Background: Physical and psychological symptoms limit physical activity for people with arthritis. This study examined if self-efficacy mediated a relationship between symptom and physical activity (PA) frequency change. Methods: This was a secondary analysis of older adults with arthritis and joint pain in a trial of a lifestyle PA program (n = 339). Measures were depressive symptoms, pain, fatigue, arthritis self-efficacy, PA self-efficacy, and PA frequency. A panel model was used to analyze relationships at baseline and changes at 20 weeks. Results: The mean age was 68.8 years. At baseline, depression and fatigue were associated with arthritis self-efficacy (beta = -.34 and -.24) and, in turn, PA self-efficacy (beta = .63); PA self-efficacy was associated with PA (beta = .15). Pain and depression changes were associated with arthritis self-efficacy change (beta = -.20 and -.21) and, in turn, PA self-efficacy (beta = .32) change; PA self-efficacy change was associated with PA change (beta = .36). Conclusion: Change in symptom severity affected change in PA frequency. These relationships appeared to operate through self-efficacy. Over time, pain appeared to have a stronger relationship than fatigue with self-efficacy and PA. These findings support strategies to help people with arthritis strengthen their confidence for symptom coping and PA participation
