Do Words Matter? Incongruent Responses to Inconsistently Worded AUDIT-C Alcohol Screening Instruments

ABSTRACT

The first 3 questions of the Alcohol Use Disorders Identification Test (AUDIT-C) are often used as a brief alcohol screening instrument. However, the implications of common modifications made to the original AUDIT questions and response options have not been considered. The authors examined existing data from a randomized controlled trial of 310 persons with human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS) that was testing the efficacy of 2 antiretroviral medication adherence interventions. Logistic regression was used to model the probability of participants having inconsistent AUDIT-C item responses. Three patterns of conflicting responses to the AUDIT-C items were identified. Common item modifications resulted in 14% (n = 48) of the parent study sample reporting conflicting responses across related AUDIT-C items. The odds of having conflicting data were 3 times greater in opioid users (odds ratio [OR] = 3.139, 95% confidence interval [CI] = 1.267–7.777, P = .01) and greater in persons with higher levels of conscientiousness (OR = 1.053, 95% CI = 1.006–1.103, P = .03). Inconsistent question format and response options may impede proper scoring and interpretation of the AUDIT-C. Further discussion and consensus building are needed on the psychometrically ideal version of the AUDIT-C.

KEYWORDS:

• Alcohol consumption
• alcohol-related disorders
• mass screening
• psychometrics

Acknowledgments

This study was financially supported by a Ruth L. Kirschstein National Research Service Award (NRSA) to Dr. Broyles from the National Institute of Nursing Research (NINR, F31NR008822) and by NINR grant R01NR004749 to Dr. Erlen. The material is the result of work supported with resources and the use of facilities at the VA Pittsburgh Healthcare System, Pittsburgh, Pennsylvania. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States government.

The authors gratefully acknowledge the editorial contributions of Dr. Kevin L. Kraemer, University of Pittsburgh School of Medicine, and the data management contributions of Ms. Lisa Tamres and Mr. Blair Powell at the University of Pittsburgh School of Nursing.

These results were previously presented at the Association for Medical Education and Research in Substance Abuse (AMERSA) Annual Conference, Bethesda, Maryland, November 5–7, 2009.