Objective. Nonsurgical interventions are recommended for osteoarthritis (OA). However, how interventions change pain and physical function is unclear. Therefore, the objectives of this scoping review were to 1) identify what potential mediators of nonsurgical interventions on pain and physical function have been evaluated and 2) summarize the findings according to intervention, joint, and outcome.Methods. We searched Medline, Embase, CINAHL, Cochrane Central Register of Controlled Trials, and Scopus databases. Studies were included if they conducted a mediation analysis on a randomized controlled trial evaluating a nonsurgical intervention on OA of any joint. Outcomes were pain and physical function.Results. Nine knee OA studies, evaluating diet plus exercise, exercise, unloading shoes, high-expectation communication during acupuncture, and telephone-based weight loss plus exercise were identified. Except for weight loss and self-efficacy, putative mediators (knee muscle perfusion/extensor strength/adduction moment, systemic inflammatory biomarkers, physical activity, dietary intake, and pain beliefs) were evaluated by single studies. Ten mediators partially mediated intervention (diet plus exercise, exercise, high-expectation communication) effects on pain and physical function. Eight mediators were common to pain and function (reduced weight, increased knee extensor strength, and increased self-efficacy). Constant knee flexor muscle perfusion partially mediated exercise effects on pain, and knee pain relief partially mediated exercise effects on function.Conclusion. In knee OA, some evidence suggests that the benefits of 1) diet and exercise are mediated through changes in body weight, systemic inflammation, and self-efficacy; 2) exercise is mediated through changes in knee muscle strength and self-efficacy; and 3) high-expectation communication style is mediated through changes in selfefficacy.