BACKGROUND: There are limited data on the characteristics of yoga users in the U.S. OBJECTIVE: To characterize yoga users, medical reasons for use, perceptions of helpfulness, and disclosure of use to medical professionals. METHODS: Utilizing cross-sectional survey data from the 2002 National Health Interview Survey (NHIS) Alternative Medicine Supplement (n = 31044), we examined correlates of yoga use for health. The estimated prevalence from 2002 NHIS of yoga for health was 5.1% corresponding to over 10 million adults. RESULTS: In 2002, yoga users were predominately Caucasian (85%) and female (76%) with a mean age of 39.5 years. Compared to non-yoga users, yoga users were more likely female (OR 3.76, 95% CI 3.11-4.33); less likely black than white (OR 0.65, 95% CI 0.53-0.80); tended to be younger; and more likely college educated (OR 2.70, 95% CI 2.37-3.08). Musculoskeletal conditions (OR 1.61, 95% CI 1.42-1.83), mental health conditions (OR 1.43, 95% CI 1.22-1.67), severe sprains in the last 12 months (OR 1.49, 95% CI 1.22-1.81), and asthma (OR 1.27, 95% CI 1.05-1.54) were independently associated with higher yoga use, while hypertension (OR 0.78, 95% CI 0.64-0.95) and chronic obstructive lung disease (OR 0.69, 95% CI 0.48-1.00) were associated with lower use. Yoga was most commonly used to treat musculoskeletal or mental health conditions, and most users reported yoga to be helpful for these conditions. A majority of yoga users (61%) felt yoga was important in maintaining health, though only 25% disclosed yoga practice to their medical professional. CONCLUSIONS: We found that yoga users are more likely to be white, female, young and college educated. Yoga users report benefit for musculoskeletal conditions and mental health, indicating that further research on the efficacy of yoga for the treatment and/or prevention of these conditions is warranted.