Construction and validation of the prediction model for kinesiophobia in older adults with chronic low back pain

Abstract

Background: Low back pain imposes a substantial burden on global healthcare systems. Kinesiophobia is highly prevalent among older adults with chronic low back pain, severely hindering effective intervention and treatment. However, current assessment of kinesiophobia in this population remain inadequate, necessitating further research. This study aimed to develop a predictive model to identify key factors influencing kinesiophobia in older adults with chronic low back pain, thereby assisting clinicians in assessment and interventions. Methods: A cross-sectional study was conducted among 386 older adults with chronic low back pain admitted to the Department of Spinal Surgery and Orthopedic Rehabilitation Center between January 2024 and December 2024. The dataset was randomly split into training (70%) and testing (30%) sets. Six machine learning models, including Logistic Regression(LR), Decision Tree(DT), Random Forest(RF), eXtreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LightGBM) and Artificial Neural Network(ANN), were developed and evaluated. Model performance was assessed using the area under the receiver operating characteristic curve (AUROC), calibration curves, and other metrics. Clinical utility was examined via decision curve analysis, and predictor significance was interpreted using Shapley additive explanations (SHAP). Results: The prevalence of kinesiophobia in the cohort was 65.3%. LightGBM and RF demonstrated robust performance across both training and testing sets. In the testing set, LightGBM achieved superior accuracy, precision, sensitivity, specificity, and F1-score. Key predictors included Oswestry Disability Index (ODI), moderate-to-severe anxiety, pain intensity, moderate-to-severe depression, frailty, smoking, and the history of falls. Conclusion: This study developed a machine learning-based predictive model for assessing kinesiophobia in older adults with chronic low back pain, which providing a reference for kinesiophobia assessment in this patients group and also identifying key intervention priorities for healthcare systems in kinesiophobia management.