Explainable Artificial Intelligence for Early Detection of Chronic Diseases in Saudi Healthcare: A Systematic Review
Keywords:
explainable artificial intelligence; chronic disease; early detection; machine learning; Saudi Arabia; digital healthAbstract
Background: Explainable artificial intelligence (XAI) may support earlier detection of chronic disease, but Saudi-specific evidence has not been systematically characterized. Objective: To synthesize Saudi patient-level studies of AI/XAI for chronic-disease prediction and assess clinical readiness.
Methods: PubMed and supplementary publisher-site, reference, and citation searches were completed on 26 August 2026. Eligible primary studies used Saudi clinical data for prediction, classification, or risk stratification. Findings were narratively synthesized and appraised with an adapted PROBAST framework.
Results: Fifty-eight records were identified, 52 screened, nine assessed in full, and four included. Studies addressed chronic kidney disease, diabetic kidney disease, hypertension, and right-sided cardiac dysfunction. Reported discrimination was promising, with explicit SHAP-based explanations in three studies; however, all lacked independent external validation and had high overall risk of bias.
Conclusion: Saudi XAI research is promising but clinically preliminary. Multicentre external validation, calibration, fairness evaluation, and prospective impact studies are required before routine deployment.
