Development and Evaluation of a Web-Based Clinical Decision Support System for Shoulder Pain Risk Stratification Integrating SPADI and Machine Learning

Shoulder pain is among the most prevalent musculoskeletal disorders and is frequently associated with functional limitations, reduced quality of life, and increased healthcare utilization. Early assessment of symptom severity and timely intervention are essential for preventing progression to chronic disability; however, access to standardized and accessible assessment tools remains limited in many healthcare settings. Advances in digital health and machine learning provide new opportunities for developing intelligent clinical decision support systems that facilitate preliminary assessment and personalized care planning. This study presents the development and evaluation of a web-based Clinical Decision Support System (CDSS) for shoulder pain risk stratification and management recommendation. The proposed framework is based on an adapted version of the validated Shoulder Pain and Disability Index (SPADI) questionnaire and combines rule-based clinical reasoning with a preliminary interpretable Logistic Regression component to demonstrate the integration of machine learning within the decision-support framework. The system integrates SPADI pain and disability scores, selected medical risk factors, and longitudinal patient information to classify individuals into low, moderate, and high-risk categories while generating personalized management recommendations. The platform was implemented using a responsive web architecture to support real-time assessment, longitudinal monitoring, and accessibility across multiple devices. To enhance transparency and interpretability, the machine learning component provides probability-based predictions and feature-level explanatory information that highlights the factors contributing to each risk classification. A preliminary technical evaluation of the machine learning component was performed using synthetically generated assessment data and standard classification metrics, including accuracy, precision, recall and F۱-score. System implementation and functional testing demonstrated successful questionnaire processing, automated score computation, risk prediction, recommendation generation, and patient progress tracking. The results indicate that the proposed framework can serve as a scalable and low-cost digital health solution for preliminary shoulder pain assessment while maintaining transparency, interpretability, and clinical usability. This work contributes to the integration of interpretable machine learning and web technologies in musculoskeletal healthcare and provides a foundation for future clinically validated decision support systems for shoulder pain assessment and management.

کلیدواژه ها:

Clinical Decision Support System ، Machine Learning ، Logistic Regression ، Digital Health ، Risk Stratification ، SPADI ، Shoulder Pain