Comparing Alcohol Screening Measures Among HIV‐Infected and ‐Uninfected Men

Abstract

Brief measures of unhealthy alcohol use have not been well validated among people with HIV. We compared the Alcohol Use Disorders Identification Test (AUDIT) to reference standards for unhealthy alcohol use based on 30-day Timeline Follow Back (TLFB) and Composite International Diagnostic Interview-Substance Abuse Module (CIDI-SAM), among 837 male HIV-infected and -uninfected patients in the Veterans Aging Cohort Study. Three reference standards were (i) Risky drinking-based on TLFB >14 drinks over 7 consecutive days or >4 drinks on 1 day; (ii) Alcohol dependence-based on a CIDI-SAM diagnosis; and (iii) Unhealthy alcohol use-risky drinking or a CIDI-SAM diagnosis of abuse or dependence. Various cutoffs for the AUDIT, AUDIT-C, and heavy episodic drinking were compared with the reference standards. Mean age of patients was 52 years, 53% (444) were HIV-infected, and 53% (444) were African American. Among HIV-infected and -uninfected patients, the prevalence of risky drinking (14 vs. 12%, respectively), alcohol dependence (8 vs. 7%), and unhealthy alcohol use (22 vs. 20%) was similar. For risky drinking and alcohol dependence, multiple cutoffs of AUDIT, AUDIT-C, and heavy episodic drinking provided good sensitivity (≥80%) and specificity (≥90%). For unhealthy alcohol use, few cutoffs provided sensitivity ≥80%; however, many cutoffs provided good specificity. For all 3 alcohol screening measures, sensitivity improved when heavy episodic drinking was included with the cutoff. Sensitivity of measures for risky drinking and unhealthy alcohol use was lower in HIV-infected than in uninfected patients. For identifying risky drinking, alcohol dependence, and unhealthy alcohol use, AUDIT-C performs as well as AUDIT and similarly in HIV-infected and -uninfected patients. Cutoffs should be based on the importance of specific operating characteristics for the intended research or clinical use. Incorporating heavy episodic drinking increased sensitivity for detecting alcohol dependence and unhealthy alcohol use.