OBJECTIVES: To determine whether ethnicity was associated with baseline and 18-month health status within a merged sample of older adults with knee osteoarthritis (OA) from the Fitness Arthritis in Seniors Trial and the Arthritis, Diet, and Activity Promotion Trial. DESIGN: Cross-sectional and prospective study. SETTING: Center-based exercise therapy at two universities. PARTICIPANTS: A total of 584 African-American (n = 143) and Caucasian-American (n = 441) adults aged 60 and older with knee OA were examined for baseline and 18-month health status. MEASUREMENTS: Six-minute-walk distance, 36-item Short Form General Health Scale (SF-36 GHS), and Physical Functioning Questionnaire index score. Ethnicity was obtained via self-report. RESULTS: Analyses of covariance testing the effect of ethnicity, adjusted for demographic and health status covariates, revealed significant effects for ethnicity upon baseline 6-minute-walk distance and SF-36 GHS, with Caucasian Americans reporting better scores (P = .001), although these differences were not significant after 18 months of exercise therapy. CONCLUSION: Ethnicity and baseline function are important factors that should not be overlooked in knee OA research involving exercise interventions. Moreover, not only should physical activity be recommended to improve functional outcomes, it may also be a useful strategy in reducing health disparities. © 2005 by the American Geriatrics Society.