Digital health is the set of technologies that digitise data, diagnosis and medical care. It includes electronic health records, AI-assisted diagnostics, telemedicine and wearable devices for continuous monitoring. This year, in particular, is shaping up as one of concrete deadlines rather than announcements. For example, Italy is working towards national interoperability of its electronic health record. Meanwhile, Europe has set its common health-data space in motion. In addition, AI-assisted diagnostics has, for the first time, passed one thousand authorisations in the United States.
Contents
- What digital health is
- Electronic Health Record 2.0: status in Italy
- The European Health Data Space (EHDS)
- AI diagnostics: where it actually works
- Telemedicine: what is reimbursed and what isn’t
- Wearables and continuous monitoring
- Digital mental health: apps and their limits
- The AI Act and high-risk medical devices
- Privacy and cybersecurity of health data
- FAQ and healthcare in 2035
What digital health is (definition)
Digital health applies information technology, artificial intelligence and connected devices to four areas of medicine. These are managing clinical data, diagnosis, remote delivery of care, and continuous monitoring of health status. However, it does not replace the doctor-patient relationship. Instead, it changes its informational support. In other words, data becomes more complete and more shared, and increasingly it is also analysed by AI systems as a second opinion.
Electronic Health Record 2.0: status in Italy
The Electronic Health Record (Fascicolo Sanitario Elettronico, FSE) is the digital container that gathers every citizen’s clinical history. It holds reports, discharge letters, prescriptions, and the new Synthetic Health Profile compiled by the treating physician. It has existed since 2012, but the original version was deeply inconsistent across Italy’s regions. As a result, a patient moving from one region to another often had to start from scratch.
Here is the status verified as of August 2026. The implementing decree set 31 March 2026 as the deadline for stricter rules to take effect. For the first time, these rules extend the obligation to feed the record to outpatient clinics, dental practices and private laboratories, not just public facilities. In practice, this means documents must be uploaded within 5 days of the service being provided. The stated target for full national interoperability across all regions is June 2026. Even so, several industry sources flag a real risk of further regional slippage. For instance, some regions, such as Tuscany, have already obtained local extensions on specific technical points. Meanwhile, citizens retain control over their data through a masking right, since they can hide individual documents from specific professionals. In addition, they can check at any time who has accessed their record and for what purpose.
The European Health Data Space (EHDS)
EHDS, the European Health Data Space, is the EU regulation that builds a common framework for sharing health data between member states. It has been in force since 26 March 2025. Specifically, it covers both direct-care purposes, known as primary use, and research and health-policy purposes, known as secondary use.
| Deadline | What happens |
|---|---|
| 26 March 2027 | General application of the regulation; core obligations for electronic health record providers |
| 26 March 2029 | Exchange of the first priority data categories (synthetic health profile, e-prescriptions) between all member states, and secondary-use rules start applying to most data |
| 26 March 2031 | Extension to the exchange of medical imaging, lab results and hospital discharge letters |
For Italy, the natural link is with the FSE 2.0 infrastructure. This infrastructure will need to become compatible with the European interoperability standards, such as the FHIR format, required by EHDS.
AI diagnostics: where it actually works
In the United States, the FDA passed 1,300 cumulative authorisations for AI-based medical devices in 2026. Notably, radiology alone accounts for around 76% of all authorisations. In 2025 alone, a record year, 258 were granted. A representative case is Aidoc, which secured authorisation for an AI triage solution. This solution combines eleven new clinical indications with three already authorised. In its registrational studies, it reached an average sensitivity of up to 98.5% and a specificity of up to 99.7%.
However, there is a flip side, documented in an analysis published in npj Digital Medicine: transparency. Almost half of FDA summaries for authorised devices do not describe the clinical study design used for approval. Moreover, more than half omit sample size. As a result, it becomes hard for doctors and hospitals to genuinely assess how reliable a given tool will be in their own clinical context. The right message for readers, then, is this: AI diagnostics works well on specific, narrow tasks, such as spotting a fracture or flagging a suspicious nodule. It is not, however, a general substitute for medical judgement.
Telemedicine: what is reimbursed and what isn’t
Telemedicine covers remote visits, remote monitoring of clinical parameters, and teleconsultations between specialists. In Italy, though, reimbursement by the National Health Service remains uneven across regions and services. For example, some follow-up teleconsultations are treated the same as an ordinary outpatient visit for co-payment purposes. By contrast, other remote-monitoring services remain tied to specific regional projects, often funded through the PNRR recovery plan, with no single national tariff. On a positive note, FSE 2.0 now includes dedicated functionality for telemedicine services. This is a sign of the direction the system is converging on.
Wearables and continuous monitoring
Smartwatches and smart rings today offer certified clinical functions in some markets. Atrial fibrillation detection is one example. Alongside these sit generic wellness metrics, such as sleep, stress and recovery scores, that do not carry the same level of medical validation. The practical distinction for readers is this: a function carrying medical-device certification has passed a verifiable regulatory pathway. A “readiness” or “stress” score generated by a proprietary algorithm, however, has not. Therefore, it should be treated as a wellness indicator, not clinical data.
Digital mental health: apps and their limits
Mental health support apps, often built around conversational chatbots, have seen rapid growth in downloads and use. Still, the scientific community remains divided on their effectiveness as a standalone tool. Generally speaking, most available studies place them as a complement to a therapy path led by a professional. In other words, they are not a substitute for therapy in cases of significant psychological distress.
The AI Act and high-risk medical devices
The EU’s AI Act, in force since August 2024, classifies AI-based medical devices as “high-risk” systems. Consequently, it introduces reinforced obligations around risk management, training-data quality and human oversight. Most obligations for high-risk AI take effect from August 2026. Full compliance for AI medical devices, however, is required by August 2027. As a result, a manufacturer placing an AI-based diagnostic tool on the EU market today already has to navigate a double regulatory track: the AI Act plus the medical device regulation (MDR).
Privacy and cybersecurity of health data
Health data remains among the most sensitive categories of personal data. It is also among the most targeted by cyberattacks. This is precisely because of its black-market value and the operational impact a systems shutdown has on a hospital. EHDS imposes security and access-tracking requirements for electronic health record systems. Even so, the operational responsibility for implementing them still rests with each individual healthcare facility. In practice, this means very different maturity levels between large hospitals and small private practices.
FAQ
Is the Electronic Health Record already complete across Italy? Not entirely yet. The stated target for full national interoperability is June 2026, but some regions have obtained local extensions on specific technical points.
What is EHDS and does it affect ordinary citizens? It’s the EU regulation on health data. Mainly, it affects citizens through rights of access and control over their own data, with the first concrete cross-border exchange features expected by March 2029.
Can an AI replace a doctor’s diagnosis? No. Even the most widely authorised AI devices, over 1,300 in the US, are validated for specific, narrow tasks, not as a general substitute for clinical judgement. In every case, they remain supervised by a professional.
Can my smartwatch really detect a heart problem? Some specific functions, such as atrial fibrillation detection, have recognised clinical validation in several markets. Other metrics, like stress or recovery, remain wellness indicators rather than diagnostic data.
Do mental health apps replace therapy? In most available studies, no. Instead, they are generally positioned as a complement to a therapy path with a professional, not a substitute in cases of significant distress.
