Rapid and accurate health literacy screening tools could assist interventions to improve care in the HIV clinic setting. Prior studies described brief screening questions for identification of persons with low health literacy, but the performance of these questions in HIV clinics is unknown. We examined the accuracy of provider perception and previously described brief screening questions for identification of low health literacy among 147 persons attending two HIV specialty clinics. Mean age was 41 and 48% were African American. Using the short Test of Functional Health Literacy in Adults (S-TOFHLA) as reference measure, the combined prevalence of marginal and inadequate health literacy (low health literacy) was 16%. Sensitivity and specificity of provider perception for identifying persons with low health literacy were 0.47 and 0.79. The screening question "How confident are you filling out medical forms by yourself?" had an area under the receiver operator characteristic curve (AUROC) of 0.67. Response of "somewhat" to this question had sensitivity 0.30, specificity 0.91, positive likelihood ratio 3.4 (95% confidence interval [CI] 1.8-6.4) and negative likelihood ratio 0.76 (95% CI 0.4-1.4). Educational attainment was a more sensitive but less specific measure of low health literacy (sensitivity 0.90, specificity 0.61 for completed high school or fewer years education), but was overall a more accurate discriminator of health literacy than either provider perception or screening questions (AUROC 0.79; p<0.05 for comparison of AUROCs). HIV providers often misjudge patients' health literacy. In the absence of rapid and accurate health literacy screening tools, clinics caring for persons with HIV/AIDS should implement clinic-wide interventions to improve health communication. Copyright © 2010, Mary Ann Liebert, Inc.