INTEGRATED DIGITAL HEALTH AND EARLY RISK PREDICTION MODEL TO ENHANCE PATIENT SAFETY AND OPERATIONAL READINESS IN RIYADH'S HOSPITAL
Keywords:
Artificial intelligence; digital health; patient safety; early warning systems; emergency department overcrowding; surgical safety; heat-related illness; Saudi Arabia; Vision 2030Abstract
Background: Rapid urbanisation in Saudi Arabia, coupled with the ambitions of Vision 2030 healthcare transformation, has intensified demands on tertiary hospitals in Riyadh to enhance patient safety and operational efficiency. However, fragmented data systems and predominantly reactive clinical decision-making continue to contribute to preventable complications and emergency department (ED) overcrowding.
Objective: This study aimed to design, implement, and evaluate an integrated digital health platform incorporating machine learning–based early risk prediction across multiple clinical and operational domains in tertiary care settings.
Methods: A prospective, multi-phase mixed-methods study was conducted at King Fahad Medical City and King Abdulaziz Medical City between January 2024 and December 2025. In Phase 1, longitudinal data were collected across six domains: emergency visits, surgical complications, vision screening, National Early Warning Score (NEWS2), staff education attendance, and heat-related illness. Phase 2 involved the development of an ensemble artificial intelligence model integrating six domain-specific algorithms. In Phase 3, the platform was deployed across relevant clinical and operational units. Phase 4 evaluated its impact using paired t-tests and a Composite Operational Readiness Score (ORS).
Results: The ensemble model demonstrated excellent predictive performance (AUC-ROC = 0.961). Following implementation, significant improvements were observed across key indicators: surgical site infection rates decreased by 39.6%, ED length of stay was reduced by 24.1%, heat-related ICU admissions declined by 54.8%, and staff training compliance increased from 61.4% to 84.7%. The ORS showed a statistically significant improvement from 64 to 82 (p < 0.001).
Conclusion: The implementation of a unified AI-driven digital health platform was associated with substantial reductions in clinical complications, improved patient flow, and enhanced operational readiness. These findings support the scalability of such integrated solutions across the Saudi Ministry of Health system and similar resource-intensive healthcare environments.
