Objective: To compare the effectiveness of a continence pessary to evidence-based behavioral therapy for stress incontinence and to assess whether combined pessary and behavioral therapy is superior to single-modality therapy. Methods: This was a multisite, randomized clinical trial (Ambulatory Treatments for Leakage Associated with Stress Incontinence [ATLAS]) that randomly assigned 446 women with stress incontinence to pessary, behavioral therapy, or combined treatment. Primary outcome measures, at 3 months, were Patient Global Impression of Improvement and the stress incontinence subscale of the Pelvic Floor Distress Inventory. A priori, to be considered clinically superior, combination therapy had to be better than both single-modality therapies. Outcome measures were repeated at 6 and 12 months. Primary analyses used an intention-to-treat approach. Results: At 3 months, scores from 40% of the pessary group and 49% of the behavioral group were "much better" or "very much better" on the Patient Global Impression of Improvement (P=.10). Compared with the pessary group, more women in the behavioral group reported having no bothersome incontinence symptoms (49% compared with 33%, P=.006) and treatment satisfaction (75% compared with 63%, P=.02). Combination therapy was significantly better than pessary as shown on the Patient Global Impression of Improvement (53%, P=.02) and Pelvic Floor Distress Inventory (44%, P=.05) but not better than behavioral therapy; it was therefore not superior to single-modality therapy. Group differences were not sustained to 12 months on any measure, and patient satisfaction remained above 50% for all treatment groups. Conclusion: Behavioral therapy resulted in greater patient satisfaction and fewer bothersome incontinence symptoms than pessary at 3 months, but differences did not persist to 12 months. Combination therapy was not superior to single-modality therapy. © 2010 by The American College of Obstetricians and Gynecologists. Published by Lippincott Williams & Wilkins.