<![CDATA[Osteoarthritis (OA) is a degenerative joint disease often accompanied by multimorbidity, particularly cardiometabolic diseases. OA is also associated with comorbidities, thus requiring an analytical approach capable of identifying patterns of relationships between diseases and rational therapies. This study aims to explore patterns of multimorbidity and treatment patterns in hospitalized patients with osteoarthritis using a machine learning (ML) approach. This study employed a retrospective design using medical records of hospitalized OA patients from January 2020 to January 2025 at Sultan Agung Islamic Hospital in Semarang. Analysis was performed using the Frequent Pattern Growth (FP-Growth) algorithm with support, confidence, and lift parameters. The minimum support value was set at 1% to identify a wider variety of patterns. A total of 25 patients were analyzed, with the majority being female (14 patients; 56%) and aged ≥59 years (14 patients;96%), with comorbidities predominantly obesity and hypertension. Association Rule Mining (ARM) results showed cardiometabolic multimorbidity patterns, with the strongest association in the combination OA+HTàDM (lift 1.93). Therapy pattern analysis indicated that disease combinations were associated with the use of therapies such as NSAIDs for OA and metformin for diabetes, as well as the addition of adjuvant therapies. Multimorbidity patterns in hospitalized OA patients are dominated by the cardiometabolic group, with complex therapeutic regimens. ML approaches are effective in identifying patterns of disease and therapy relationships, therapy supporting more rational clinical decision-making.]]>
