OBJECTIVE: To test the efficacy of a questionnaire for tuberculosis screening in an Alaskan homeless population. METHODS: Simultaneous PPD skin testing and questionnaire administration was performed at a free-meal facility to test the efficacy of a questionnaire for tuberculosis screening. "Positive" questionnaires were based on responses to symptom-based and risk-based questions. RESULTS: Demographic data was similar to pre-existing data for the Anchorage homeless population. Of the 64 participants, 3 were excluded and 47 (77%) returned for skin test readings. There were 8 (17%) positive, and 39 (83%) negative PPD tests, 17 (36%) positive and 30 (64%) negative symptom-based questionnaires (Sensitivity 25%, specificity 51%. PPV 13%, NPV 80%) and 28 (60%) positive, and 19 (40%) negative (n = 61) risk-based questionnaires (Sensitivity 63%, specificity 41%, PPV 18%, NVP 84%). CONCLUSION: The prevalence of TB warrants screening and treatment of the Anchorage, Alaska homeless population. However, this questionnaire was not useful for this population and therefore screening should be done using PPD skin tests.