A systematic review was conducted to identify and describe existing models for predicting knee pain in patients with knee osteoarthritis; the study involved searching multiple electronic databases up to May 2023, resulting in the identification of 2,693 records. After screening, 16 articles reporting on 26 prediction models were included. These models targeted various aspects of knee pain, including occurrence, progression, persistence, incident pain, frequent pain, and flares. The most common predictors in these models were age, BMI, gender, baseline pain, and joint space width. However, most studies (94%) were found to be at high risk of bias, and the quality of evidence for most types of knee pain predictions was low, with only frequent knee pain showing moderate-quality evidence.
The review concludes that while there are many prediction models for knee pain in patients with knee osteoarthritis that show potential, their clinical applicability and interpretability need further consideration. The models’ accuracy, as measured by AUROC, ranged from 0.62 to 0.81, indicating varying levels of predictive performance.
Reference: Tong B, Chen H, Wang C, et al. Clinical prediction models for knee pain in patients with knee osteoarthritis: a systematic review. Skeletal Radiol. 2024 Jun;53(6):1045-1059. doi: 10.1007/s00256-024-04590-x. Epub 2024 Jan 24. PMID: 38265451.
