pression, nonpharmacological methods have been shown to de- crease depressive symptomology as well as provide valuable skills for individuals. Resilience training (RT) is a program re- cently developed at the Penny George Institute for Health and Healing by holistic psychiatrist and author of Chemistry of Joy, Dr Henry Emmons. The RT program is an eight-week, skills build- ing, group-based program conducted at the George Institute Outpatient Clinic for persons with chronic depression. Mind- fulness meditation, nutritional counseling, and exercise assess- ments and recommendations are the tenets of the RT. The pur- pose of this study is to evaluate the psychological impact of the RT on Perceived Stress Scale (PSS), Center for Epidemiologic Study-Depression (CES-D), Spielberger’s State-Trait Anxiety In- ventory (STAI), and single-item questions related to sleep qual- ity and fatigue. A recent protocol amendment ensures future collection of biological parameters, including high-sensitivity C-reactive protein; vitamins D, B 6 , and B 12 ; essential fatty acids; DHEA, DHEA-S, testosterone; red blood cell folate and serum folate; and thyrotropin will be presented at the meeting. Evalu- ations of possible correlations between psychological and bio- logical measures will also be assessed in the forthcoming cohorts. Through September 30, 2008, 47 subjects have completed the pretraining and posttraining psychological questionnaires. Anal- yses indicate a significant reduction on PSS (9.1 (3.5 SD) to  5.4 (2.9 SD); P  .001), CES-D (16.1 (7.1 SD) - 5.5 (6.5); P  .001), State anxiety (48.2 (12.9 SD) - 37.7 (12.5 SD); P  .001), Trait anxiety (53.6 (13.3 SD) - 40.7 (11.0 SD); P  .001), sleep quality (5.1 (2.1 SD) - 6.7 (1.7); P  .001) and fatigue (5.9 (1.9 SD) - 4.4 (2.0 SD); P  .001. Results from this clinical inter- vention indicate that a skills building program improves psycho- logical functioning in moderate to severely depressed individu- als. Although not a randomized controlled trial, these results indicate significant improvements on self-reported measures and provide effect size estimates for future controlled studies. Robert Saper, Diana Cullum-Dugan, Larry Culpepper, Roger Davis, Paula Gardiner, Surya Karri, Russell Phillips, and Karen Sherman: PILOT RANDOMIZED CONTROLLED TRIAL OF HATHA YOGA FOR CHRONIC LOW BACK PAIN IN A PREDOMINANT MINORITY POPULATION The purpose of this study was to collect preliminary data on the feasibility, efficacy, and safety of hatha yoga for chronic low back pain in a predominantly urban minority population. We conducted a pilot randomized controlled trial of hatha yoga for 30 English-speaking adults recruited from two urban commu- nity health centers who had nonspecific low back pain lasting more than three months. An expert panel developed a standard- ized hatha yoga protocol by consensus after a systematic review of the lay and scientific literature on yoga and low back pain. All participants received an educational handout on low back pain and continued to receive their usual medical care. Subjects were block randomized to a weekly 75-minute yoga class for 12 weeks or a wait-list control. Home practice was encouraged through providing a CD of the protocol recorded for the study, hand- book with instructions and diagrams of all poses, and yoga sup- plies such as a mat, belt, and block. Subjects completed home practice logs and adverse-event report forms weekly. Feasibility outcomes included duration of recruitment, proportion of ra- cial/ethnic minority subjects, and retention rate. Primary effi- cacy outcomes at 12 weeks were self-reported low back pain score by using an 11-point numerical rating scale and function mea- sured by the 23-point modified Roland-Morris Low Back Pain Disability Questionnaire. Secondary outcomes measured at 12 weeks included global improvement and pain medication use during the previous week. Pain and function change scores (12 week minus baseline measurements) for the two groups were compared using the Wilcoxon rank sum test. Categorical data were compared with the Fisher exact test. All analyses were in- tention-to-treat by using last value carried forward. Thirty sub- jects were recruited, enrolled, and randomized in two months. Mean age was 44 years (SD 11.7), 83% were female, 83% were racial/ethnic minorities, 48% had annual household incomes $30,000, and 33% had high school education or less. Subject retention was 97%. Yoga subjects attended a median of seven classes (range, 1-11). Subjects’ mean low back pain and Roland- Morris disability scores at baseline were 7.1 (SD 1.7) and 15.3 (SD 4.5), respectively. Yoga subjects had a 2.3 (SD 2.1) point mean decrease in pain at 12 weeks compared with 0.4 (SD 1.8) points for the wait-list control group (P = .02). For the Roland- Morris disability questionnaire, the yoga group improved by 6.3 (SD 6.9) points compared with 3.6 (SD 5.0) points for controls (P = .26). In the yoga group, 73% reported their back condition was somewhat or very improved overall compared with 27% of controls (P = .03). At 12 weeks, 13% of yoga participants com- pared with 73% of controls reported pain medication use during the previous week (P = .003). One yoga participant discontinued classes after six weeks due to transient worsening of back pain from yoga. Her 12-week pain and function scores were improved compared with baseline. No other significant adverse events were reported. Conducting a randomized controlled trial of hatha yoga in an urban, predominantly minority, adult popula- tion with chronic low back pain is feasible. Yoga was relatively safe and more effective than usual care in reducing low back pain and pain medication use. Jeffrey Greeson, Jeffrey Brantley, Andrew Ekblad, Janna Fikkan, Thomas Lynch, Moria Smoski, and Ruth Wolever: MECHANISMS LINKING MINDFULNESS, EMOTIONAL EXPERIENCE, AND SLEEP QUALITY Restful sleep is instrumental in promoting mental, physical, and behavioral health. However, sleep can easily be disrupted by feelings of stress, anxiety, and depression. This cross-sectional study used structural equation modeling to test the hypothesis that (1) higher dispositional mindfulness—measured using the Cognitive and Affective Mindfulness Scale-Revised—is associ- ated with better sleep quality—measured by the Pittsburgh Sleep Quality Index, and (2) this relationship is mediated by emo- tional experience. The sample included 103 adults (mean age = 45, SD 12], 74% women, 87% white, 57% graduate level educa- tion) enrolled in a mindfulness-based stress reduction (MBSR) program at a large academic medical center. Data from precourse assessments were used for analysis. Bivariate correlations re- vealed a relatively strong relationship between higher levels of dispositional mindfulness and better sleep quality (ie, less sleep 161 Abstracts EXPLORE May/June 2009, Vol. 5, No. 3