican American).
Description of tests and diagnostic standard
The 3 screening tests used were the CAGE questionnaire, the Self-Administered Alcoholism Screening Test (SAAST), and the Alcohol Use Disorders Identification Test (AUDIT). The diagnostic standard, an interview using DSM-IV criteria, was administered by a project interviewer and scored by computer algorithm.
Main outcome measures
Positive and negative likelihood ratios (LRs).
The CAGE questionnaire and the SAAST performed poorly for African-American and Mexican-American women. +LRs for the AUDIT (cut-point 5) were similar to or higher than those for the other screening tests (cut-points 2 for the CAGE and 3 for the SAAST), whereas -LRs were lowest for the AUDIT (< 0.33), indicating the superiority of this test in ruling out alcohol abuse (Table).
The Alcohol Use Disorders Identification Test was not affected by ethnic and sex bias when screening for alcohol use disorders in primary care.
Sources of funding: In part, National Institute on Alcohol Abuse and Alcoholism and Bureau of Health Professions, Health Resources and Services Administration.
For correspondence: Dr. R.J. Volk, Department of Family Medicine and Community Medicine, Baylor College of Medicine, 5510 Greenbriar, Houston, TX 77005. FAX 713-798-7940.
Table. Positive and negative likelihood ratios (LRs) for 3 screening tests for alcohol use disorders (diagnostic standard DSM-IV )
|Patient group||+LRs* (95% CI)||-LRs* (CI)|
|White men||3.0 (2.0 to 4.4)||2.8 (1.9 to 4.1)||3.6 (2.6 to 4.8)||0.4 (0.2 to 0.7)||0.4 (0.2 to 0.7)||0.1 (0.03 to 0.3)|
|Women||4.8 (2.8 to 7.7)||2.4 (1.2 to 4.5)||10.3 (6.4 to 16.4)||0.6 (0.4 to 0.8)||0.8 (0.7 to 1.0)||0.3 (0.2 to 0.5)|
|African-American men||1.9 (1.0 to 3.2)||1.9 (1.2 to 2.9)||5.5 (3.3 to 9.1)||0.7 (0.4 to 1.0)||0.6 (0.3 to 0.9)||0.2 (0.1 to 0.5)|
|Women||9.4 (5.5 to 15.2)||12.3 (6.8 to 21.7)||13.5 (8.2 to 21.9)||0.4 (0.2 to 0.6)||0.4 (0.3 to 0.6)||0.2 (0.1 to 0.4)|
|Mexican-American men||3.4 (1.6 to 7.0)||2.9 (1.6 to 5.3)||3.4 (2.3 to 5.1)||0.7 (0.4 to 0.9)||0.6 (0.4 to 0.8)||0.1 (0.03 to 0.4)|
|Women||4.7 (1.7 to 11.6)||1.8 (0.6 to 4.9)||6.0 (3.9 to 9.0)||0.8 (0.6 to 1.0)||0.9 (0.7 to 1.0)||0.3 (0.1 to 0.5)|
*LRs defined in Glossary.
Steinbauer and colleagues provide a convincing study supporting the superiority of the AUDIT over the CAGE and the SAAST in detecting alcohol problems among white, African-American, and Mexican-American primary care patients. The AUDIT has other advantages: It is brief, it can be used as a pencil-paper waiting room screen, both "at risk" alcohol users and alcohol-dependent users are identified, the questions address issues of both consumption and harmful use, and it assesses both current and lifetime diagnostic time frames.
This study seems carefully planned and implemented and has relatively few methodologic flaws. The results will help clinicians working with white, African-American, or Mexican-American patients in most primary care settings. However, a broader challenge that this study does not address is the diagnosis of occult alcohol problems. Many patients with alcohol use disorders minimize or deny consumption or related problems to avoid adverse social consequences. All 3 screening tests and the diagnostic standard used in this study rely heavily on the integrity of patients' self-reports. They may well offer reproducible or comparable (i.e., reliable) results. However, their validity may be differentially diminished in ways this study (and most others) would not detect when used for patients with occult alcohol problems. Reduced validity may occur in occupational medicine settings (e.g., military medical settings) where open and honest disclosure may adversely affect the respondent. To improve detection of occult alcohol problems, a diagnostic standard should use multiple detection methods, such as patient reporting, biochemical markers, longitudinal assessment, and corroborating historians. Such studies are seldom done, however, as they are expensive and difficult and present ethical dilemmas. Without such studies, clinicians can place greater faith in the value of these self-report measures when they yield positive results. Negative results must be viewed with skepticism.
Rena Ferguson, MD, MPH
Charles C. Engel Jr., MD, MPHUniformed Services UniversityWashington, D.C., USA