Comparing short versions of the Alcohol Use Disorders Identification Test (AUDIT) in a military cohort

Jason Watterson, Belinda J. Gabbe, Paul Dietze, Anna L. Bowring, Jeffrey V. RosenfeldJournal of the Royal Army Medical Corps 2018 · DOI

Abstract: ACTRN12614001332617.

Introduction

Alcohol misuse represents a significant public health challenge within military populations, frequently contributing to accidents, injuries, and long-term health complications. While the 10-item Alcohol Use Disorders Identification Test (AUDIT-10) is considered the gold standard for screening, its length can be prohibitive during rapid medical intakes or in comprehensive health surveys. Existing shortened versions, such as the three-item AUDIT-C, are widely utilized but may not be optimized for the unique risk profiles of young military personnel, whose drinking patterns often involve acute "binge" episodes rather than chronic dependence.

Methodology

This study involved a secondary analysis of data from 952 Royal Australian Navy trainees (80% male, median age 20.4). Researchers systematically evaluated all 330 possible three- and four-item combinations of the AUDIT-10 to identify the most efficient screening tool for this demographic.

To ensure clinical relevance, the evaluated versions were required to cover three domains: Alcohol Consumption (C), Alcohol Dependence (D), and Harmful Alcohol Use (H). Furthermore, based on the high risk of acute trauma in young populations, combinations including Item 9 ("Have you or someone else been injured because of your drinking?") were prioritized.

Table 1: AUDIT-10 questions and item inclusion in common established shortened versions

Performance was assessed using Cronbach’s alpha for internal consistency, the $R^2$ statistic to determine explained variance of the full AUDIT-10 score, and Pearson’s correlation coefficient for concurrent validity.

Results

The analysis identified a novel four-item combination—comprising Items 3, 4, 8, and 9—as the top-performing variant. This combination addresses the frequency of heavy drinking (6+ drinks), inability to stop drinking, memory loss (blackouts), and alcohol-related injuries.

Table 2: Performance metrics for selected novel and established AUDIT combinations

As shown in Table 2, this novel variant achieved an $R^2$ of 0.85 and a Pearson’s correlation of 0.92, significantly outperforming established short forms like the FAST and AUDIT-C. While the full AUDIT-10 had an alpha of 0.80, the top 4-item variant maintained a respectable alpha of 0.63. Within this cohort, 36% of trainees were classified as hazardous drinkers (score $\ge$ 8), highlighting the critical need for such rapid screening tools.

Discussion and Conclusion

The study demonstrates that a four-item screening tool can reduce the administrative burden by 60% while retaining approximately 85% of the statistical signal of the full AUDIT-10. The inclusion of Item 9 (alcohol-related injury) proved essential for capturing the specific risks faced by young military recruits.

However, the study is limited by the lack of a "gold standard" clinical diagnosis (e.g., DSM-5) for comparison and the inherent statistical overlap between the short versions and the full test. Future research should validate these novel variants against independent measures of alcohol use disorder to confirm their predictive validity in diverse military and veteran populations.

Summary updated

Abstract

ACTRN12614001332617.
BibTeX
@article{watterson2018comparing,
  author = {Jason Watterson and Belinda J. Gabbe and Paul Dietze and Anna L. Bowring and Jeffrey V. Rosenfeld},
  title = {{Comparing short versions of the Alcohol Use Disorders Identification Test (AUDIT) in a military cohort}},
  journal = {Journal of the Royal Army Medical Corps},
  year = {2018},
  doi = {10.1136/jramc-2018-001024},
}
Figure