A Machine Learning Model for Predicting Depression in Moroccan Rheumatoid Arthritis Patients

Area of Science:

• Rheumatology
• Artificial Intelligence
• Psychiatry

Background:

• Rheumatoid arthritis (RA) is a chronic inflammatory disease significantly affecting patient quality of life.
• Depression is a common comorbidity in RA, worsening pain and hindering remission.
• Predicting depression in RA is challenging due to resource limitations.

Purpose of the Study:

• To develop and evaluate machine learning (ML) models for predicting depression in RA patients.
• To identify key clinical features associated with depression in RA.
• To improve early detection and management of depression in RA.

Main Methods:

• A cohort of 112 RA patients was studied.
• Depression was assessed using the Hospital Anxiety and Depression Scale (HADS).
• Five ML models (SVM, RF, DT, LR, GBC) were developed using 12 clinical features, with data preprocessing including normalization and encoding. Performance was evaluated using accuracy, precision, recall, F1 score, and AUC.

Main Results:

• The logistic regression (LR) model demonstrated the highest predictive performance.
• The LR model achieved 76.5% accuracy, 72.2% precision, 81.2% recall, an F1 score of 0.765, and an ROC AUC of 0.767.
• ML-based feature selection identified optimal predictors for depression in RA.

Conclusions:

• ML models, particularly logistic regression, show significant promise for predicting depression in RA patients.
• The identified features and the LR model offer a valuable tool for early risk identification.
• Further validation and development of advanced ML models are recommended to enhance RA patient care and psychological support.