MACHINE learning models can accurately predict which patients with painful knee osteoarthritis will benefit from short-term NSAID plus paracetamol therapy, according to new research. The findings highlight the promise of integrating biological, psychological, and clinical markers to improve personalized pain management.
Researchers assessed 101 patients undergoing three weeks of combined NSAID and paracetamol treatment. Before therapy, participants completed pain sensitivity testing using cuff algometry, psychological questionnaires such as the Hospital Anxiety and Depression Scale and Pain Catastrophizing Scale, and quality-of-life measures including EQ-5D-3L. Blood samples were also collected to evaluate inflammatory biomarkers and microRNA profiles.
Treatment response was measured using the pain subscale of the Knee Injury and Osteoarthritis Outcome Score before and after the three-week regimen. Data were analyzed with a multifactorial machine learning framework known as Data Integration Analysis for Biomarker discovery using Latent cOmponents (DIABLO). This approach integrated information from four domains: pain sensitization, psychological status, inflammatory response, and epigenetic mechanisms.
The model identified 30 key variables predictive of analgesic benefit. After cross-validation, the algorithm achieved strong performance, with an area under the precision-recall curve of 85%, sensitivity of 83%, specificity of 87%, and balanced accuracy of 85%.
The authors concluded that multifactorial modeling offers a promising strategy to identify patients most likely to respond to analgesic therapy, advancing the development of precision pain medicine. By combining markers of pain processing, inflammation, epigenetics, and psychological health, the algorithm moves beyond single-domain predictors to provide a more reliable assessment of treatment potential.
The study demonstrates that machine learning tools can harness complex biological and psychosocial data to support clinical decision-making. For patients with osteoarthritis, this approach may help target effective treatment, avoid ineffective regimens, and ultimately improve outcomes in a condition where pain control is often inconsistent.
Reference: Giordano R et al. Multifactorial machine learning algorithm integration of pain mechanisms can predict the efficacy of 3-week NSAID plus paracetamol in patients with painful knee osteoarthritis. Eur J Pain. 2025;29(10):e70140.
