Purposes of the study and background. Smaller studies indicate that yoga may be an effective treatment for chronic low back pain. We conducted a randomised trial to evaluate if yoga compared to usual care improves back function in patients with chronic or recurrent low back pain. Summary of the methods used and the results. Outcomes were assessed by postal questionnaires. The setting was 13 non-National Health Service premises. We recruited 313 adults with chronic or recurrent low back pain from primary care. 157 were randomised to usual care. 156 were randomised to a 12-class, gradually-progressing programme of yoga delivered by 12 teachers over three months. All received The Back Book. Primary outcome was back function (Roland Morris Disability Score) at three months. Secondary outcomes: back function at six and 12 months, back pain, pain self-efficacy and general health. Back function improved more in the yoga group: mean difference in changes from baseline at three (−2.17, 95% CI −3.31 to −1.03, p<0.001), six (−1.48, 95% CI −2.62 to −0.33, p=0.011) and 12 months (−1.57, −2.71 to −0.42, p=0.007). Improvement in pain self-efficacy at three and six months in the yoga group. No differences in general health and pain reduction. Two adverse events were reported by controls and 12 by the yoga group – 8 out of 12 reported pain which may have been due to yoga. 63 (40%) were not fully compliant with treatment and 23 (15%) did not attend any yoga classes. Conclusion. At all time points Yoga was more effective than usual care at improving back function. Conflicts of Interest. AT, AS and JDA are self-employed yoga practitioners and may benefit if there is a demand for the yoga programme. AT, AS, the funders and the University of York would share any royalties from the future sale of the yoga manuals and CD.” AT, AS AND JDA had no involvement in the collection of data and in the analysis of the trial. Source of funding. Arthritis Research UK