Modifiable Determinants of Exercise Use in a Diverse Ethnic Population With Osteoarthritis.

OBJECTIVE: To determine the extent of ethnic differences in the use of exercise for therapy and identify relevant modifiable determinants of exercise use among patients with knee/hip osteoarthritis (OA). METHODS: Knee/hip OA study participants were identified. Surveys were administered to collect patient sociodemographic and clinical information, and beliefs and attitudes about providers and treatments. Final multivariable logistic regression models were created using a fully conditional method. RESULTS: Hispanic participants (n = 130), compared to non-Hispanic participants (n = 232), were less likely to have private medical insurance (9.2% versus 31.0%) or to report having excellent/very good health (40.7% versus 52.6%). They were also less likely to report using exercise for OA treatment in the last 6 months (56% versus 73%; P = 0.003). When adjusted for age and disease severity, the difference in exercise use among ethnicities remained significant (odds ratio [OR] 0.59 [95% confidence interval (95% CI) 0.36-0.99]). In a multivariable logistic regression model designed to determine the most important determinants of exercise use for OA treatment, in the last 6 months the following were all associated with exercise use: having knee instead of hip OA (OR 2.83 [95% CI 1.51-5.29]), having family/friends who exercise (OR 3.20 [95% CI 1.76-5.84]), having a good understanding of what happens after exercise (OR 2.19 [95 CI 1.15-4.19]), and higher perceived benefit of exercise (OR 2.24 [95% CI 1.64-3.04]). CONCLUSION: Among patients with knee/hip OA, Hispanics were less likely to exercise for OA treatment. Increased knowledge about the benefits of exercise for treatment and improved familiarity with exercise as treatment for OA may increase exercise use.