Decreasing sensitivity of clinical alcohol screening with the AUDIT-C after repeated negative screens in VA clinics.

Abstract

Routine screening for unhealthy alcohol use is widely recommended in primary care settings. However, the validity of repeat screening among patients who have previously screened negative remains unknown. This study aims to evaluate the performance of a clinical alcohol screen compared to a confidential comparison alcohol screen among patients with previous negative alcohol screens. This study included four nested samples of Veteran Health Administration (VA) outpatients with at least one (N=18,493) and up to four (N=714) prior negative annual clinical AUDIT-C screens who completed the AUDIT-C the following year, both in a VA clinic (clinical screen) and on a confidential mailed survey (comparison screen). AUDIT-C screens were categorized as either negative (0-3 points men; 0-2 women) or positive (≥4 men; ≥3 women). For each sample, the performance of the clinical screen was compared to the comparison screen, the reference measure for unhealthy alcohol use. The sensitivity of clinical screens decreased as the number of prior negative screens in a sample increased (40.0-17.4%) for patients with 1-4 negative screens. The positive predictive value also decreased as the number of prior negative screens in a sample increased (67.7-33.3%) while specificity was consistently high for all samples (≥97.8%). Repeat clinical alcohol screens became progressively less sensitive for identifying unhealthy alcohol use among patients who repeatedly screened negative over several years. Alternative approaches for assessing unhealthy alcohol use may be needed for these patients.