JMIR Publications

ABSTRACT

Background:

Patients with rheumatoid arthritis (RA) face higher risks of major adverse cardiovascular events (MACE) than the general population. While machine learning offers powerful predictive capabilities, its clinical adoption is hindered by the "black-box" nature of complex algorithms.

Objective:

This study aimed to develop interpretable survival models to predict the risk of MACE in patients with RA, providing transparent and actionable insights for personalized clinical prognosis management.

Methods:

Utilizing data from the Taipei Medical University Clinical Research Database (2011–2022) for 2,461 patients with RA, machine learning survival models including Random Survival Forest (RSF), DeepSurv, and Cox-Time were compared with the traditional Cox-PH model. Performance was evaluated using the C-index and Integrated Brier Score (IBS). Permutation importance and SHAPley Additive exPlanations (SHAP) analyses were integrated to provide explainability for individual-level risk predictions.

Results:

RSF demonstrated superior performance, achieving a C-index of 0.8771 and an IBS of 0.0775. Permutation importance identified key features including creatinine (CRE), conventional synthetic disease-modifying antirheumatic drugs (csDMARDs), C-reactive protein (CRP), alanine aminotransferase (ALT), and age at diagnosis. SHAP analysis further quantified feature-specific effects, revealing both protective and risk-increasing associations of medications and laboratory indicators.

Conclusions:

RSF outperformed traditional methods by capturing non-linear and time-dependent relationships, supporting its potential for clinical decision support. Integrating SHAP enabled transparent and personalized risk interpretation, translating complex models into actionable insights. This approach empowers clinicians to identify high-risk individuals and advances precision medicine in rheumatology. Future work should focus on cross-database validation across diverse populations to enhance generalizability.