Background Osteoarthritis (OA) is highly prevalent in older adults and a growing cause of disability. Easily accessible first-line treatment of OA is increasingly important. Digital self-management programs have in recent years become available. Evidence of short-term effects of such programs are abundant, yet reports on long-term benefits and adherence to treatment are scarce. The current study’s objective was to investigate the long-term pain and function outcomes of people with hip or knee OA participating in a digital self-management programme. Methods and findings In this longitudinal cohort study, individuals with hip and knee OA, from the register of a digital self-management program and with 0-24-week (n = 499) or 0-48-week adherence (n = 138), were included. The treatment effect in terms of monthly pain (NRS, 0–10 worst to best) and physical function (30-second chair stand test (30CST), number of repetitions) change were investigated using a mixed model, controlling for the effect of age, body mass index (BMI), gender and index joint. For the 24-week sub-sample, pain NRS decreased monthly by -0.43 units (95% CI -0.51, -0.35, mean knee pain from 5.6 to 3.1, and hip pain from 5.9 to 3.8) and 30CST repetitions increased monthly by 0.76 repetitions (95% CI 0.64, 0.89 mean for knee from 10.0 to 14.3, and for hip from 10.9 to 14.8). For the 48-week sub-sample, pain decreased monthly by -0.39 units (95% CI -0.43, -0.36, mean knee pain from 5.7 to 3.2, and hip pain from 5.8 to 3.8), and repetitions increased by 0.72 repetitions (95% CI 0.65, 0.79, mean repetitions for knee from 10.3 to 14.4, and for hip from 11.1 to 14.9). There were no clinically relevant effects on the improvement of pain or function by any covariate (age, sex, index joint). The lack of a control group and randomization limit our ability to explain the mechanisms of the observed results. Conclusions Continuously participating in a digital OA treatment program for 6 or 12 months was associated with a clinically important decrease in joint pain and increased physical function, in hip and knee OA. Future research should follow OA-related outcomes in participants that end their treatment to explore when and why that decision was made.