Healthcare infrastructure and data governance: the silent transformation currently underway

In fact, infrastructure remains central. On the contrary: infrastructure continues to be the main tool through which data governance is actually exercised today. Because infrastructure is not neutral. It defines standards, access methods, interoperability, security and the practical possibilities for using data.

And this is where the transformation takes on a systemic dimension. EDS is, in fact, one of the most significant changes introduced in recent years. Not only because it builds a new technological infrastructure, but because it redefines the way in which the system conceives the relationship between data, institutions and decision-making processes.

For the first time, the National Health Service is moving towards a truly federated approach on a national scale. Information from electronic health records, laboratories, diagnostic services, pharmacies, local health systems, population registers and general practice is gradually being integrated into a common architecture based on shared standards and interoperability. The significance of this transition, however, goes far beyond the technical aspect.

The real innovation is that data is no longer confined to individual hospitals or healthcare organisations. It becomes a link between citizens, organisations, regions and the national level. And this profoundly changes the nature of the interdependencies within the healthcare system. The challenge, therefore, is not simply to ‘have data’, but to build a system capable of circulating, interpreting and using it in a coherent manner.

This is why the issue of federated infrastructure becomes a crucial one. Distributed data governance can only work if there are common standards, shared rules and architectures capable of maintaining consistency amongst different stakeholders. From this perspective, the EDS represents the first concrete attempt to build a federated national platform that is intended to gradually integrate with the European framework defined by the EHDS.

However, the real issue is not merely about collecting or accessing information. The key point is to understand how data generates value. In public debate, there is often an implicit conflation between the availability of data and the system’s capacity for transformation. But this is not the case. The availability of data represents only the first step in a much more complex chain. We can envisage this evolution as a veritable data value ladder.

At the core lies data collection: integration of source systems, standardisation of coding, interoperability and feeding data into shared platforms. This is the level in which the Italian healthcare system is currently investing most heavily. The second level is visualisation. Making data accessible to citizens, professionals and institutions is already a crucial step: summary health profiles, medication records, access to medical reports, monitoring dashboards and the integration of clinical information from different facilities. But data collection and visualisation alone are not enough. The real leap forward occurs when the system is able to analyse the data. It is here that the data takes on meaning – not as isolated information, but as an element which, when combined with other data, enables us to generate insights, identify patterns, make predictions, anticipate risks and support decision-making. Consider the possibility of estimating the risk of readmission, identifying undiagnosed chronic conditions at an early stage, or predicting trends in waiting lists and A&E attendance. Here, the data is no longer merely documentation. It becomes operational knowledge. But the most advanced level is that which triggers action.

This is the point at which the system no longer merely represents or interprets reality, but intervenes directly in processes: it organises discharges, reallocates resources, coordinates hospital beds, updates care pathways, reschedules surgical procedures or triggers automatic notifications. It is precisely at this level that AI and GenAI find their most transformative role. The public debate on artificial intelligence in healthcare often tends to focus on the tools: diagnostic algorithms, chatbots, clinical support and document automation. But the real strategic crux is not the individual technological application. AI and GenAI only work if they are underpinned by a mature, interoperable and governed data ecosystem. Without data quality, without shared standards, without common models and without consistent organisational processes, artificial intelligence risks producing fragile and poorly scalable automation.

International experience clearly illustrates this point. In the United States, Kaiser Permanente uses predictive systems to identify patients at risk of clinical deterioration up to twelve hours before their condition worsens. In the UK, the NHS has developed a Federated Data Platform capable of coordinating waiting lists, operating theatres, beds and discharges via a single platform. But what is striking is not merely the use of advanced technologies. It is the fact that these systems are built around a shared organisational model. In the UK’s case, the real strategic element is the NHS Canonical Data Model: a common definition of the healthcare system’s operational entities – patient, admission, waiting list, operating theatre slot, hospital bed – and the relationships linking them. It is this that enables the platform to manage processes in a coordinated manner. When an operating theatre is cancelled, the system does not merely record an event. It automatically re-prioritises patients according to clinical priority and guaranteed times, proposes new allocations, updates waiting lists, notifies staff and members of the public, and synchronises different systems. The same applies to discharge planning: the system automatically coordinates with families, GPs, transport services, care homes and bed availability.

Here, data is not just for ‘seeing’. It is for taking action. And this is precisely the crux of the matter for the Italia system.

In many respects, Italia is not starting from scratch. The source systems already exist. Healthcare organisations have electronic health records, administrative systems, tools for managing waiting lists and digital patient registers. Language standards are gradually converging and the federated architecture of the EDS has been defined.

But one crucial element is still missing: a common ‘grammar’ for the healthcare system. It is not enough, in fact, for the systems to ‘speak the same language’ from a technical point of view. They must also share the operational meaning of the objects and actions they describe. What exactly does ‘waiting list’ mean? When can a patient be considered ready for discharge? What links exist between the care pathway, local care provision and bed availability? What actions should be triggered automatically when a patient’s status changes? These are not merely IT issues. They are managerial, clinical and policy decisions.

Defining a common ontology means codifying how services operate. It means building a shared representation of the objects, relationships and actions that define how the National Health Service operates. Without this level of shared definition, AI and automation cannot truly scale.

And this is also why the discussion on data governance cannot be confined solely to ICT departments. The transformation currently under way requires, in fact, a profound change in the skills sets within healthcare organisations as well. For years, the issue of digitalisation has been viewed primarily as a technological matter: acquiring software, ensuring interoperability, and managing infrastructure.

But a data-driven system requires much broader skills. IT skills are certainly needed, but they are no longer enough. We need professionals capable of analysing data, interpreting decision-making processes, building operational models and understanding how healthcare services actually work. Anyone building a predictive algorithm to identify patients at risk is not merely carrying out a technical exercise. They are intervening in a clinical and organisational process that has operational, allocative and decision-making consequences. The same applies to platforms for managing waiting lists or discharges: they only work if they incorporate a genuine understanding of healthcare processes. Transforming data into action therefore requires new hybrid skills, capable of linking technology, organisation and the governance of services.

Above all, however, it requires strategic direction. The scale of the transformation currently under way is enormous. Not only because of the magnitude of the potential impacts – reducing inefficiencies, improving the appropriateness of care, enhancing predictive capabilities, personalising services and improving coordination between hospitals and the local community – but also because it changes the very way in which decisions are made.

This is why the issue of data governance cannot be treated as a mere technological project. It is a strategic matter of corporate governance. The third-oldest country in the world cannot afford new platforms without a genuine transformation of its processes. The issue of data and its management could form the basis of a new alliance between institutions, because today the point is not simply to digitise information in the healthcare sector. The point is to understand that data is becoming the key element upon which a multitude of other innovations are built.

* Cergas and SDA Bocconi