In Reply: We agree with Drs Dervaux and Laqueille that using a score of 2 or more for the CAGE questionnaire increases specificity while decreasing sensitivity when screening patients for alcohol abuse or dependence.1,2 In a study of alcohol dependence, higher specificity may be more desirable. However, the purpose of our study was to investigate changes in alcohol use behaviors and alcohol-related problems before and after combat deployment, using the metrics of heavy weekly drinking, binge drinking, and alcohol-related problems defined by the Patient Health Questionnaire.3 The CAGE screening questions in our study were used to adjust for confounding related to the existence of potential alcohol-related problems prior to the baseline questionnaire. We elected to use a more sensitive definition of 1 or more affirmative responses on the CAGE, which was the cutoff established when the questions were first developed and has been deemed useful in other studies,4,5 to improve our ability to identify individuals possibly at increased risk for the 3 primary outcomes studied. Thus, we were careful to use the term potential alcohol dependence because of the sensitive criteria involved in using a tool for screening (rather than a thorough clinical evaluation).
Jacobson IG, Smith TC, Bell NS. Military Combat Deployment and Alcohol Use—Reply. JAMA. 2008;300(22):2606-2607. doi:10.1001/jama.2008.765