BACKGROUND: Exercise is recommended for patients with hip/knee osteoarthritis (OA), however, adherence to exercise tends to be poor. This study aimed to (1) compare in-person versus app-based adherence to exercise therapy, (2) identify patient characteristics associated with adherence, and (3) examine associations between adherence and change in disease-specific outcomes in hip/knee OA patients. METHOD: Data were collected as part of a randomized controlled trial. In-person physiotherapy treatment involved supervised exercise therapy twice weekly for 6 weeks, complemented by an additional weekly home-exercise session. The app group received an individually tailored exercise program in the Virtual Training-app for 6 weeks and were advised to exercise three times weekly. Adherence was recorded as number of exercise sessions attended. Patient characteristics were reported at baseline. Changes in disease-specific outcomes were evaluated at 6-weeks follow-up using Hip/Knee Injury and Osteoarthritis Outcome Score. Associations were assessed using multiple logistic and linear regression analyses. RESULTS: In total 68 patients, 34 in each group were included (63 years, 69% female). The odds of adherence were 4.2 times higher when exercise was supervised by physiotherapists (95% CI 1.5, 12.4; p = 0.008). High fatigue was associated with lower adherence (OR 0.8, p = 0.049), whereas higher education (OR 5.2, p = 0.02) and self-efficacy (OR 3.2, p = 0.008) were associated with higher adherence. No significant associations were found between disease-specific outcomes and adherence. CONCLUSION: App-based exercise did not improve adherence compared to supervised exercise. Apart from fatigue, education, and self-efficacy, no factors were found to influence adherence to exercise in hip/knee OA patients. TRIAL REGISTRATION: ClinicalTrials.gov (NCT04767854).