BACKGROUND:The recommended initial treatment for multidirectional instability (MDI) of the shoulder is a rehabilitation program, yet there is very low-quality evidence to support this approach. Purpose/Hypothesis: The purpose was to compare the Watson MDI program and Rockwood Instability program among patients with nontraumatic, nonstructural MDI. The hypothesis was that the Watson MDI program would produce clinically and statistically superior outcomes over the Rockwood Instability program. STUDY DESIGN:Randomized controlled trial; Level of evidence, 2. METHODS:Forty-one participants with MDI were randomly allocated to the Watson MDI or Rockwood Instability program. Participants attended 12 weekly physiotherapy sessions for exercise prescription. Outcomes were assessed at baseline and 6, 12, and 24 weeks after randomization. Primary outcomes were the Melbourne Instability Shoulder Score (MISS) and the Western Ontario Shoulder Index (WOSI). Secondary outcomes included the Orebro Musculoskeletal Pain Questionnaire, pain, muscle strength, scapular upward rotation, scapular coordinates, global rating of change, satisfaction scales, limiting angle in abduction range, limiting factor in abduction range, and incidence of dislocation. Primary analysis was by intention to treat based on linear mixed models. RESULTS:Between-group differences showed significant effects favoring the Watson program for the WOSI (effect size [ES], 11.1; 95% CI, 1.9-20.2; P = .018) and for the limiting factor in abduction (ES, 0.1; 95% CI, 0.0-1.6; P = .023) at 12 weeks, and for the WOSI (ES, 12.6; 95% CI, 3.4-21.9; P =. 008), MISS (ES, 15.4; 95% CI, 5.9-24.8; P = .002), and pain (ES, -2.0; CI: -2.3 to -0.7, P = .003) at 24 weeks. CONCLUSION:For people with MDI, 12 sessions of the Watson MDI program were more effective than the Rockwood program at 12- and 24-week follow-up. Registration: ACTRN12613001240730 (Australian New Zealand Clinical Trials Registry).