OBJECTIVE: To explore how lifestyle and demographic, socioeconomic, and disease-related factors are associated with supervised exercise adherence in an osteoarthritis (OA) management program and the ability of these factors to explain exercise adherence. METHODS: A cohort register-based study on participants from the Swedish Osteoarthritis Registry who attended the exercise part of a nationwide Swedish OA management program. We ran a multinomial logistic regression to determine the association of exercise adherence with the abovementioned factors. We calculated their ability to explain exercise adherence with the McFadden R . RESULTS: Our sample comprises 19,750 participants (73% female, mean ± SD age 67 ± 8.9 years). Among them, 5,862 (30%) reached a low level of adherence, 3,947 (20%) a medium level, and 9,941 (50%) a high level. After a listwise deletion, the analysis was run on 16,685 participants (85%), with low levels of adherence as the reference category. Some factors were positively associated with high levels of adherence, such as older age (relative risk ratio [RRR] 1.01 [95% confidence interval (95% CI) 1.01-1.02] per year), and the arthritis-specific self-efficacy (RRR 1.04 [95% CI 1.02-1.07] per 10-point increase). Others were negatively associated with high levels of adherence, such as female sex (RRR 0.82 [95% CI 0.75-0.89]), having a medium (RRR 0.89 [95% CI 0.81-0.98] or a high level of education (RRR 0.84 [95% CI 0.76-0.94]). Nevertheless, the investigating factors could explain 1% of the variability in exercise adherence (R = 0.012). CONCLUSION: Despite the associations reported above, the poorly explained variability suggests that strategies based on lifestyle and demographic, socioeconomic, and disease-related factors are unlikely to improve exercise adherence significantly.