This study was undertaken to determine the diagnostic test characteristics of the alcohol screening strategy recommended in the National Institute on Alcoholism and Alcohol Abuse (NIAAA) "Physicians' Guide to Helping Patients with Alcohol Problems." A research interview was performed on patients who presented to one urban emergency department (N = 395; 61% women). It asked three alcohol consumption questions, the CAGE questionnaire, and about past alcohol problems. The NIAAA-recommended screen was considered positive for alcohol consumption in excess of 14 drinks per week or 4 drinks per occasion for men, or 7 drinks per week or 3 drinks per occasion for women, or a CAGE score of 1 or greater. A sample of patients (n = 250) received the Composite International Diagnostic Interview substance abuse module, a gold standard interview, to determine lifetime or prior 12-month alcohol abuse or dependence; results were adjusted for verification bias. The prevalence of lifetime:alcohol abuse or dependence was 13%, for which the NIAAA strategy was 81% sensitive and 80% specific. The prevalence of alcohol abuse or dependence in the prior 12 months was 10%, for which the strategy was 83% sensitive and 84% specific. Its positive likelihood ratio exceeded that of the CAGE, augmented CAGE or consumption questions alone, and its negative likelihood ratio was the lowest. The screening strategy combining alcohol consumption and CAGE questions recommended in the NIAAA "Physicians' Guide" is valid, and has superior test characteristics compared to the CAGE alone, in this predominantly black (86%) emergency department population. Its brevity and simple interpretation recommend wider dissemination of the NIAAA "Physicians' Guide," although future research should examine its test characteristics in other clinical settings and with other populations.