INTRODUCTION: Education and exercise are poorly implemented in osteoarthritis care. The purpose of the present study was to identify predictors of effectiveness at one year from education and exercise in patients with knee or hip pain in clinical practice. MATERIAL AND METHODS: Good Life with Arthritis in Denmark is an implementation initiative consisting of education and 12 sessions of neuromuscular exercise delivered by trained physiotherapists. Pain (visual analogue scale 0-100) and quality of life (EQ-5D) were assessed at baseline and after three and 12 months. Changes in outcome from baseline to one year were dependent variables in regression analyses, while changes from baseline to 3 months and absolute results at three months in physical performance (30-second chair stand test) and self-efficacy were predictor variables. RESULTS: A total of 79/82 patients completed the one-year follow-up. Improvements in pain and EQ-5D at three months were maintained at one year (p < 0.006). Change in self-efficacy from baseline to three months (Beta = -0.369) and 30-second chair stand test (Beta = -0.251) and self-efficacy at three months (Beta = -0.492) were predictors of one-year improvement in pain (p < 0.05). Furthermore, self-efficacy at three months (Beta = 0.304) was a predictor of one-year improvement in EQ-5D (p = 0.01). CONCLUSION: The identified predictors highlight the importance of combining education and exercise. This study indicates that good long-term treatment results are achievable in clinical practice.