The aim of this study was to examine optimal self-management in osteoarthritis and its association with patient-reported outcomes. We recruited a population-based sample of Medicare beneficiaries (n=551) residing in Allegheny County, PA, USA and elicited an expanded set of self-management behaviors using open-ended inquiry. We defined optimal self-management according to clinical recommendations, including use of hot compresses on affected joints, alteration of activity, and exercise. Only 20% practiced optimal self-management as defined by two or more of these criteria. Optimal and suboptimal self-managers did not differ in sociodemographic features. Both white and African–Americans who practiced optimal self-management reported significantly less pain, but the benefit was greatest in severe disease for whites and for mild-moderate disease among African–Americans. This backdrop of naturally occurring self-management behaviors may be important to recognize in planning programs that seek to bolster self-management skills.