Artificial Intelligence and Digital Health Worldwide – Building Connected and Resilient Hospitals and Healthcare Centres

Three Criteria for Understanding Digital Maturity

A healthcare organisation may have sophisticated software yet struggle to exchange information with other providers. Understanding a health system’s digital maturity requires examining three complementary capabilities: the digital systems available to healthcare facilities, patients’ access to their records, and the availability and secure use of health data.

International indicators shed light on these capabilities, but their scope, reference years and methodologies differ. A digital-access score does not necessarily measure actual patient use. An assessment of institutional readiness does not establish how widely hospitals are equipped with digital systems. Among the references presented, there is therefore no recent global ranking that combines all three criteria using a single methodology. For decision-makers, the priority is to compare specific capabilities and identify improvements relevant to their context.

Five Foundations That Must Work Together

• Electronic health records
• Hospital information systems
• Interoperability
• Data sovereignty and security
• Responsible AI

Electronic health records, hospital information systems, interoperability, data sovereignty and security, and responsible AI form an interconnected foundation. Structured records support continuity of care; hospital systems coordinate activities; interoperability enables information to move between organisations; governance safeguards its use; and AI can then support clearly defined, evaluated tasks.

Software selection must therefore be part of a sustainable architecture. Open-source solutions, international proprietary platforms, regional products and national systems address different needs. Their evaluation should consider total cost of ownership, maintenance, training, data exchange, service continuity and local capabilities. Strategy, processes and responsibilities must be clarified before procurement.

Practical Experience Shows Why Context Matters

The experiences presented in the white paper span different stages of digital transformation. In Canada, service continuity and the management of IT environments are central priorities. In Burkina Faso, the feasibility study for Hôpital Saint Camille de Ouagadougou (HOSCO) establishes an architecture and roadmap ahead of deployment. In Togo, the digitalisation of five pilot hospitals connects clinical and administrative functions.

The exploratory mission to South Korea and the European dialogue in Belgium add research and governance perspectives. These initiatives have different objectives and levels of progress: a preliminary assessment, an exploratory mission and an operational deployment do not represent equivalent outcomes.

Where connectivity or electricity is unreliable, systems that can operate offline and synchronise data securely when connections become available are particularly valuable.

AI Must Demonstrate Clinical and Operational Value

Clinical documentation, image analysis, decision support, triage and monitoring of at-risk patients are among the most promising applications. Their evaluation should focus on access to care, time saved, quality and safety. In OntarioMD’s 2024 study, participants using AI scribes reported saving three to four hours of administrative work per week. These are self-reported gains within that study, rather than a guaranteed outcome for every physician. ^[1]

Data representativeness is equally important. A review published in August 2026 examined 38 publicly available cardiac imaging datasets and found that nearly 80% originated in high-income countries. None of the datasets identified came from Africa or South America. This finding applies to that specific collection; it highlights the need for local data and validation tailored to the populations concerned. ^[2]

Ten Leading Hospitals: Learning from Pioneers While Recognising Different Levels of Maturity

The new edition profiles ten hospitals among the world’s most advanced in AI integration in 2026:

• Sheba
• Mayo Clinic
• University Health Network (UHN)
• Charité Berlin
• Renji
• Albert Einstein
• Seoul National University Hospital
• Fortis Memorial Gurugram
• National University Hospital (NUH) Singapore
• JCHO Osaka

The selection reflects documented initiatives, a broad range of applications and geographical diversity. It is not a global ranking based on comparable scores.

Three lessons stand out. In Singapore, NUH reports 115 AI applications in use: scaling requires integration into everyday roles and workflows. ^[4] In Berlin, the Dragon Copilot documentation pilot requires physicians to review the generated notes: time savings must be accompanied by explicit professional oversight. ^[5] Meanwhile, the Clinical Environment Simulator developed by researchers in Seoul and at Harvard enables the consequences of AI decisions to be studied in a virtual hospital. This is a research tool, not a substitute for validation in real-world care. ^[6] These examples underline the importance of distinguishing operational deployments, pilots and announced projects before making comparisons.

Security and Skills: Plan for Both from the Start

Digitalisation increases dependence on information systems and exposure to incidents. The 2026 IBM/Ponemon report estimates the average cost of a healthcare data breach at US$6.64 million across the incidents studied. Access controls, backups, data protection and incident response must be built into projects from the outset. ^[3]

Success also depends on people: initial training, ongoing support, knowledge transfer and change management. Project funding must cover the operational lifecycle, with clear responsibilities and teams capable of maintaining and improving the systems over time.