Background: Patients commonly experience impaired sleep throughout cancer treatment and for years into survivorship. Impaired sleep may mediate other cancer-related symptoms and can lead to the inability to complete daily activities and lower quality of life. More effective non-pharmacological treatment options for impaired sleep are needed. We conducted a nationwide, multicenter, phase III randomized controlled trial (RCT) comparing the effects of yoga (Yoga for Cancer Survivors; YOCAS), cognitive behavioral therapy for insomnia (CBT-I), and a behavioral placebo on impaired sleep in cancer survivors. Methods: This RCT was conducted via the URCC NCORP Research Base. Participants were cancer survivors 2-60 months post-treatment with insomnia. They were randomized to 1) YOCAS (75-min session biweekly for 4 wks), 2) CBT-I (90-min session weekly for 8 wks), and 3) behavioral placebo (survivorship health education per ASCO guidelines; 75-min session biweekly for 4 wks). Sleep efficiency, sleep duration, wake after sleep onset (WASO), and sleep latency were assessed via actigraphy at baseline and post-intervention. Actigraphs were worn on the non-dominant wrist 24 hours a day for 7 days. Linear mixed models were used to assess intervention effects on sleep outcomes. Results: 740 survivors were enrolled (93% female, mean age 56±11 years, 73% breast cancer). Results revealed significant group differences among survivors in the 3 arms in sleep efficiency, sleep duration, and WASO (all p<0.05), but not in sleep latency (p>0.05). YOCAS and CBT-I subjects maintained sleep efficiency (mean change= -0.8% and -0.03%, respectively, all p>0.05) while behavioral placebo subjects significantly reduced sleep efficiency (mean change= -3.4%, p<0.01). When controlling for baseline, YOCAS and CBT-I subjects demonstrated better sleep efficiency compared to behavioral placebo subjects at post-intervention (all p<0.05). YOCAS subjects also maintained sleep duration (mean change= -3.5 minutes, p>0.05) while CBT-I and behavioral placebo subjects significantly reduced sleep duration (mean change= -20.3 minutes and -26.6 minutes, respectively, all p<0.01). When controlling for baseline, YOCAS subjects demonstrated longer sleep duration compared to CBT-I and behavioral placebo subjects at post-intervention (all p<0.05). There were no significant within-group changes in WASO over time in the 3 arms. When controlling for baseline, CBT-I subjects demonstrated a trend toward lower WASO compared to YOCAS (p=0.07) and behavioral placebo (p=0.05) subjects at post-intervention. Conclusions: Both YOCAS and CBT-I maintained sleep efficiency and/or sleep duration among cancer survivors. Oncologists should consider prescribing yoga and CBT-I for treating impaired sleep in cancer survivors. Funding: NCI UG1CA189961, R01CA181064, T32CA102618. Clinical trial information: NCT02613364