PURPOSE: To determine if a pre-operative group rehabilitation programme can improve arthritis self-efficacy for people with severe hip and knee osteoarthritis. METHOD: Single group, repeated measures design: 4-week baseline phase followed by a 6-week intervention phase of water exercise, and education with self-management strategies. The primary outcome was arthritis self-efficacy. The secondary outcomes were measures of pain (WOMAC), activity limitation (WOMAC), activity performance (30 s chair stand test, 10 m walk test) and health-related quality of life (EuroQol). RESULTS: Twenty participants (10 knee osteoarthritis and 10 hip osteoarthritis) with a mean age of 71 years (SD 7) attended 92% (SD 10%) of the scheduled sessions. All measures demonstrated baseline stability between two time points for measurements at week 1 and measurements at week 4. After the 6-week intervention programme there were no significant improvements for arthritis self-efficacy. There was a 12% increase for fast walking speed (mean increase of 0.14 m/s, 95% CI 0.07, 0.22). There were no significant improvements for other secondary outcomes. CONCLUSIONS: A pre-operative water-based exercise and educational programme did not improve arthritis self-efficacy, self-reported pain and activity limitation, and health-related quality of life for people with hip and knee osteoarthritis who were candidates for joint replacement. While there was a significant increase in one measure of activity performance (walking speed), these findings suggest the current programme may be of little value. Implications for Rehabilitation This pre-operative group rehabilitation programme for people with severe hip and knee osteoarthritis did not change arthritis self-efficacy, pain, activity limitation and health-related quality of life. This programme may have little value in preparing people for joint replacement surgery. The optimal pre-operative programme requires further design and investigation.