Aim: The aim of this study was to examine the screening accuracy of four brief alcohol screening instruments in a general hospital setting. Methods: Ten outpatient clinics were recruited to ensure a representative mix of demographics (e.g. sex, age and diagnosis). The staff at the reception desk handed out pre-sorted envelopes with questionnaires and information letters to the visitors. The questionnaires consisted of the 10-item Alcohol Use Disorders Identification Test (AUDIT), the Weekly Consumption Question (WCQ), the Heavy Episodic Drinking (HED) question and two questions on sex and age. Sensitivity and specificity were calculated for the AUDIT-C, AUDIT-3, WCQ and HED using the full 10-item AUDIT for comparison, with cut-off points of >=6 for women and >=8 for men. Results: In all, 898 questionnaires were included (52% women). According to the full AUDIT, 12.0% of the women and 14.8% of the men were drinking above the hazardous level. Corresponding percentages for the brief screening instruments for women and men, respectively, were as follows: AUDIT-C: 17.2% and 27.4%; the AUDIT-3: 6% and 16.2%; the WCQ: 2% and 1.6%; and the HED screener: 12.9% and 21.2%. Conclusions: The AUDIT-C may be used as a brief screener in a general hospital setting. The WCQ, as a stand-alone screening tool, may underestimate hazardous drinking habits. Screening results from the AUDIT-3 and the HED should be interpreted with caution when applied to women because of the risk of underestimation.