Digital health technologies are reshaping cardiovascular disease (CVD) prevention, diagnosis, and management. Tools such as electronic health records (EHRs), clinical decision support systems, telemedicine, mobile health apps, and wearable devices support risk identification, real-time monitoring, and tailored interventions. These innovations promise improved care co‑ordination and reduced avoidable morbidity, yet evidence of sustained impact on long-term outcomes remains limited. EHRs and decision support systems can close treatment gaps, but interoperability and governance challenges persist. Telemedicine has expanded access, particularly for stroke and chronic CVD care, and remote monitoring and consumer wearables offer opportunities for self-management – though uptake is uneven and integration into health systems remains limited. Artificial intelligence and big data analytics can further enhance diagnosis and risk prediction but also present concerns related to bias, transparency, and privacy. Realising the full potential of digital health requires robust infrastructure, inclusive design, and harmonised data governance.
The State of Cardiovascular Health in the European Union
Abstract
Digital health technologies are transforming the way cardiovascular disease (CVD) is prevented, diagnosed, and managed. Tools such as electronic health records (EHRs), clinical decision support systems (CDSS), telemedicine platforms, mobile health apps, and wearable monitoring devices are enabling more proactive, co‑ordinated, and patient-centred care. These technologies can help identify individuals at high risk, monitor chronic conditions in real time, and deliver tailored interventions – potentially reducing avoidable morbidity and mortality. However, while process improvements are well documented, evidence of sustained improvements in long-term cardiovascular outcomes remains limited and context dependent.
EHR systems now underpin much of the information infrastructure in high-income countries, allowing for better care co‑ordination, patient tracking, and data-driven insights. When linked with CDSS, they can prompt evidence‑based decisions, close treatment gaps, and reduce variation in care. Yet, challenges around usability, interoperability across local and organisational EHR platforms, and trust in data governance continue to limit their full potential. Results from the 2025 OECD Cardiovascular Policy and Data Survey suggest that countries may not be positioned to leverage EHR data for prevention, early detection and management of CVD.
Telemedicine has emerged as a core component of cardiovascular care, especially since the COVID‑19 pandemic. Remote consultations, virtual rehabilitation, and telemonitoring have been associated with better access to services and improved outcomes in some conditions, provided they are accompanied by clinical follow-up and intervention. In terms of acute intervention, six of 18 countries report having in place telemedicine facilities that connect stroke patients in remote regions with specialist teams, while seven countries report regulatory frameworks for teleconsultations are either “fully implemented” or “in development”. Ten countries report that ambulances have telemedicine capabilities that connect paramedics with physicians and/or specialist teams.
Remote monitoring devices and implantable sensors enable earlier detection of deterioration, and studies show they can reduce hospitalisations and enhance treatment adherence when combined with structured care programmes. Consumer-facing technologies, including smartwatches, fitness trackers, and mobile health apps, offer new opportunities for self-management and early detection. These tools can help patients track activity, monitor heart rhythms, and receive coaching or medication reminders. However, their use remains skewed towards younger, more affluent populations, with limited integration into mainstream health services. Questions remain over their reliability as well as regulating their integration into mainstream services. Currently, few countries have policies or infrastructure in place to leverage wearable and mobile technologies for national-level CVD monitoring or prevention. Bridging these gaps – through inclusive design, policy development, and better data systems – will be essential to realising the full promise of digital innovation in cardiovascular health.
Machine learning, artificial intelligence (AI), foundational models and big data analytics play a key role in powering new monitoring tools. They are reshaping CVD management by enhancing diagnostics, risk prediction, and treatment personalisation. AI is showing promise in interpreting medical imaging ECGs, detecting conditions such as asymptomatic left ventricular dysfunction from routine ECGs or coronary artery calcium from unrelated computed tomography (CT) scans. With more robust trials and evaluation, these tools may enhance the ability to uncover hidden cardiovascular risks and guide early intervention. AI is also being embedded into wearables and mobile health apps, allowing continuous, real-time analysis of large data volumes from devices. Machine learning models are already helping clinicians predict patient outcomes, optimise treatment plans, and support clinical decision making. With over 1 000 FDA-approved tools in 2024 – many in cardiology – these AI-based technologies are increasingly used in practice. Meanwhile, big data integration across EHRs, registries, wearables, and genetic databases is powering large‑scale research and CVD surveillance. The EU has introduced an Act to help regulate this burgeoning industry – the first jurisdiction in the world to do so.
Despite the promise, implementation of AI and big data faces challenges including data bias, algorithm transparency, data privacy, and automation bias. Ensuring equitable access and building clinician trust are essential for responsible adoption. Trust can only be built through high data quality, adequate training and supervision, robust clinical evidence and transparency. While these technologies can enable a more proactive and personalised approach to cardiovascular care, their implementation requires careful consideration of evidence, ethics, and inclusivity.
Leveraging routinely collected health data is key to maximising the potential of digital technologies and advancing next-generation biomedical innovations in CVD. EHRs, with their longitudinal patient histories, support both clinical care and research (although most OECD countries are yet to fully leverage their potential). Disease‑specific registries enable benchmarking, real-world evaluation of care, and inclusion of underrepresented patient groups. Administrative datasets offer system-level insights into care delivery and costs, while mortality databases enable tracking trends and outcomes. These individual sources are powerful, but their true potential is realised when data infrastructure can be linked and made interoperable to provide a continuous, granular view of the patient journey, support better prevention strategies, and allow benchmarking across systems and borders.
Interoperability allows for richer analytics – such as calculating survival rates or identifying high-risk patients before adverse events occur. Cross-border data integration in Europe is also key, as it enables pooled analyses, enhances statistical power, and facilitates benchmarking. Initiatives like the European Society of Cardiology’s (ESC) EuroHeart aim to standardise cardiovascular data across countries to improve policy and practice. However, data silos, incompatible standards, lack of unique patient identifiers, and varying privacy regulations remain substantial barriers. Overcoming these challenges is critical to realising the full value of electronic data for CVD prevention and control.
Achieving health data interoperability requires co‑ordinated action on several fronts. Technical barriers include differing data formats and the absence of a shared patient ID across Europe. Legal and ethical hurdles stem from varying interpretations of General Data Protection Regulation (GDPR), inconsistent consent rules, and public trust concerns. Technical and governance structures like the European Health Data Space (EHDS) offer promise by creating national competent authorities facilitating data access to health data (Health Data Access Bodies), bodies, common standards, and by bringing legal clarity. Strong governance involves patient representation, transparent policies, and robust measures for data quality and security. In addition, investment in infrastructure and workforce development is essential. Skilled professionals, from informaticians to clinicians trained in data literacy, are essential to fully exploit digital health tools. Incentives, mandates can ensure comprehensive and high-quality data collection. A fully interoperable system can serve as the backbone of learning health systems across Europe.
Many cases of CVD are linked to modifiable risk factors and could be prevented through better early intervention, prompting health systems and policymakers to leverage digital innovations to strengthen prevention and management efforts. Recent advances in information and communication technologies (ICT) offer novel ways to prevent disease, detect problems sooner, and manage chronic cardiac conditions more effectively. Digital health technologies – encompassing electronic records, connected devices, and patient-facing applications – are increasingly used to prevent, detect and manage CVD. These technologies range from tools used primarily by healthcare professionals (such as electronic health records (EHRs), clinical decision support systems, and telemedicine platforms) to consumer-facing devices and apps like wearable fitness trackers, home monitoring devices, Holter monitors, connected wearables and mobile health (mHealth). Each category of technology can play a distinct role in CVD management, and evidence is rapidly emerging on their uptake, benefits, and limitations.
Prevention is key, and digital technology can facilitate moving care upstream to fundamentally modify cardiovascular risk factors such as diet, obesity, inactivity and chronic stress. In Europe, 11.4 million deaths from CVD were deemed avoidable between 1995 and 2020 – resulting in a potentially avoidable loss of 213.1 life years (Avi Cherla, 2024[1]). As much as 80% of the current stroke burden is linked to 10 modifiable risk factors including high blood pressure (BP), smoking, cholesterol, and salt, alcohol and sugar intake, as well as environmental factors such as air pollution, see more on risk factors for CVD in Chapter 3 (Feigin et al., 2025[2]). Similar prevention initiatives exist across EU Member States, supported by the European Society of Cardiology (ESC) and national digital health programmes – for example, the ESC’s EuroHeart registry, which as of 2024 covers 14 countries and more than 60 000 patients.
Primary preventative interventions include encouraging and supporting lifestyle interventions, building health literacy, and promoting screening and management of associated health conditions. Secondary prevention is also critical with 25% of stroke victims likely to have another stroke within five years (Mohan et al., 2011[3]). Most interventions that have been evaluated in the academic literature have focussed on dietary habits; adherence; medication use; waist circumference and quality of life (Qi et al., 2025[4]). Evidence of the long-term impact of these interventions on key cardiovascular outcomes is still relatively undeveloped. Plus, given that 20% to 40% of heart attacks occur in patients previously undiagnosed with CVD, monitoring could prevent much morbidity and mortality (McClellan et al., 2019[5]). Digital technology can be useful to monitor people routinely and generate alerts when action is needed, as, for example, proposed under the EU CVD Health Check (Section 5.1.5).
5.1.1. Clinical Decision Support Systems (CDSS) improve clinical processes but evidence for improved outcomes remains elusive
Clinical decision support systems are computerised tools that assist clinicians by integrating patient data with evidence‑based guidelines or algorithms to inform decision making at the point of care. CDSSs targeting CVD prevention include: assessing a patient’s risk for developing CVD based on their history, risk factors, and clinical test results; recommending health behaviour changes, such as quitting smoking, increasing physical activity, and reducing excessive salt intake; tailored reminders to screen for risk factors and preventive care, clinical tests, and treatments; alerts when indicators for CVD risk factors are suboptimal; and evidence‑based treatments to prevent CVD, including intensification of existing management (Njie et al., 2015[6]). CDSS include both rule‑based tools that apply predefined guidelines and emerging AI-enabled systems that learn from clinical data to provide predictive or personalised recommendations.
In cardiovascular care, CDSS often take the form of risk calculators, alert systems, or care pathway prompts embedded in electronic health record interfaces (Qi et al., 2025[4]). For example, a CDSS might alert a primary care provider that a patient’s blood pressure is uncontrolled and suggest guideline‑recommended therapy adjustments, or it might calculate a patient’s ten‑year risk of heart attack based on their profile and prompt a statin prescription if appropriate. By providing tailored recommendations based on individual data, CDSS aim to enhance clinician oversight and reduce “clinical inertia” in managing CVD risk factors (Njie et al., 2015[6]). The ESC’s Chatbot ESC Chat platform offers instant and personalised answers to clinical queries. Answers are exclusively based on current ESC Clinical Practice Guidelines (ESC, 2025[7]). Decision support tools for acute ischaemic stroke are also proliferating, although the quality of the evidence is variable and challenges to clinical translation are reported (Ela Marie Z. Akay, 2023[8]).
Virtual, computational representations of individual patients built from real-world clinical, physiological, and genetic data also hold potential for improving CVD prevention and care. Such virtual human digital twins integrate information from imaging, EHRs, wearable data profiles to simulate the dynamic behaviour of the cardiovascular system under different conditions or interventions. This enables personalised risk prediction, early detection of disease trajectories, and the capacity to test therapeutic strategies in silico before applying them in clinical practice (Viceconti, 2016[9]). In CVD care, digital twins could continuously update as new data are collected, allowing clinicians to model how a patient’s heart might respond to medication changes, revascularisation, or lifestyle modification, thereby supporting truly adaptive and preventive medicine. Beyond individual care, population-level digital twins could help identify vulnerable groups, optimise resource allocation, and accelerate the evaluation of new treatments. Digital twin frameworks exemplify the convergence of AI, systems biology, and precision medicine. Despite challenges and limitations (data quality, model complexity), they offer a pathway toward continuously learning, data-driven cardiovascular care (Corral-Acero et al., 2020[10]).
Over the past decade, CDSS capabilities have been increasingly incorporated into care – often as part of broader health IT systems – especially in hospitals and large practices. Many OECD health systems now use CDSS for preventive care reminders (e.g. prompting cholesterol or diabetes screening) or for acute care decision support (e.g. alerts for possible drug interactions in cardiac medications). The adoption of CDSS parallels the adoption of EHRs, with which they are frequently integrated, which expanded dramatically after 2010 due to policy incentives (in the United States, for instance, CDSS use rose with near‑90% EHR adoption in hospitals (Qi et al., 2025[4]). Precise utilisation rates vary by setting and country, but CDSS are commonly available in modern cardiology practice environments – from specialised algorithms in catheterisation labs to primary care tools for cardiovascular risk assessment (Njie et al., 2015[6]). Machine‑learning models are increasingly explored to support cardiovascular decision making, with early studies suggesting potential efficiency gains; however, systematic evidence on clinical outcomes remains limited. With over 1 000 FDA-approved AI-based tools by 2024 – many in cardiology – these technologies are increasingly used in practice, promising faster workflows and more targeted care (Grant et al., 2024[11]; Singh et al., 2025[12]).1
Evidence suggests CDSS can modestly improve the quality of CVD care processes. A systematic review by the U.S. Community Preventive Services Task Force found that CDSS implementations led to significant increases in clinicians’ performance of recommended CVD preventive services, risk factor screenings, and treatments (median improvements on the order of 2‑4 percentage points in adherence to recommended care) (Njie et al., 2015[6]). For example, the presence of a CDSS might increase the likelihood that a patient due for a cholesterol test actually gets one, or improve rates of appropriate antihypertensive prescribing by a few percentage points. Across large populations, such incremental gains can be meaningful. However, the impact on patient outcomes (like actual reductions in blood pressure, cholesterol levels, or CVD events) was inconsistent with some studies finding evidence of improvement and others failing to. Overall, reviewers concluded that CDSSs can improve process measures of care, but more evidence is needed to demonstrate clear reductions in CVD morbidity and mortality attributable to these decision support tools (Njie et al., 2015[6]).
Individual studies suggest promise in specific contexts: for instance, a CDSS in community clinics significantly improved control of cardiovascular risk factors in socio‑economically vulnerable patients (Gold et al., 2022[13]). In the management of heart failure (HF), for example, integrating CDSS into care has been associated with better adherence to guideline‑directed therapies and even improved clinical outcomes in certain cases. A review noted that while broad EHR adoption alone did not automatically HF outcomes, targeted EHR-based decision tools (a form of CDSS) did help increase guideline‑concordant care and were linked to better patient self-management and clinical results (Kao, 2022[14]).
Based on the available evidence, CDSSs can help to standardise care and reduce gaps between evidence and practice. They can remind busy clinicians of best practice, potentially preventing errors of oversight. By processing large amounts of clinical data, advanced decision support (increasingly augmented by AI – see below) can also identify non-obvious patterns – for example, flagging a combination of subtle ECG changes and lab results that suggest HF risk.
Nevertheless, limitations exist. Early-generation CDSS often produced alert fatigue (too many pop-up alerts leading providers to override or ignore them). Integration challenges and poor usability have sometimes hindered uptake, with resistance to tools that disrupt workflow. Moreover, if underlying clinical guidelines are outdated or not applicable to a particular patient – say an elderly multimorbid patient not fitting trial populations – CDSS advice can be less useful or even misleading. AI-ready guidelines are thus urgently needed to ensure correct evidence‑to-CDSS mapping. This requires digital libraries containing phenotype definition to ensure harmonisations of definitions and alignment across guidelines. Furthermore, guidelines adapted to local circumstances (e.g. resources, language, culture) are needed before wide‑scale implementation.
Gaps also exist in evidence on long-term outcomes. Many studies suggest improved care processes but evidence on meaningful improvements in patient outcomes (e.g. fewer heart attacks or strokes) remain elusive (Kao, 2022[14]). Good implementation and continuous evaluation and updating of CDSSs are pivotal. When implemented well, a CDSS can effectively prompt better CVD preventive care, but they work best as part of a comprehensive strategy including provider education and patient engagement (Kao, 2022[14]).
5.1.2. EHRs and related data systems enable prevention, early detection and management
EHRs are primarily clinical information systems used to document and exchange patient data. On this foundation, additional functionalities – such as dashboards, decision-support modules, and analytics – can support prevention and management of CVD (Slawomirski et al., 2023[15]; Kao, 2022[14]). The past decade has seen near-universal EHR adoption in some high-income countries. Countries including Denmark and Estonia have implemented comprehensive, national EHR systems enabling seamless data flow across care settings. The widespread use of EHRs means that most patients with CVD have their histories, medications, and investigations recorded electronically, which can theoretically support co‑ordinated prevention and management of CVD.
EHR platforms also enable large‑scale data aggregation – facilitating national CVD registries and research networks. For instance, the European Heart Rhythm Association’s Atrial Fibrillation (AF) Ablation Pilot Study leveraged a multi-country registry to collect real-world data on 1 410 patients undergoing AF ablation across 72 centres, yielding insights into outcomes and practice variation (European Heart Rhythm Association, 2014[16]). Of course, such initiatives that access and link data for direct and indirect purposes are predicated on good EHR infrastructure (OECD, 2019[17]).
The impact of EHRs on cardiovascular care can take several forms. Firstly, access to EHR by relevant providers improves information continuity – all members of a patient’s care team (cardiologists, primary care physicians, emergency department staff) can view a patient’s history, medications, allergies, and test results. This is expected to reduce errors such as duplicate medications or overlooked contraindications and supports safer care transitions for CVD patients. For example, when a HF patient is discharged from hospital, their outpatient provider can see the discharge summary without delay and adjust treatment accordingly.
EHRs also enable population management: using EHR data, clinics can identify patients overdue for cholesterol checks or vaccinations, supporting preventive care outreach. Some evidence suggests that specific EHR-based interventions can improve care quality. A review noted that employing EHR tools – like automated reminders for physicians about therapy optimisations or patient portals for self-monitoring weight and symptoms – was associated with better adherence to HF guidelines and even improved patient outcomes in some instances (Kao, 2022[14]). Similarly, EHR data analytics have been used to flag care gaps, such as finding coronary artery disease patients not on statins, and prompting targeted improvements (Kao, 2022[14]).
The rich data in EHRs are increasingly being mined to develop predictive models – for example, algorithms that predict which hypertensive patients are at highest risk of a cardiovascular event within five years, so that clinicians can intervene proactively (Grant et al., 2024[11]). Research leveraging EHR databases has led to discoveries in CVD trends, treatment effectiveness, and identification of at-risk cohorts (Dhingra et al., 2023[18]). The FIND-AF algorithm, developed using machine learning models, which scans primary care medical records for red flags that a patient might develop atrial fibrillation, which, when undiagnosed can lead to stroke, is being piloted (Nadarajah et al., 2023[19]). Such initiatives depend on robust governance frameworks addressing privacy, public engagement and cybersecurity.
Despite the obvious potential, broad studies on EHR adoption have revealed modest direct impact in outcomes. Early analyses comparing hospitals with and without EHRs found mixed results in CVD care quality. One study noted that EHR adoption was not by itself associated with major improvements in HF outcomes or 30‑day readmission rates (Kao, 2022[14]). A 2023 survey of OECD countries found persistent fragmentation of EHRs, with only 15 of the 27 responding countries reporting a single, unified EHR covering the entire healthcare system (Slawomirski et al., 2023[15]).
Better information does not translate to improved outcomes organically or automatically. Much depends on how effectively the EHR is used. Clinician users often report that current EHR interfaces are cumbersome, consuming time that could be spent with patients. Usability issues and alert fatigue can hinder the effective use of EHR features. Interoperability is another challenge: if data cannot easily flow between different healthcare institutions, the continuity benefits are undercut. Concerns about data privacy and security remain salient. The OECD survey of 2023 found ongoing challenges in the governance of harnessing EHR data for analytics and research. Many of these challenges have been reported since 2012. They include legal barriers, lack of resourcing, and ongoing resistance from providers. A lack of social consensus, license and trust is a key barrier to using EHR data for these purposes (Slawomirski et al., 2023[15]). This is also explored further in Section 5.3.
The potential of EHRs in preventing and managing CVD is nevertheless considerable. When optimised, they standardise best practices (partly through embedded CDSS alerts discussed above), track outcomes, and enable large-scale quality improvement initiatives. For example, integrated EHR data allowed the American Heart Association’s Get With The Guidelines program to monitor and improve adherence to acute myocardial infarction treatment metrics nationwide (Njie et al., 2015[6]). As such, EHRs can form the backbone of digital cardiology if universally adopted and used for direct patient care as well as research, and a consensus exists that they are a necessary foundation for improving CVD management. To truly realise benefits, ongoing efforts to enhance their usability, interoperability, and incorporation into clinical workflow are needed. When coupled with other digital tools ‑ and applied in a patient-centred way ‑ EHR systems can meaningfully contribute to better prevention and management.
Description: In 2011, the City of Oulu, Finland, scaled-up its digital patient-provider portal – the Self Care Service (SCS) – to all residents. SCS offers patients a range of online primary care services such as online appointment booking, messaging with professionals and ePrescriptions. SCS is tethered to each individual’s electronic health record (EHR) to ensure health professionals have ready access to patient data. For health professionals, SCS provider guidelines and care pathways based on individual patient data. SCS is voluntary and free-of-charge.
Best practice assessment:
Enhancement options: To enhance effectiveness, policy makers can continue efforts to boost population health literacy to ensure patients understand the information they receive and therefore appreciate the usefulness of SCS. To enhance equity, plans to expand the number of languages available on SCS can be prioritised given languages such as Dari and Somali are spoken by refuges who typically have worse health outcomes. To enhance the evidence-base, researchers can capitalise on the high-quality data stored within Finland’s national EHR system by evaluating the impact of SCS on outcomes and healthcare utilisation. Several options to enhance the extent of coverage are available such as encouraging health professionals to promote SCS to patients.
Transferability: SCS has been transferred from Oulu to other Finnish municipalities. It has not been transferred to other countries, however, many OECD and EU countries allow patients to access their EHR via a patient portal (or have plans to). Results from the transferability assessment using publicly available data revealed SCS is most suited to other Nordic countries, which have digitally advanced healthcare systems.
Conclusion: SCS is a patient-provider portal offering residents of Oulu access to a wide-range of primary care services online. SCS is a global leader in this area, however, further enhancements are possible, as outlined in this case study. A high-level transferability assessment revealed Nordic countries are most suited to SCS, nevertheless, there is political interest among a number of countries to improve patient access to their data.
Source: OECD (2023[20]), Integrating Care to Prevent and Manage Chronic Diseases: Best Practices in Public Health, https://doi.org/10.1787/9acc1b1d-en.
5.1.3. Telemedicine and telehealth in cardiovascular care has been shown to improve access and outcomes
Telemedicine refers to the remote provision of clinical care through digital communication technologies, while telehealth encompasses a broader range of remote health-related services including rehabilitation, follow-up and patient education. Telemedicine has moved from a niche offering to a mainstream component of cardiovascular care in recent years. Modalities include video or phone consultations, remote cardiac monitoring, and virtual clinics and virtual wards for chronic disease management. The COVID-19 pandemic in 2020 dramatically accelerated telemedicine adoption across OECD health systems, as regulatory barriers were lowered and both patients and providers became more familiar with remote consultations (OECD, 2023[21]; Eze, Mateus and Cravo Oliveira Hashiguchi, 2020[22]). In many countries, cardiology services rapidly shifted to telehealth for routine check-ups, follow-ups, and rehab sessions when in-person visits were limited. Even post-pandemic, telemedicine remains a key part of care delivery for CVD, valued for its convenience and ability to reach patients who otherwise face access barriers.
Pre-2020, telemedicine use in CVD care was “not used at scale” in most OECD Member countries (Eze, Mateus and Cravo Oliveira Hashiguchi, 2020[22]), though pockets of innovation existed such as telestroke networks in some regions, or remote HF monitoring programmes. The pandemic forced countries to adapt, leading to the volume of teleconsultations surging by several-fold within months. In 2020, in-person cardiology visits dropped significantly in many countries, replaced by telephone or video visits. OECD analysis found in-person consults per capita fell by up to 30%, reflecting substitution by remote care (OECD, 2023[21]). By 2021, telehealth stabilised but remained far above pre‑pandemic levels; in the United States, for instance, telehealth accounted for around 4‑5% of all outpatient claim submissions in late 2021 (up from virtually 0% before) (Edmiston and AlZuBi, 2022[23]), and it is now common for cardiology practices to offer a mix of in-person and telemedicine appointments.
Remote consultations are especially used for follow-up (e.g. reviewing blood pressure or cholesterol lab results), cardiac rehabilitation sessions, and managing stable HF patients who require frequent check-ins. Telemedicine’s uptake is likely a function of high patient satisfaction and acceptance. Surveys show most patients find virtual cardiac care acceptable and would use it again (Eze, Mateus and Cravo Oliveira Hashiguchi, 2020[22]). An example from Ireland is the Virtual Atrial Fibrillation Care at Home Pathway, where – following an assessment in the Emergency Department – eligible patients are provided with a home blood pressure monitor and access to Fibricheck (a certified smartphone App that monitors heart rate and rhythm) and are then managed at home (HSE, 2025[24]).
A growing evidence base suggests that telemedicine can enhance outcomes in addition to traditional face‑to-face care in cardiology. A 2024 systematic review found that it was associated with significant improvement in patient access, satisfaction and patient outcomes in chronic disease management, especially diabetes (Ezeamii et al., 2024[25]), while an umbrella review of 98 systematic reviews concluded that 83% of reviews found telemedicine “at least as effective as” in-person care for clinical outcomes across various conditions (Eze, Mateus and Cravo Oliveira Hashiguchi, 2020[22]). In the CVD context, studies and trials have demonstrated telehealth’s effectiveness. For example, multiple randomised trials of telemedicine‑based cardiac rehabilitation (where patients do rehab exercises and counselling at home with remote supervision) have shown equivalent or better improvements in exercise capacity and risk factor control compared to conventional face‑to-face rehabilitation (Gallegos-Rejas et al., 2024[26]). This has considerable import, since tele‑rehabilitation expands access for patients who face geographic or mobility barriers.
Another area is chronic HF management: remote monitoring of patients’ weight, blood pressure, and symptoms combined with scheduled teleconsultations has led to reduced hospitalisations. A 2022 meta‑analysis in found that in HF patients, a telemedicine intervention consisting of combined remote monitoring plus consultation was associated with a 29% reduction in cardiovascular-related hospitalisations and a significant reduction in cardiovascular mortality (risk ratio ~0.83) compared to usual care (Kuan et al., 2022[27]). Notably, the benefits were observed when telemedicine was comprehensive (monitoring + active management). Remote consultations alone without data monitoring did not significantly change outcomes (Kuan et al., 2022[27]). Telehealth has also shown to be effective in hypertension management – patients receiving remote coaching and telemonitoring achieved slightly better blood pressure control than those with usual office care – and in secondary prevention. One study reported small but significant improvements in systolic blood pressure through telemedicine with patients recovering from myocardial infarction (Kuan et al., 2022[27]).
Overall, evidence suggests that telemedicine can be a useful addition to face‑to-face management of cardiovascular conditions, while often improving patient convenience and adherence. In fact, the OECD review concluded telemedicine interventions have demonstrated concrete benefits such as reducing chronic HF mortality and improving patients’ physical activity levels (Eze, Mateus and Cravo Oliveira Hashiguchi, 2020[22]).
The key advantages of telemedicine in cardiology include improved access (patients in remote or underserved areas can consult specialists without travel), timeliness (early follow-ups can be done quickly via video), and potentially cost savings for both patient and system, if hospital admissions are avoided (Eze, Mateus and Cravo Oliveira Hashiguchi, 2020[22]). Telehealth can also promote multidisciplinary care. For example, a virtual heart team conference can bring together cardiologists, dietitians, and nurses to counsel a patient with coronary disease.
However, there are limitations. Not all evaluations or treatments can be done remotely. Certain visits, like initial diagnostic workups or procedures (stress tests, echocardiograms), still require in-person attendance. Another challenge is technology access and literacy. Older CVD patients may struggle with video platforms or may not have high-speed internet, which can exacerbate disparities. Digital divide concerns have arisen. A study found older and lower-income cardiac patients were less likely to engage in telehealth, requiring strategies to improve usability and access (Dhingra et al., 2023[28]). Reimbursement and regulatory frameworks are still catching up – not all countries reimburse telecardiology at parity with in-person care, and cross-jurisdiction care (important for rural areas near borders) can face licensing hurdles.
Despite these issues, the use of telemedicine is growing in cardiovascular care delivery. The general consensus from outcomes studies is encouraging: telehealth is effective and safe for managing many CVD patients, with particular success in chronic disease management (e.g. HF) and preventive coaching, as long as it is used in appropriate contexts (Eze, Mateus and Cravo Oliveira Hashiguchi, 2020[22]; Kuan et al., 2022[27]). The focus now is on refining hybrid care models that optimally mix telemedicine with hands-on care, and ensuring no patient populations are left behind as care increasingly goes digital.
The survey results from the 2025 OECD Cardiovascular Policy and Data Survey paint a different picture, however. Only six countries (Croatia, Estonia, Germany, Ireland, Norway and Slovenia) reported having in place telemedicine facilities that connect stroke patients in remote regions with specialist stroke teams. Eight (Estonia, France, Germany, Ireland, Latvia, Norway, Slovenia and the United Kingdom) reported that regulatory frameworks for teleconsultations were either “fully implemented” or “in development”. Meanwhile nine countries (Canada, Czechia, Estonia, Finland, Germany, Iceland, Ireland, Norway and the United Kingdom) reported that ambulances have telemedicine capabilities that connect paramedics with physicians and/or specialist stroke teams.
5.1.4. Remote monitoring and home‑based cardiac devices can enhance care quality, but managing the volume of data is essential
Remote patient monitoring (RPM) refers to the use of connected medical or consumer devices to collect and transmit cardiovascular parameters – such as blood pressure, weight, or heart rhythm – from patients’ homes to healthcare providers. This approach enables continuous monitoring outside traditional clinical settings, supporting early detection of deterioration and proactive management of chronic conditions. In high-income countries, RPM programmes have expanded for hypertension, heart failure, and arrhythmia detection, using technologies that range from regulated medical-grade devices (e.g. Bluetooth-enabled blood pressure monitors, ECG sensors) to consumer-grade wearables that provide complementary data but may vary in accuracy. By integrating both types of devices, health systems can enhance patient engagement and improve care co‑ordination.
Hypertension is an example where home monitoring has made an impact. Patients measure their blood pressure at home using digital sphygmomanometers, and readings can be automatically sent via smartphone apps or telehealth platforms to their clinic. This not only provides a more accurate picture (since home readings avoid “white coat” effects) but also allows providers to adjust medications between office visits and has been shown to overcome therapeutic inertia and yield meaningful improvements in BP levels. A meta‑analysis of 46 trials (covering over 13 000 patients) found that home blood pressure telemonitoring (HBPT) significantly improved blood pressure control compared to usual care, lowering systolic BP by about 4 mmHg on average (Duan et al., 2017[29]). Moreover, patients doing HBPT were more likely to achieve blood pressure targets than those with office measurements alone. The effect was even greater when home monitoring was combined with proactive support such as counselling, education, or medication management from care teams (Duan et al., 2017[29]). While simply giving patients a BP cuff helps, giving them a BP cuff plus active telehealth follow-up helps more.
Similar home monitoring approaches exist for other risk factors: diabetics use glucometers feeding data to digital logs; patients trying to lose weight use Wi-Fi scales; those with hyperlipidaemia might use digital pill bottles or apps to ensure medication adherence. While these devices by themselves are not treatments, they facilitate behaviour change and early intervention. If a hypertensive patient’s readings creep up, for example, a nurse or pharmacist can intervene via a phone call rather than waiting for the next clinic visit potentially avoiding months of uncontrolled pressure. The overall evidence supports that home monitoring, when coupled with clinical response, leads to better risk factor management and could reduce cardiovascular events (Duan et al., 2017[29]).
Patients with HF benefit from at-home monitoring of weight, blood pressure, heart rate, and symptoms. Sudden weight gain can indicate fluid retention and impending decompensation; detecting this early via a connected scale allows for diuretic adjustments and averts hospitalisation. Trials like TIM-HF and others demonstrated that remote monitoring of HF patients (with daily weight and symptom reporting, plus nurse follow-up) reduced HF hospitalisations and improved survival in some studies (Eze, Mateus and Cravo Oliveira Hashiguchi, 2020[22]). As mentioned above, a programme of remote monitoring plus teleconsultation was associated with a 29% reduction in HF hospitalisations, based on a meta‑analysis (Kuan et al., 2022[27]). Other successful examples include the STOP-HF screening programme in Ireland (HeartBeat Trust, 2025[30]).
Description: TeleHomeCare is a digital intervention designed to support home care through telemonitoring and teleconsultation for patients who suffer from one or more of the following chronic diseases: HF, chronic obstructive pulmonary diseases (COPD) and diabetes. TeleHomeCare was initially developed in Ceglie Messapica (a small town near Brindisi, Italy). The intervention involves the patient, caregivers of patients, General Practitioners (GPs), specialists, and nurses working in the area. The device installed at the patient’s home – called Hospital-at-Home (H@H) – allows the patient to monitor physiological parameters, share measurements with control room operators and care providers. All clinical parameters of the patients based at home are centralised in the hospital, which respect all privacy laws. The device allows doctors to have remote consultations with patients via video.
Best practice assessment:
Enhancement options: Monitoring and evaluating clinical outcomes of TeleHomeCare are needed to enhance the effectiveness. While the intervention was evaluated to cost more than it saves, future evaluation of TeleHomeCare can envisage taking a broader perspective, valuing improved quality of life of patients, reduced waiting and travelling times, reduced workload of healthcare workers, and higher work productivity of patients. Efforts can focus on enhancing the internet network to enable access to TeleHomeCare technology and improve access to population groups who are at risk of digital exclusion, in particular older people, disabled people, people in remote locations and those on low incomes.
Transferability: TeleHomeCare is likely to be transferable, since telemonitoring is experimented with in many countries, either at the national, regional or local level. In addition, there is political support given most countries have a national eHealth and telehealth policy or strategy. However, health- and digital literacy are important considerations.
Conclusion: By favouring continuity of care from hospital to the home setting, TeleHomeCare has the potential to reduce costs due to long hospital stays and emergency services use. Further evaluations on what aspects of the intervention work well and do not work well are needed to improve effectiveness.
Source: OECD (2023[20]), Integrating Care to Prevent and Manage Chronic Diseases: Best Practices in Public Health, https://doi.org/10.1787/9acc1b1d-en.
Some HF programmes also use patch-based sensors that track respiratory rate or thoracic impedance (as a proxy for lung fluid) to detect decompensation. One of the more cutting-edge technologies is the CardioMEMS device – an implanted pulmonary artery pressure sensor. This tiny sensor is placed in a patient’s pulmonary artery via catheter and transmits daily pressure readings wirelessly to their doctor. The CHAMPION trial showed that managing HF based on this continuous pressure data led to nearly a 30% reduction in HF hospitalisations over 15 months (Eze, Mateus and Cravo Oliveira Hashiguchi, 2020[22]), an example of an invasive but efficacious digital monitoring tool. While not a “wearable” per se, it exemplifies how digital tech can directly improve management by providing actionable, real-time data from outside the clinic.
Detecting cardiac arrhythmias, such as atrial fibrillation (AF) or sporadic tachycardias, traditionally relied on Holter monitors – portable devices with electrodes that record ECG for 24 to 48 hours. Holter monitors are a well-established tool, and essentially an early form of wearable tech prescribed by healthcare professionals. They have helped diagnose countless cases of intermittent arrhythmia. However, their limited recording duration can miss infrequent events. Today’s digital health toolkit has expanded to include extended continuous ECG monitors, such as adhesive patch monitors, event recorders, and even implantable loop recorders for long-term monitoring.
Studies comparing newer wearable ECG patches to traditional Holters show striking differences in detection yield. In one head-to-head study, a 14‑day continuous ECG patch detected arrhythmias in 66% of patients, whereas a standard 24‑hour Holter caught arrhythmias in only 9% of the same patients (Chua et al., 2020[31]). The patch, worn for 2 weeks, can reveal episodes of atrial fibrillation, flutter, or tachycardia that a one‑day recording missed. Specifically, the patch detected atrial fibrillation in 22% of patients (capturing 202 AF episodes), while the Holter detected AF in just 3% (only 1 episode) in that study. Improved sensitivity is important because finding atrial fibrillation, even asymptomatic, can prompt anticoagulation to prevent stroke. Indeed, prolonged monitoring is now recommended in guidelines for patients with cryptogenic strokes to screen for occult AF. Wearable ECG patches and implantable loop recorders (which can monitor for a year or more) have improved detection elusive arrhythmias, leading to earlier and more definitive treatment. While a higher diagnostic yield means fewer missed clinically significant arrhythmias, detecting short term episodes may add little therapeutic value while creating additional work for providers.
Regulatory frameworks to enable safe, high-quality home‑based monitoring and care for CVD and other high-impact diseases can be further promoted and harmonised across the EU. Eight countries responding to the OECD CVD survey reported developing regulatory frameworks for remote monitoring after acute coronary syndrome and cardiovascular procedures. Only Estonia reported having implemented home‑based care regulation, while four reported having one in development. In terms of telehealth, Estonia, France, Germany and the United Kingdom reported fully implemented regulation (while Ireland, Norway, Latvia and Slovenia reported these to be in development) (Figure 5.1).
Note: N=18 EU+2 and OECD G20 countries (Austria, Canada, Czechia, Estonia, Finland, France, Germany, Iceland, Ireland, Japan, Latvia, Luxembourg, the Netherlands, Norway, Slovenia, Sweden, Türkiye, the United Kingdom).
Source: 2025 OCED Cardiovascular Policy and Data Survey.
The downside of longer monitoring is the volume of data produced – hundreds of hours of ECG that must be analysed. The limitation mainly involves ensuring the data are acted upon (having care team infrastructure to respond to alerts) and dealing with potential false alarms or patient anxiety from constant monitoring. Still, remote monitoring stands out as one of the most tangible successes in digital cardiology, directly linking patients’ daily health metrics to timely clinical interventions. Algorithmic analysis (often AI-powered) is being applied to filter and identify episodes of concern, helping clinicians manage the data deluge.
5.1.5. Wearable devices and smartwatches show promise but uptake skews towards healthier and wealthier users
Wearable eHhealth devices – especially smartwatches and fitness trackers – have seen explosive growth among consumers, and they are increasingly relevant to cardiovascular health. These gadgets blur the line between medical device and lifestyle accessory. Modern wearables come equipped with sensors that measure heart rate, physical activity, and even perform single‑lead ECGs or blood oxygen saturation readings. In the context of CVD prevention and detection, wearables offer a unique opportunity: they allow continuous or frequent monitoring in everyday life, potentially catching early signs of disease or motivating healthier behaviours.
Wearables have seen rapid adoption across high-income countries. Recent market data show that as of 2023, roughly one in three adults in the United States uses some form of wearable health tracker (smartwatch or fitness band) (Dhingra et al., 2023[28]). Global estimates indicate over 400 million people worldwide use smartwatches, with the user base expected to approach half a billion in 2025, with Poland, Germany and France among the top ten adopters (Kumar, 2025[32]). This popularity stems from the appeal of these devices as consumer electronics for fitness and communication, as well as increasing health awareness. Cardiologists see patients already armed with data from their wearable device and health systems have begun pilots to integrate patient-generated wearable data into electronic records for monitoring of conditions like atrial fibrillation or post-operative recovery. As consumer wearables become ubiquitous, their influence on CVD care is rising.
However, disparity in their usage has been observed and, paradoxically, those who might benefit most (older, less healthy individuals) are less likely to be using wearables, whereas younger, health-conscious individuals lead adoption. Although this trend may benefit the latter group as they age, it is important to ensure that current non-adopters are not overlooked. In the United States, fewer than one in four individuals with, or at high risk of, CVD use a wearable device. Those with established CVD who do own a wearable, only about 38% use it regularly, compared to almost half of healthy adults (Dhingra et al., 2023[28]). Ensuring accessibility for older adults, low-income users and people with limited digital literacy is critical to prevent widening cardiovascular health inequalities.
Wearables can also contribute to better cardiovascular health by encouraging physical activity and healthier lifestyles. Fitness trackers count steps, measure exercise duration, and often allow goal-setting with feedback, which can motivate users to move more. There is evidence that these devices succeed in boosting activity levels. A 2022 meta-review encompassing 163 000 participants found that wearable activity trackers (including pedometers and smartwatches) significantly increased physical activity, adding approximately 1 800 extra steps per day and about 40 minutes more walking per day on average compared to controls. Users also saw improvements in weight and body composition – about a 1 kg more weight loss on average and reductions in BMI and waist circumference – when using wearables to support lifestyle change (Ferguson et al., 2022[33]). These modest effects can be clinically meaningful: even a ~2 mmHg blood pressure reduction or a kilogram of weight loss can contribute to lower CVD risk. The review concluded that activity trackers are effective at increasing physical activity across age groups and populations, with benefits sustained over time, and recommended their use as a public health intervention [ibid.].
[My sports monitor] provides me with recommendations regarding physical activities, heart rate, exercise level, [and] sleep/sleep assessment. In addition, I use a health watch to find out about my hypertension, ECG, blood numbers on cholesterol and others.
Ulrich, 65, retired military officer, heart attack and transient ischaemic attack survivor, living with high cholesterol, arrhythmia and heart failure.
Other studies reflect these findings, showing improvements in exercise capacity (e.g. better six‑minute walk test distances) among cardiac rehab patients given wearable monitors and app-based coaching (Zhu, Zhao and Wu, 2024[34]). Wearables can also incorporate features like reminding users to stand if they’ve been sedentary too long, monitoring sleep (poor sleep is a risk factor for hypertension), and tracking stress or heart rate variability – all data that can potentially nudge users toward healthier behaviour or alert them to trends. There is growing interest in leveraging these devices in preventive cardiology programmes; for example, some insurers provide fitness trackers to members as part of wellness incentive programmes, banking on the evidence that increased activity translates into healthier outcomes.
One of the more attention-grabbing impacts of smartwatches has been in the detection of atrial fibrillation and other arrhythmias. Several smartwatch models now have EU-cleared features that can passively monitor pulse irregularity via photoplethysmography (PPG) and actively record a single‑lead ECG. These capabilities turn the watch into a basic heart rhythm monitor.
The Apple Heart Study illustrated the potential smartwatch potential. This study enrolled over 419 000 participants to assess the Apple Watch’s ability to identify undiagnosed AF. Over about four months of monitoring, ~0.5% of participants received an alert of an irregular pulse, and of those who followed up with an ECG patch, 34% were confirmed to have atrial fibrillation. The positive predictive value was 84% for the watch’s irregular rhythm notification – i.e. most notified cases were true AF upon further evaluation (Perez et al., 2019[35]). While the yield of screening was relatively low (only a half-percent got flagged), the system was reasonably accurate and could safely identify new AF cases that might have gone otherwise undetected. No serious adverse events were reported in the study, and about 57% of those notified sought medical attention as advised (Perez et al., 2019[35]). Similar studies, including Fitbit’s Heart Study and various smaller trials have also suggested that wearables can pick up asymptomatic AF with a decent reliability. However, the role and benefit of opportunistic passive smartwatch AF screening in initiating anticoagulation therapy remains to be seen.
The benefits of wearables include empowering individuals to take charge of their health metrics daily. For clinicians, patient-generated data from wearables can provide insights that one‑off clinic measurements cannot, capturing sporadic events or lifestyle patterns. For instance, knowing a patient’s step count trend can contextualise their weight and blood pressure trends.
However, there are important limitations. False positives and over-diagnosis are a concern; an irregular pulse alert may lead to anxiety and a cascade of tests in some cases where no significant pathology exists. The specificity of algorithms can be high but is far from perfect, and remain to be validated, while benign variants like premature beats can trigger alerts. This requires the healthcare system to be ready to process these outputs – triaging which alerts need follow-up. Interpreting consumer ECGs also adds to clinicians’ workloads, and responsibilities and costs associated with possible technical failures may discourage adoption in the medical and pharmaceutical industries. Additionally, there is no guarantee of sustained use. People may abandon their wearable devices after the novelty wears off. Long-term adherence to using the device (and acting on its data) is variable. Privacy and data security of wearable data is another consideration, as these devices often sync to phone apps and cloud services. As mentioned, wearables currently skew toward younger, wealthier users. Older patients with CVD use wearables at lower rates due to cost, access, or technology comfort (Dhingra et al., 2023[28]). This could widen health disparities and thwart the goal of participatory, equitable and effective surveillance of CVD (and other diseases).
While I appreciate the advancements in wearable devices and health apps, I believe there is still a gap in their accessibility and usability for older or less tech-savvy patients. Personalization and user-friendly designs are crucial for these tools to be effective.
Caius, artist, researcher, patient advocate, and heart attack survivor.
Despite these challenges, the trend is that wearables and smartwatches are becoming more sophisticated each year. Some regulators have embraced a pathway for clearing software as medical devices, which has allowed algorithms on consumer devices (such as Apple’s AF detection or AliveCor’s smartphone ECG) to be clinically validated and approved (Grant et al., 2024[11]). As accuracy improves, wearables may contribute not just to detection of disease but to management – for example, guiding titration of exercise in rehab, or monitoring adherence to medications via physiological markers.
In the EU, wearable eHealth devices fall under the Medical Device Regulation (MDR) umbrella and are covered under the EU Regulation on Health Technology Assessment (HTA), under which electronic wearable products face additional scrutiny (EU, 2021[36]). Differences in classification, assessment, and oversight are due not only because of the intangible nature of software, but also because they, by definition, involve the collection of sensitive health data. The additional requirements therefore centre on data interoperability, cybersecurity, and their potential use in machine learning / AI (see Section 5.2). In addition, the EU Regulation on MDR, establishes a framework for the systematic clinical evaluation of high-risk medical devices such as implantable devices or active devices intended to administer or remove medications (ibid.).
Countries report limited use and integration of wearables into national CVD health monitoring (Figure 5.2). Only three EEA countries (Czechia, France and Norway) reported having policies in place to promote the adoption of medical devices such as wearable fitness trackers, home ECG-enabled arrhythmia detection devices for CVD prevention and monitoring. Only one of these (France) reports having comprehensive data infrastructure to support the transmission of data from wearables to a patient’s EHR. Elsewhere, Singapore reports having policies to support the use of medical devices for CVD prevention and monitoring. The Health Sciences Authority (HSA) regulates these devices using a risk-based system, covering both CVD and telehealth tools to ensure safety.
Note: N=19 EU+2 and OECD G20 countries (Austria, Canada, Croatia, Czechia, Estonia, Finland, France, Germany, Iceland, Ireland, Japan, Latvia, Luxembourg, the Netherlands, Norway, Slovenia, Sweden, Türkiye, the United Kingdom).
Source: 2025 OCED Cardiovascular Policy and Data Survey.
Nevertheless, wearables represent a potentially powerful public-facing technology for CVD health – they actively engage people in prevention (by promoting exercise and healthy habits) and provide an adjunct diagnostic tool for early detection of problems like arrhythmias. The empirical evidence shows tangible benefits, but realising their full potential will require addressing usage gaps and integrating these tools responsibly into care pathways. Additional areas that need to be carefully addressed include representativeness of the subjects and data used to evaluate these technologies, the interoperability of their data with health information systems, and – as always – the risk of entrenching inequalities. Attention is needed to governance, privacy and capacity of the health system to absorb additional data generated.
Digital health tools such as remote monitoring and online patient portals have been beneficial in managing my conditions, allowing me to track vital signs and communicate with providers more efficiently. However, there are usability challenges, and not all providers use these tools uniformly. Telemedicine has offered convenience but sometimes lacks the personal touch necessary for complex care discussions.
Antonis, 58, congenital heart disease patient and advocate for digital health and patient empowerment.
5.1.6. Some smartphone apps and mHealth platforms can help but their effectiveness varies widely
Alongside hardware devices, software in the form of mobile health apps and digital platforms is transforming CVD management and prevention. Smartphone apps targeting CVD are incredibly diverse – ranging from simple activity trackers, medication reminder apps and blood pressure diaries to comprehensive cardiac rehabilitation programmes and interactive coaching platforms. These apps are accessible to the public via app stores – for better or worse as their quality and effectiveness can vary considerably. Additionally, SMS text message‑based programmes (an early form of mHealth) have demonstrated some success in behaviour change for CVD prevention, especially in low-resource settings.
Numerous apps aim to improve lifestyle factors that contribute to CVD. For instance, there are apps for smoking cessation (providing tips and tracking quit progress), dietary improvement (logging food intake and providing heart-healthy recipe suggestions), and increasing physical activity (workout planners, virtual coaches). Many of these apps use techniques like goal setting, gamification (e.g. earning points for each healthy meal), and social support (connecting with peers). One example is the TEXT ME study from Australia, where patients with coronary disease received regular motivational and educational text messages; after six months, those patients had significantly lower LDL cholesterol, blood pressure, and body mass index compared to controls (Qi et al., 2025[4]), illustrating that even a relatively low-tech intervention (text messages) can improve multiple cardiovascular risk factors and medication adherence.
A 2024 systematic review and meta‑analysis of mobile health apps for coronary heart disease (CHD) patients found that, across 34 randomised trials with over 5 300 participants, use of cardiac mobile apps led to a 32% reduction in the incidence of major adverse cardiac events (MACE) (RR = 0.68) and a 44% reduction in hospital readmission rates (RR = 0.56) compared to usual care (Zhu, Zhao and Wu, 2024[34]). These apps were also found to improve patients’ exercise capacity (VO₂ max), ejection fraction, six‑minute walk distances, and medication adherence rates. Psychosocial outcomes also improved with users reporting lowered anxiety and depression scores (Zhu, Zhao and Wu, 2024[34]). Such findings suggest that well-designed digital interventions can keep patients out of the hospital by better managing their condition. Notably, the meta‑analysis did not find a significant impact on blood pressure or LDL cholesterol in aggregate (Zhu, Zhao and Wu, 2024[34]), suggesting some risk factors still require more intensive intervention beyond what apps achieved, or perhaps that usual care was already effective in those areas.
My smartwatch counts the steps I take, encouraging me to walk more. I exercise at home using an App. I control my blood pressure through my mobile phone.
Diana, 48, stroke survivor.
In HF, some trials where apps have been used to monitor symptoms and weight and have shown improvements in self-care behaviour and reductions in HF exacerbations (Choi et al., 2023[37]). These tools are designed to help healthcare providers make evidence‑based decisions. Increasingly, consumer apps are bridging to clinical management. Some allow patients to transmit data to clinics (e.g. an app paired with a blood pressure cuff sending readings to a hypertension clinic where clinicians review trends). Electronic cardiac diaries allow patients to log symptoms (chest pain episodes, palpitations) which can be reviewed in follow-up to tailor treatment. On the provider side, digital platforms used by clinicians include apps that calculate risk scores and apps that help guide acute treatment.
Smartphone apps have the advantage of being widely accessible – nearly everyone carries a mobile phone, and app-based interventions are relatively low-cost to scale. They can deliver rich multimedia content (videos on heart-healthy living, interactive quizzes to reinforce knowledge) and can be personalised (for example, adjusting goals based on a user’s progress). They also facilitate frequent interaction: daily or weekly messages keep patients engaged in their care continuously, as opposed to the episodic reinforcement that happens only during clinic visits. This “always-on” support is a key reason why digital interventions often improve adherence – patients are repeatedly reminded and educated, which helps sustain behaviour change or medication routines (Zhu, Zhao and Wu, 2024[34]).
However, not all health apps are effective, and their quality varies considerably. The market is flooded with thousands of “heart health” apps, but only a fraction has been formally evaluated in clinical studies. Many users download health apps but stop using them after a short period (retention is a challenge unless the app is very engaging or the user is highly motivated). Integration can be problematic, where people use apps in isolation and the data may not reach their healthcare providers. For maximum benefit, especially in disease management, apps can be incorporated into programmes where clinicians or coaches actively follow the data and give feedback. Some successful studies provided human support in conjunction with the app (e.g. nurse phone calls when certain thresholds are crossed), making it hard to disentangle the effect of the app alone versus the app plus the human touch.
As always, data privacy and ensuring regulatory compliance (for example, when apps provide medical advice, they need regulatory clearance) are additional considerations. Nevertheless, many organisations are moving towards prescribing vetted apps as adjunct therapy. For instance, the National Health Service (NHS) in the United Kingdom has an app library including ones for managing hypertension or depression, and some insurers in the United States cover programmes like Omada (a digital behaviour change programme for diabetes/CVD risk). As digital health matures, more standardisation and clinical validation of apps is expected, although more high-quality trials are needed to accommodate these tools into mainstream practice (Liu, 2024[38]).
Evidence to date, especially the meta‑analytic data showing reduced cardiac events (Zhu, Zhao and Wu, 2024[34]), underscores that when used properly, mobile apps can measurably improve patient outcomes in CVD. Going forward, integration of apps with wearables (e.g. an app that not only provides coaching but also takes in data from a smartwatch to tailor recommendations) and with healthcare provider dashboards will likely enhance their impact even further.
Health technologies have played a valuable role in helping me manage my complex medical profile more efficiently. I regularly use a mobile app… which consolidates all my medical records, test results, and reports in one accessible platform. This centralized access has made it significantly easier to track my health history, prepare for appointments, and consult with multiple specialists—especially important given the number of chronic conditions I manage.
I have also seen the potential of wearable technology in improving patient comfort and compliance. For example, during one cardiac assessment, I used a smartwatch-based Holter monitor instead of the traditional wired version. This was by far the most comfortable and unobtrusive experience I’ve had when monitoring heart activity, and it allowed me to maintain a normal routine while still collecting accurate data. Such advancements in digital health and wearable diagnostics could greatly improve the quality of care and patient adherence, especially for those with long-term or complex conditions like mine.
Angela, 53, mother, in menopause and living with multiple chronic conditions.
AI algorithms have shown impressive performance in interpreting medical data relevant to cardiology (Figure 5.3). AI-enabled supports for stroke detection and care have been outlined previously. In medical imaging, for instance, deep learning models can analyse echocardiograms with accuracy approaching that of experienced sonographers. In one blinded study, an AI system automatically assessed left ventricular ejection fraction from echocardiographic images and its initial readings were found to be non-inferior to those of expert human sonographers (He et al., 2023[39]). AI may soon assist (or even partially automate) the interpretation of routine cardiac ultrasounds, potentially speeding up workflows and improving reliability (Hirata and Kusunose, 2025[40]). AI has also been applied to analyse 12‑lead ECGs to detect subtle patterns; notably, researchers at the Mayo Clinic developed a neural network that can identify patients with asymptomatic left ventricular dysfunction (low ejection fraction) just from their surface ECG – essentially using AI to find a HF signal in what appears to be a normal ECG (Grant et al., 2024[11]). Such an algorithm can flag patients who would benefit from further echocardiographic evaluation and early intervention.
Another example is the use of AI on chest CT scans performed for other reasons. Here, algorithms can quantify coronary artery calcium on a routine CT scan and thereby estimate CVD risk, even if the scan was done for respiratory problems, for example (Grant et al., 2024[11]). AI-driven analysis of retinal photographs has even been shown to predict cardiovascular risk factors (blood pressure or smoking status) and risk of cardiac events, illustrating the breadth of AI’s pattern-recognition power. Importantly, AI is also being embedded in wearables and apps – for instance, algorithms that analyse PPG waveforms from a smartwatch to detect not just AF but potentially other arrhythmias or even atrial premature contractions burden. AI can also be deployed to parse the large volumes of data generated by wearables and apps, which would overwhelm any human analyst, thus helping to overcome a major limitation of consumer-facing technologies.
Beyond diagnostics, AI is being used to optimise treatment decisions. For example, machine learning models can predict which HF patients are at highest risk of hospitalisation in the next 30 days, helping clinicians prioritise interventions (this is already used in some hospital readmission reduction programmes). AI-based clinical decision support, as an evolution of the rule‑based CDSS discussed earlier, can integrate far more data (genetics, labs, imaging, sensor data) to provide nuanced recommendations – essentially an advanced form of CDSS. Examples include algorithms that analyse coronary angiography images to assist in lesion assessment or help in cardiac MRI interpretation. With these tools, clinicians may gain decision support such as “This patient’s profile suggests a 90% likelihood of responding to catheter ablation for their arrhythmia” or “AI analysis of this echo suggests the patient’s aortic stenosis is severe,” which can supplement clinical judgment.
5.2.1. Large language models have the potential to improve the management of CVD
Another emerging application is large language models (LLMs) being tested to enhance patient communication and health coaching (Grant et al., 2024[11]). Early experiments suggest LLMs can draft answers to patient questions, potentially helping to keep patients informed and engaged outside of brief doctor visits. However, this technology is still nascent and requires further rigorous investigation.
The amalgamation of large datasets from EHRs, registries, wearables, and genomics is allowing for predictive analytics at scale. In research contexts, projects like the UK Biobank and the All of Us project in the United States are correlating genetic profiles with lifetime cardiovascular outcomes to find new risk factors. Clinically, integrated data networks enable something like a national CVD surveillance system – using EHR data across hospitals to track trends in heart attacks, HF admissions, and outcomes in near real-time (Williams et al., 2022[41]). This can guide public health interventions and resource allocation.
With improved interoperability, cardiologists will have more complete data on their patients (for example, knowing about that emergency visit in a different city, or being alerted if their patient’s wearable detects an issue). Algorithms could potentially help tailor prevention strategies to individual risk profiles far beyond traditional risk scores. It could also reduce diagnostic errors and free up clinicians from mundane tasks (like measurements on imaging) to focus on complex decision making and patient counselling. The ESC‑EHRA EORP Atrial Fibrillation Ablation Long‑Term Registry – a prospective observational study conducted across 27 European countries and 104 centres, enrolling consecutive patients undergoing catheter ablation for atrial fibrillation – serves as an example of harnessing data from multiple countries and centres to generate real-world evidence in a way a single centre never could.
Source: Authors based on information from Poterucha, T. et al. (2025[42]), “Detecting structural heart disease from electrocardiograms using AI”, https://doi.org/10.1038/s41586-025-09227-0; Singh, M. et al. (2024[43]), “Artificial intelligence for cardiovascular disease risk assessment in personalised framework: a scoping review”, https://doi.org/10.1016/j.eclinm.2024.102660; Chowdhury, M. et al. (2025[44]), “The Heart of Transformation: Exploring Artificial Intelligence in Cardiovascular Disease”, https://doi.org/10.3390/biomedicines13020427.
However, these technologies bring challenges. Validation and accuracy are paramount – an AI algorithm is only as good as the data it’s trained on, and there have been instances of bias (for example, an algorithm underperforming in minority populations if trained predominantly on white patients). Thus, rigorous testing and regulatory oversight are needed to ensure AI tools are safe and effective across diverse populations (Grant et al., 2024[11]; Oliveira Hashiguchi, Slawomirski and Oderkirk, 2021[45]). Clinician acceptance is another hurdle – some providers are wary of “black box” algorithms. Explainable AI (making AI outputs interpretable) is an area of active development aimed at addressing this concern, but it is not a substitute for trust and transparency. As the use of personal data increases, concerns about privacy are growing; policies need to ensure that patient data remains protected while enabling meaningful insights. There is also a risk of both over- and under-reliance – AI can serve as a valuable support tool, but it is not intended to replace human judgment. Lastly, there’s an equity concern: if sophisticated AI tools only benefit those in high-tech health systems, disparities could widen (Oliveira Hashiguchi, Slawomirski and Oderkirk, 2021[45]; OECD, 2019[17]). Democratising access (for example, deploying AI ECG interpretation even in rural clinics via cloud services) is important. The key with these technologies is implementing them thoughtfully, with attention to evidence, ethics, and equity.
Regulation is also a critical aspect to help this emerging technology advance social and public policy objectives. The European Union Artificial Intelligence Act, which came into force in 2025, is the world’s first comprehensive legal framework governing AI. It adopts a risk-based approach to regulate AI systems, aiming to ensure safety, transparency, and fundamental rights protection while fostering innovation. The Act categorises models based on the level of risk they pose. It also introduces specific provisions for General Purpose AI, including LLMs. Full compliance obligations take effect in August 2026, with non-compliance resulting in significant fines (EU, 2024[46]). The European Commission has developed resources to assist developers navigate the Medical Device Regulation and the AI Act (European Comission, 2025[47]). Like the General Data Protection Regulation (GDPR), the Act has extraterritorial reach because it applies to any system that affects individuals within the EU, regardless of where the provider is based. This positions the Act as a potential global standard for AI regulation.
A range of routinely collected health data can be leveraged to improve CVD prevention and management both directly, and by underpinning the digital technologies outlined above. Each data source – including EHRs, registries, administrative datasets, mortality data and (increasingly) alternative sources like patient-generated information – provides unique value for guiding interventions and policy. However, it is by linking these data, and making them interoperable across jurisdictions and sectors, that their value can be amplified exponentially, and be harnessed to develop and refine potentially game‑changing technologies to prevent and manage CVD.
EHR systems were discussed above mainly in terms of directly improving patient care and outcomes, such as identifying high-risk patients and improving care co‑ordination. But because EHRs contain longitudinal patient data from routine clinical encounters, offering a comprehensive view of individual health histories, their data can be put to work in a number of productive ways. They provide a rich source of information for research on CVD prevention, treatments and outcomes that can enable large‑scale analysis of real-world evidence – although caveats do apply (Davidson et al., 2020[48]).
Seven of the 19 participating countries in the 2025 OECD Cardiovascular Policy and Data Survey reported using EHRs to monitor CVD burden and outcomes (Croatia, Estonia, France, the Netherlands, Norway, Türkiye and the United Kingdom). For comparison, 17 countries report using administrative datasets and mortality data for this purpose (Figure 5.4). This suggests that countries may not be positioned to leverage EHR data for CVD management and research.
Note: N=19 EU+2 and OECD G20 countries (Austria, Canada, Croatia, Czechia, Estonia, Finland, France, Germany, Iceland, Ireland, Japan, Latvia, Luxembourg, the Netherlands, Norway, Slovenia, Sweden, Türkiye, the United Kingdom).
Source: 2025 OECD Cardiovascular Policy and Data Survey.
Disease‑specific registries collect standardised clinical information on people who suffer from a specific condition. Registries record how patients are managed and their outcomes, monitor the quality of care, and provide a rich resource for research. They help track disease patterns, evaluate treatments and adherence to guidelines, and assess the effectiveness of preventive measures in real-world practice. For example, the European Society of Cardiology (ESC) has developed over 20 registries (enrolling 140 000+ patients) through its EURObservational Research Programme to continuously improve patient care and outcomes (ESC[49]). These support quality benchmarking across hospitals and countries, and they enable studies that include patient groups often excluded from clinical trials such as the elderly or those with complex co-morbidities (ESC[49]). Among the countries responding to the 2025 OECD survey, 11 (about half) report having a national Stroke registry and PCI registry, while fewer report having registries for HF, cardiac arrest, acute coronary syndrome and diabetes (Figure 5.5). Only three components of stroke registries – morbidity, mortality and demographic information – feature in every registry reported (Figure 5.6).
Registries can also enable randomised clinical trials. For example, Swedeheart (the Swedish CVD registry) has pioneered the concept of registry-based randomised clinical trials, leveraging high-quality, population-level data infrastructure for efficient and pragmatic research (Doherty DA, 2023[50]; Erlinge, 2021[51]). By integrating trial randomisation and follow-up directly into the existing registry platform, Swedeheart allows investigators to enrol large numbers of patients, capture outcomes through routine clinical data, and minimise the need for costly parallel data collection systems. This approach has produced by the TASTE and VALIDATE‑SWEDEHEART trials, which enrolled thousands of patients across Sweden and produced findings comparable to traditional RCTs at a fraction of the cost (Ole Fröbert, 2013[52]; Erlinge D, 2017[53]). This illustrates how registries can serve not only as surveillance and quality-improvement tools but also as powerful research infrastructure, bridging the gap between clinical research and real-world care delivery.
Note: N=19 EU+2 and OECD G20 countries (Austria, Canada, Croatia, Czechia, Estonia, Finland, France, Germany, Iceland, Ireland, Japan, Latvia, Luxembourg, the Netherlands, Norway, Slovenia, Sweden, Türkiye, the United Kingdom).
Source: 2025 OECD Cardiovascular Policy and Data Survey.
Note: N=19 EU+2 and OECD G20 countries (Austria, Canada, Croatia, Czechia, Estonia, Finland, France, Germany, Iceland, Ireland, Japan, Latvia, Luxembourg, the Netherlands, Norway, Slovenia, Sweden, Türkiye, the United Kingdom).
Source: 2025 OECD Cardiovascular Policy and Data Survey.
Administrative health datasets – such as hospital discharge databases and insurance claims – capture information on healthcare utilisation, procedures, and costs. Although collected for non-clinical purposes, these data can be repurposed to monitor CVD care at the system level. Administrative data allow efficient acquisition of large volumes of patient data, providing a big-picture view of disease burden and treatment. Health services can use such data to calculate indicators like hospital readmission rates after heart attacks or procedure volumes and outcomes by hospital. In Europe, routine hospital admission and procedure data are already used for CVD indicators – for example, tracking rates of acute myocardial infarction admissions or cardiac surgeries across countries (ESC[49]). Administrative data can lack clinical detail, but linking them with clinical registries or EHRs can yield powerful insights (Maciejewski et al., 2022[54]). In many countries, administrative datasets are the only or main source available because registries are expensive and often absent. While registries provide a more complete picture and enable better monitoring, reliance on administrative data alone can pose challenges for accurately assessing CVD care and outcomes.
Civil registration and mortality databases record causes of death, providing an indispensable health outcome measure for CVD. Mortality statistics are a fundamental tool to assess long-term trends and the effectiveness of prevention efforts (Palmieri et al., 2018[55]). The decline in cardiovascular mortality in many countries over recent decades reflects significant progress in prevention and treatment, although disparities remain. These data help identify where CVD continues to be a leading cause of premature death and inform policy priorities. Linking mortality records with patient registries or EHRs enables calculation of survival rates and case‑fatality for conditions such as heart attack or heart failure, offering valuable insights into healthcare system performance. However, mortality data are not always fully accurate, as the underlying cause of death is often based on clinical judgment rather than thorough investigation, which can lead to misclassification.
Additional routinely collected data can further support CVD prevention and management. These include risk-factor surveillance systems that track population-level indicators such as blood pressure, smoking rates, and cholesterol levels, as well as prescription registries. Such data help identify emerging trends, such as rising obesity or diabetes prevalence, that influence cardiovascular risk, and they enable monitoring of preventive interventions – for instance, tracking uptake of hypertension treatment or use of cholesterol-lowering medications. Israel also maintains national registries on CVD-related interventions, including cardiac surgery, transcatheter aortic valve implantation (TAVI), and atrial fibrillation (AF) ablation, which provide valuable insights for improving care and outcomes. As outlined above, patient-generated data from wearables or mobile apps are also becoming more common; when carefully integrated into health systems, these can complement traditional data sources by enabling real-time monitoring of health metrics.
Tracking CVD epidemiology is a cornerstone of effective prevention and management, yet responses to the 2025 OECD survey reveal significant gaps. Seventeen countries reported breaking down epidemiological data by geographic region, and 11 by clinical characteristics such as severity or risk factors. However, only eight countries include socio‑economic indicators like income or educational attainment – critical factors for understanding health inequalities and targeting interventions (Figure 5.7).
Note: N=19 EU+2 and OECD G20 countries (Austria, Canada, Croatia, Czechia, Estonia, Finland, France, Germany, Iceland, Ireland, Japan, Latvia, Luxembourg, the Netherlands, Norway, Slovenia, Sweden, Türkiye, the United Kingdom).
Source: 2025 OECD Cardiovascular Policy and Data Survey.
5.3.1. Data linkage and interoperability across systems is essential but challenging
Linking data from different sources is essential for a comprehensive understanding of CVD and for co‑ordinated care, as no single dataset provides the full picture. Each source captures only part of the story – without integration, critical risk factors, interventions, and outcomes may be overlooked. Data linkage refers to the technical process of connecting datasets, interoperability is the ability of systems to exchange and interpret these data reliably, and governance frameworks establish the legal and organisational conditions for such exchanges (Palmieri et al., 2018[55]). For example, linking a myocardial infarction registry to administrative hospitalisation records and mortality data can derive incidence of heart attacks, 30‑day and 1‑year survival rates, and identify gaps in follow-up care. Data linkage also supports prevention. Combining primary care risk-factor data with hospital records can identify patients who are not yet in registries but are at high risk and need preventive interventions. As outlined above, can also enables research such as clinical trials using large populations in a more cost-effective way. In short, interoperability multiplies the value of individual datasets, enabling more effective CVD research, detection and management across the care continuum.
I use the remote monitoring tool to send the hospital data from my subcutaneous implantable cardioverter-defibrillator once a week. I also use a watch to keep control of my heart beats during exercise. I think they can be useful but always under the supervision of an expert.
Francesca, 34, female living with hypertrophic cardiomyopathy and an implantable cardioverter-defibrillator.
This remains challenging. The 2023 survey of 27 OECD countries found “ongoing challenges in the governance of harnessing EHR data for analytics and research. Many of these challenges have been reported since 2012. They include legal barriers, lack of resourcing, and ongoing resistance from providers. A lack of social consensus, license and trust is a key barrier to using EHR data for these purposes (Slawomirski et al., 2023[15]). While critical issues remain, the last two years have seen encouraging initiatives and investment to improve and to enhance information infrastructure and to address the lack of social consensus and improvement of data use. For example, the recent European Court of Justice decision regarding the secondary use of health data (see Section 5.3.3).
Beyond integration within a country, interoperability across countries has high importance. Cross-border data sharing permits pooling of information from different populations, which is especially valuable for comparing outcomes and best practices. Large differences in CVD mortality and incidence exist between European member states (OECD/European Commission, 2024[56]). Yet it may be unclear how much of this variation stems from risk factors versus differences in healthcare quality or access – or indeed differences in clinical interpretation and coding. Harmonised data collection and exchange allow for benchmarking, which enables national health systems to learn from one other and inform adjustments to policy and practice. Moreover, researchers studying cardiovascular outcomes benefit from pan-European datasets that increase statistical power, which helps to study less common can enable studying uncommon conditions or treatments. Collaborative registries like EuroHeart are encouraging. EuroHart is an ESC project that connects national cardiovascular quality registries across Europe, providing the infrastructure for collecting large volumes of high-quality patient data, conducting research, and monitoring new drugs and devices. The explicit aim is to produce “comparable real-world data in cardiovascular disease… at European level,” by using common datasets across countries, which can help identify effective interventions and support continuous improvement across Europe (ESC[49]).
Achieving such linkage and interoperability is not straightforward. Key challenges include technical and semantic incompatibility, a lack of unique identifiers, legal and privacy constraints, and organisational and cultural barriers (see Box 5.3). Despite these challenges, the imperative for interoperability is widely recognised. The cost of fragmented data is missed opportunities – missed chances to prevent strokes and heart attacks by intervening earlier, and missed insights into which treatments work best in routine care. Strengthening data linkage infrastructure is therefore a high-yield investment for CVD health outcomes. Overcoming these barriers requires clear incentives, leadership, and demonstrating the mutual benefits of data sharing.
Health data are still too often siloed and incompatible. Different hospitals and countries collect data using varying definitions and standards, impeding direct comparison and linkage. The current arrangements mean that many national CVD registries cannot simply be combined because of non-uniform data formats and variables (OECD, 2022[57]) (ESC/EHN[58]). A more practical challenge is reliably matching patient data across databases. There is no common patient identifier across European health systems. Policy experts have noted the need for “the introduction of a unique European patient identification number” to facilitate cross-system data linkage (ESC/EHN[58]).. In the absence of such an ID, probabilistic matching or legal workarounds are used, but these are imperfect and resource‑intensive.
Data protection rules and privacy concerns limit data sharing. To be clear, strong privacy safeguards are essential, variation in how EU member states implement these rules creates uncertainty. For example, interpretations of the EU General Data Protection Regulation (GDPR) differ by country, leading to a fragmented landscape that complicates cross-border health data exchange. Navigating multi-country data projects can be onerous due to differing consent requirements or approval processes in each jurisdiction (OECD, 2022[57]; ESC/EHN, n.d.[58]). Building trust with the public is also paramount; patients and providers need to be confident that data sharing will occur securely and ethically. This is a political challenge, not a technical one.
Health institutions may be reluctant to share data due to concerns about data ownership, potential misuse, or simply the lack of a collaborative culture (OECD, 2022[57]). Across Europe, member states have historically been protective of health data systems (healthcare is largely managed nationally), which can hinder central or regional initiatives. Efforts to standardise data collection, such as the CARDS project establishing common cardiology data definitions, saw limited uptake without strong national buy-in (ESC/EHN[58]).
5.3.2. Interoperable data systems through standards, governance and capacity
Moving toward interoperable, linked health data in Europe will require co‑ordinated action on technical standards, governance frameworks, and workforce capacity. Clear strategies and policy measures can guide this implementation. Establishing and mandating common standards for health data is foundational. Common standards ensure that data collected in different hospitals or countries are comparable and can be aggregated for a European-level view of CVD outcomes, and are composed of semantic standards, technical protocols, and organisational frameworks. Past collaborative efforts like the ESC’s Cardiology Audit and Registration Data Standards (CARDS) defined core datasets for key cardiac conditions and more recently the EuroHeart project is using common data elements across national registries (ESC/EHN[59]). Policymakers can support the development and adoption of standards like standardised EHR data fields, coding terminologies like SNOMED CT or ICD, interoperability protocols like HL7 FHIR, as well as promoting consistent vocabulary in emerging research approaches like the GREG (Guidance and Tools for Real-World Evidence Generation and Use for Decision-Making in Europe) consortium. In practice, this might include EU-wide guidelines or requirements for vendors to ensure EHR systems are compatible with agreed standards.
Robust governance is essential to ensure that electronic health data are accessed and reused responsibly for secondary purposes. The European Health Data Space (EHDS) establishes a common EU framework for this secondary use. Each Member State will designate Health Data Access Bodies (HDAB) responsible for organising access to health data, evaluating data‑access applications, granting permits for authorised purposes, maintaining a dataset catalogue, and ensuring that access to data takes place in Secure Processing Environments (SPEs) (EUR-Lex, 2025[60]).
Through the HealthData@EU infrastructure, the national catalogues are federated to form a single European entry point for dataset discovery. Data holders will supply these catalogues with dataset descriptions prepared in line with the common European framework and technical specifications, allowing authorised users to locate relevant datasets (Raffaelli et al., 2025[61]). Once an HDAB grants a permit, data are made available within an SPE following the common European technical specifications adopted under the EHDS. Several safeguard ensure that data remain protected: processing is limited to anonymised or pseudonymised data and permitted use, activities are logged, and outputs are checked before extraction. Together, the HDABs, the HealthData Infrastructure, and the SPE provide the operational backbone of the EHDS – enabling secure and efficient secondary use of health data for research, innovation, and public-health purposes while maintaining a high level of trust, security, and data-protection.
Compliance with GDPR and other privacy laws is non-negotiable, but greater harmonisation is needed so that researchers and healthcare providers face consistent rules across Europe. The current implementation of GDPR rules covering health research at member state level is fragmented, posing a barrier to multi-country studies policymakers can work toward aligned interpretations. For example, a pan-EU Code of Conduct for health research under GDPR has been recommended (ESC[49]). Such a code could clarify permissible data uses (e.g. defining the public interest in using health data for research) and streamline approval processes. Additionally, investing in privacy-preserving technologies – such as data anonymisation/pseudonymisation techniques and federated analytics that allow analysis without raw data leaving its home country – can help reconcile data sharing with strict privacy requirements (OECD, 2022[57]). The goal is a governance regime where data can be confidently shared for legitimate purposes, under uniform safeguards across the EU.
Description: The Norwegian Cardiovascular Disease Registry (NCDR) is a national health registry established in 2012 to monitor and improve cardiovascular health outcomes in Norway. It collects data on all individuals treated for CVD in specialised healthcare services, as well as those who died from CVD. The registry consist of a core registry and eight condition-specific medical quality registries, some of which are managed by engaged clinicians across different regions. The core registry is maintained the Norwegian Institute of Public Health and includes personal identifiers, administrative data, enabling linkage across datasets. The medical quality registries include detailed information about specific cardiovascular conditions and include patient-reported outcome measures (PROMs) submitted through standardised questionnaires.
Best practice assessment:
Enhancement options: To further enhance the NCDR, integration of real-time data feeds, better integration with electronic health record to inform on comorbidity and expansion to include primary care data could improve timeliness and comprehensiveness of the data. Additionally, incorporating patient-reported experience measures (PREMs) in the registries and social determinants of health would enrich the dataset.
Transferability: The Norwegian model is highly transferable to other countries with centralised health systems and national identification numbers. Key success factors include legal frameworks for data sharing, robust data infrastructure, and stakeholder collaboration. Transferability may be limited in settings with fragmented healthcare systems or weaker digital infrastructure. The registry’s integration with PROMs and public reporting mechanisms requires cultural and institutional readiness to prioritise transparency and patient-centred care.
Conclusion: The Norwegian Cardiovascular Disease Registry exemplifies a best practice in monitoring the quality of cardiovascular care delivery through its comprehensive, validated, and actionable data. The registry supports epidemiological research, quality improvement, and health policy planning by providing high-quality and person-identifiable data, comprising a valuable tool to improve cardiovascular health in Norway.
Source: NIPH (2024[62]), About the Norwegian Cardiovascular Disease Registry – NIPH, https://www.fhi.no/en/nc/cardiovascular-disease-registry/about-the-norwegian-cardiovascular-disease-registry/ (accessed on 11 July 2025).
Interoperability is also about systems and people. Many health systems need upgrades to their IT infrastructure to collect and exchange data efficiently. The ESC has urged that “Member States should be supported to increase the digital maturity of their systems,” backed by EU funding for developing digital infrastructure (ESC[49]). This might include modernising hospital EHR systems, implementing national data integration platforms, and ensuring connectivity for health facilities. Alongside technology, building workforce capacity is critical. Health services need skilled professionals such as medical informaticians, data analysts, and cybersecurity experts to manage and utilise these data systems. Clinicians and public health workers also require training in data literacy – understanding data standards, interpreting real-world data, and using data tools for decision making. Developing this workforce may involve updating medical and nursing curricula to include clinical informatics (the study of how information, data, and technology can be used to improve clinical care), as well as on-the‑job training programmes.
The ESC is working with experts from national registries and clinical cardiology to develop a “suite of data standards for common cardiovascular diseases which are being used by national registries to capture information about patients with cardiovascular disease” (ESC[63]). This EuroHeart initiative aims to support the improvement of cardiovascular care quality and new avenues for medical research. The standards will comprise internationally agreed standardised variables and corresponding definitions. In addition, these data standards aim to align ESC quality indicators with the ESC Clinical Practice Guidelines.
Policy levers and incentives can encourage or require participation in data reporting to ensure comprehensive data capture. For example, contributions to certain disease registries can be made mandatory for hospitals, or financial incentives for reporting quality indicators can be used (“pay for data”). An EU-wide CVD data system will require buy-in from all member states and their health institutions. Continuous data collection can be facilitated through compulsory data collection or incentives (ESC/EHN[58]). Policymakers could, for example, consider tying funding or accreditation to data submission for CVD care, thus embedding data collection into routine practice.
Better data standards and infrastructure will lead to higher-quality data, which in turn enable more insightful analysis and feedback to clinicians and health managers, driving further improvements in care. Over time, an interoperable data system under strong governance can become a powerful engine for learning health systems – where policies and clinical practices continually adapt based on evidence from real-world data.
5.3.3. The necessary policy and regulatory frameworks are being instituted
Policymakers have recognised that data-driven healthcare requires supportive policy frameworks at both EU-wide and health-sector levels. Over the past decade, a combination of horizontal and sector-specific instruments has been developed to enable the secure and effective use of health data for cardiovascular and other conditions entered into force in 2025. The European Health Data Space (EHDS) Regulation establishes a common European framework for sharing and using electronic health data. It pursues two complementary objectives:
1. Empowering individuals to access and share their own health data for continuity of care across borders (primary use); and
2. Enabling secondary use of electronic health data under strict conditions for purposes such as research, innovation, statistics, and public-health policymaking.
For this second pillar, the EHDS creates a harmonised governance model: each Member State designates a Health Data Access Body (HDAB) to evaluate applications and issue permits for authorised purposes; data holders describe their datasets in national catalogues connected to the HealthData@EU infrastructure; and access is granted within Secure Processing Environments (SPEs) operating under common European technical specifications the EHDS introduces mandatory self-certification of Electronic Health Record (EHR) systems against common European specifications, ensuring interoperability, security and portability of health data throughout the EU (Figure 5.9).
Source: Authors, (EHDS[64]), The European Health Data Space (EHDS), https://www.european-health-data-space.com/.
In practical terms, this means relevant healthcare professionals involved in the care of a patient in one country could, in the future, quickly obtain that patient’s medical data from another country via a common protocol, while a researcher could request access to anonymised datasets on heart disease outcomes from multiple sources through a single application process. As the EHDS is implemented, it is expected to streamline data sharing and break down many of the interoperability barriers described earlier, provided that if member states and stakeholders invest in the necessary capacities and trust frameworks.
EHDS is a sectoral component of the broader European Data Strategy. The upcoming implementation of the Data Governance Act and Data Act will also interact with health data by setting rules for data sharing and ensuring fairness in use of data (though not specific to health, they provide general frameworks for data access and portability that complement health-specific laws). GDPR, implemented in 2018, sets uniform requirements for data privacy and security across all member states. This matters for maintaining public trust in health data use. Under GDPR, health data are classified as sensitive personal data, subject to strict protection. The regulation does, however, explicitly allow certain uses of health data for public health and research – for instance, processing is permitted if it is “necessary for reasons of public interest in the area of public health, such as protecting against serious cross-border threats to health…or ensuring high standards of quality and safety of healthcare” (GDPR Art. 9(2)(i)), or if it is for “scientific research in the public interest” under defined safeguards (Art. 9(2)(j)). These provisions mean that valuable activities like disease registries, surveillance, and research on CVD can be lawful, as long as appropriate protections such as ethics approvals, pseudonymisation of data, and data minimisation are in place. These frameworks are distinct but complementary, applying in parallel within their respective scopes: the GDPR governs personal-data protection; the EHDS regulates access and re‑use of electronic health data for both primary and secondary purposes; and the DGA and DA provide cross-sectoral mechanisms on data sharing and fairness. None overrides the others. The EHDS relies on consistency with the GDPR but does not act as lex specialis.
GDPR has provided a common baseline that enables data use by clarifying consent and security requirements, and by mandating Data Protection Officers and impact assessments for large health databases, improve how data is handled. However, as noted above, a challenge has been divergence in how each country implements some of GDPR’s clauses related to research – as well as further conditions, including limitations, with regard to the processing of genetic data, biometric data or data concerning health that may be present at the national level. National legislations and ethics committees interpret differently what is “in the public interest” or what other legal basis can be used for registry-based research. Policymakers are addressing this by seeking greater harmonisation: the European Data Protection Board has been working on guidance to unify the interpretation of GDPR for health research, and the European Commission has been urged to use both soft law (guidelines, codes of conduct) and even new regulations if necessary to reduce fragmentation (ESC[49]; EC, 2025[65]). The EHDS responds to these issues by providing a single, EU-level legal framework for secondary use, replacing the need for separate national interpretations when authorised purposes and safeguards are met.
A recent European Court of Justice ruling (C‑413/23 P, 4 September 2025) has potentially helped to clarify some of the questions raised above. The ruling stated that pseudonymised information is classified “personal data” if the recipient of these data can realistically re‑identify the individual. This implies that if this is not realistically possible, such pseudonymised data may fall outside the full scope of GDPR as long as the data controllers (who hold the encryption key to re‑identification) have implemented all necessary safeguards stipulated under the GDPR (InfoCruia, 2025[66]). While this ruling does is not a “free pass” for use of health data, it offers more clarity under what circumstances these data can be shared and analysed.
Several other EU policies support the creation of a coherent digital-health ecosystem. The Cross-Border Healthcare Directive laid early groundwork for patient data exchange, which the EHDS now consolidates and extends. The EU’s Digital and Data Strategies promote common data spaces built on shared values, while health-specific programmes such as EU4Health and Horizon Europe fund projects that implement EHDS priorities – developing federated data networks, improving data quality, and strengthening the health-data workforce. Several European countries have their own national digital health strategies that align with these EU initiatives. For instance, some Nordic countries have long-established personal identity numbers and linked health registries that serve as models for how data can drive quality improvement in CVD care. The EU’s role is to facilitate the sharing of such best practices and ensure that less digitally advanced health systems receive support to catch up.
Digital technologies are redefining how we prevent, detect, and manage CVD. Across OECD countries, what was once considered futuristic – remote monitoring of a HF patient from their home, a watch detecting an arrhythmia, or an AI helping interpret an echocardiogram – is increasingly part of everyday practice. This chapter reviewed major categories of digital technologies by type, providing evidence of their uptake and effectiveness. Clinical decision support systems and electronic records are improving adherence to best practices and enabling data-driven care (albeit with incremental gains and a need for further outcome improvements). Telemedicine has proven its worth by expanding access and showing outcomes on par with in-person care in many scenarios – even reducing mortality and hospitalisations in HF when used to its full potential. Remote monitoring devices and wearables empower continuous vigilance over health, catching problems early and nudging healthier behaviours; they have demonstrably improved surrogate outcomes like blood pressure control, arrhythmia detection rates, and physical activity levels. Mobile apps and digital interventions are engaging patients beyond clinic walls, leading to better adherence, fitness, and even fewer major cardiac events in studies. And on the horizon, AI and big data promise a new level of personalised, precision cardiology that could further enhance outcomes while potentially streamlining care delivery.
These innovations come with important challenges and limitations and need appropriate clinical evidence for their efficacy and safety. Not all patients or health systems adopt technology at the same pace – issues of access, digital literacy, and reimbursement need ongoing attention so that benefits are widespread. Moreover, technology is never a panacea. It works best in conjunction with, not in lieu of, a strong evidence base, well-resourced and trained clinical teams and lots of patient engagement. Many digital tools generate vast amounts of data, necessitating intelligent filtering (often by other digital means like AI) to avoid overwhelm. Privacy and security need to be rigorously safeguarded to maintain trust in digital health solutions. Policy and regulation can help integrate the public health perspective with the market perspective. Despite these caveats, the overall trajectory is encouraging.
Early empirical results suggest that these digital tools that show potential to improve efficiency, enable more proactive care, and support patients in leading healthier lives with CVD or at risk for it. For instance, something as simple as a text-message programme or fitness tracker can lead to better blood pressure and weight control, which in turn translates to fewer heart attacks and strokes. On a system level, digital connectivity allows for learning and quality improvement at scale. If implemented thoughtfully and safely, digital health technologies can help healthcare professionals and the public. By continuing to rigorously evaluate these tools, address implementation challenges, and ensure equitable access, stakeholders can benefit.
The value of these and other technologies is underpinned by leveraging routinely collected health data: registries, EHRs, administrative and mortality records, and (increasingly) patient-generated data, which offer significant potential to improve CVD prevention and care. When linked and made interoperable, these data sources provide a comprehensive view of patient journeys, support large‑scale population health insights, and enable proactive, data-driven care. EHRs contribute detailed clinical information, while registries offer quality benchmarking and real-world outcomes for underrepresented groups. Administrative and mortality data help track service use, outcomes, and disease trends. However, countries vary in their capacity to leverage these resources, and the full value of health data can only be realised through robust linkage, cross-system interoperability, and harmonised frameworks across jurisdictions.
Achieving this vision requires overcoming persistent technical, legal, and cultural challenges. Incompatible data systems, lack of unique patient identifiers, and fragmented interpretations of data protection regulations (like GDPR) hinder progress. At European level, the European Health Data Space (EHDS) Regulation provides the co‑ordinated framework needed to realise this potential. It creates a common structure for the primary use of health data, the self-certification of EHR systems – and for the secondary use of health data for research, innovation and public-health purposes under clearly defined authorised purposes. Health Data Access Bodies, federated catalogues and Secure Processing Environments ensure that data can be used efficiently and safely across borders, building the trust required for sustained digital transformation. Combined with investments in digital infrastructure, workforce capacity, and common data models and exchange protocols (e.g. ICD, SNOMED CT, HL7 FHIR) these measures can transform routine health data into the backbone of a self-learning health systems – supporting personalised prevention, evidence‑based policymaking and continuous improvement in cardiovascular outcomes across Europe.
References
[1] Avi Cherla, I. (2024), “Trends in avoidable mortality from cardiovascular diseases in the European Union, 1995–2020: a retrospective secondary data analysis”, The Lancet Regional Health - Europe, Vol. 47, https://doi.org/10.1016/j.lanepe.2024.101079.
[22] Carter, H. (ed.) (2020), “Telemedicine in the OECD: An umbrella review of clinical and cost-effectiveness, patient experience and implementation”, PLOS ONE, Vol. 15/8, p. e0237585, https://doi.org/10.1371/journal.pone.0237585.
[37] Choi, E. et al. (2023), “Heart Failure-Smart Life: a randomized controlled trial of a mobile app for self-management in patients with heart failure”, BMC Cardiovascular Disorders, Vol. 23/1, https://doi.org/10.1186/s12872-023-03039-8.
[44] Chowdhury, M. et al. (2025), “The Heart of Transformation: Exploring Artificial Intelligence in Cardiovascular Disease”, Biomedicines, Vol. 13/2, p. 427, https://doi.org/10.3390/biomedicines13020427.
[31] Chua, S. et al. (2020), Comparison of Arrhythmia Detection by 24-Hour Holter and 14-Day Continuous Electrocardiography Patch Monitoring, Acta Cardiologica Sinica, https://www.tsoc.org.tw/upload/journal/1/20200531/acs-36-251.pdf.
[10] Corral-Acero, J. et al. (2020), “The ‘Digital Twin’ to enable the vision of precision cardiology”, European Heart Journal, Vol. 41/48, pp. 4556-4564, https://doi.org/10.1093/eurheartj/ehaa159.
[48] Davidson, J. et al. (2020), “<p>Validity of Acute Cardiovascular Outcome Diagnoses Recorded in European Electronic Health Records: A Systematic Review</p>”, Clinical Epidemiology, Vol. Volume 12, pp. 1095-1111, https://doi.org/10.2147/clep.s265619.
[28] Dhingra, L. et al. (2023), “Use of Wearable Devices in Individuals With or at Risk for Cardiovascular Disease in the US, 2019 to 2020”, JAMA Network Open, Vol. 6/6, p. e2316634, https://doi.org/10.1001/jamanetworkopen.2023.16634.
[18] Dhingra, L. et al. (2023), “Cardiovascular Care Innovation through Data-Driven Discoveries in the Electronic Health Record”, The American Journal of Cardiology, Vol. 203, pp. 136-148, https://doi.org/10.1016/j.amjcard.2023.06.104.
[50] Doherty DA, T. (2023), “Registry randomised trials: a methodological perspective”, BMJ Open, Vol. 13/3, https://doi.org/10.1136/bmjopen-2022-068057.
[29] Duan, Y. et al. (2017), “Effectiveness of home blood pressure telemonitoring: a systematic review and meta-analysis of randomised controlled studies”, Journal of Human Hypertension, Vol. 31/7, pp. 427-437, https://doi.org/10.1038/jhh.2016.99.
[65] EC (2025), European Health Data Space Regulation (EHDS), https://health.ec.europa.eu/ehealth-digital-health-and-care/european-health-data-space-regulation-ehds_en.
[23] Edmiston, K. and J. AlZuBi (2022), Trends in telehealth and its implications for health disparities, Center for Insurance Policy and Research, https://doi.org/10.52227/25241.2022.
[64] EHDS (n.d.), The European Health Data Space (EHDS), https://www.european-health-data-space.com/ (accessed on 11 July 2025).
[8] Ela Marie Z. Akay, A. (2023), “Artificial Intelligence for Clinical Decision Support in Acute Ischemic Stroke: A Systematic Review”, Stroke, Vol. 54/6, https://doi.org/10.1161/STROKEAHA.122.041442.
[53] Erlinge D, O. (2017), “Bivalirudin versus Heparin Monotherapy in Myocardial Infarction”, N Engl J Med, Vol. 377/12, pp. 1132-1142, https://doi.org/10.1056/NEJMoa1706443.
[51] Erlinge, D. (2021), “Registry-based randomized clinical trials in cardiology: opportunities and challenges”, REC Interv Cardiol, Vol. 3, pp. 157-160, https://doi.org/10.24875/RECICE.M21000205.
[7] ESC (2025), ESC Chat - Your Guidelines Companion.
[63] ESC (n.d.), EuroHeart Project Development and Data Standards, https://www.escardio.org/Research/euroheart/EuroHeart-Project-Development-and-Data-Standards# (accessed on 11 July 2025).
[49] ESC (n.d.), The European Society of Cardiology welcomes the European Commission’s initiative to create a European Health Data Space for better healthcare, research, and policymaking, https://www.escardio.org/static-file/Escardio/Advocacy/Documents/European%20Health%20Data%20Space%20Final%20Clean.pdf.
[58] ESC/EHN (n.d.), The development of a centralised cardiovascular data collection across the European Union member states, https://www.escardio.org/static-file/Escardio/EU-Affairs/centralised-cardiovascular-collection-europe.pdf.
[59] ESC/EHN (n.d.), The development of a centralised cardiovascular data collection across the European Union member states, https://www.escardio.org/static-file/Escardio/EU-Affairs/centralised-cardiovascular-collection-europe.pdf.
[46] EU (2024), Regulation (eu) 2024/1689 of the European Parliament and of the Council.
[36] EU (2021), Regulation (EU) 2021/2282 of the European Parliament and of the Council of 15 December 2021 on health technology assessment and amending Directive 2011/24/EU (Text with EEA relevance), https://eur-lex.europa.eu/eli/reg/2021/2282/oj/eng (accessed on 11 July 2025).
[60] EUR-Lex (2025), Regulation (EU) 2025/327 of the European Parliament and of the Council of 11 February 2025 on the European Health Data Space and amending Directive 2011/24/EU and Regulation (EU) 2024/2847 (Text with EEA relevance), https://eur-lex.europa.eu/eli/reg/2025/327/oj/eng (accessed on 30 November 2025).
[47] European Comission (2025), MDCG 2025-6 - FAQ on Interplay between the Medical Devices Regulation & In vitro Diagnostic Medical Devices Regulation and the Artificial Intelligence Act.
[16] European Heart Rhythm Association (2014), “The Atrial Fibrillation Ablation Pilot Study: an European Survey on Methodology and results of catheter ablation for atrial fibrillation conducted by the European Heart Rhythm Association”, European Heart Journal, Vol. 35/22, pp. 1466-1478, https://doi.org/10.1093/eurheartj/ehu001.
[25] Ezeamii, V. et al. (2024), “Revolutionizing Healthcare: How Telemedicine Is Improving Patient Outcomes and Expanding Access to Care”, Cureus, https://doi.org/10.7759/cureus.63881.
[2] Feigin, V. et al. (2025), “World Stroke Organization: Global Stroke Fact Sheet 2025”, International Journal of Stroke, Vol. 20/2, pp. 132-144, https://doi.org/10.1177/17474930241308142.
[33] Ferguson, T. et al. (2022), “Effectiveness of wearable activity trackers to increase physical activity and improve health: a systematic review of systematic reviews and meta-analyses”, The Lancet Digital Health, Vol. 4/8, pp. e615-e626, https://doi.org/10.1016/s2589-7500(22)00111-x.
[26] Gallegos-Rejas, V. et al. (2024), “Key features in telehealth-delivered cardiac rehabilitation required to optimize cardiovascular health in coronary heart disease: a systematic review and realist synthesis”, European Heart Journal - Digital Health, Vol. 5/3, pp. 208-218, https://doi.org/10.1093/ehjdh/ztad080.
[13] Gold, R. et al. (2022), “Effect of Clinical Decision Support at Community Health Centers on the Risk of Cardiovascular Disease”, JAMA Network Open, Vol. 5/2, p. e2146519, https://doi.org/10.1001/jamanetworkopen.2021.46519.
[11] Grant, J. et al. (2024), “Digital health innovation and artificial intelligence in cardiovascular care: a case-based review”, npj Cardiovascular Health, Vol. 1/1, https://doi.org/10.1038/s44325-024-00020-y.
[30] HeartBeat Trust (2025), What is STOPHF?.
[39] He, B. et al. (2023), “Blinded, randomized trial of sonographer versus AI cardiac function assessment”, Nature, Vol. 616/7957, pp. 520-524, https://doi.org/10.1038/s41586-023-05947-3.
[40] Hirata, Y. and K. Kusunose (2025), “AI in Echocardiography: State-of-the-art Automated Measurement Techniques and Clinical Applications”, JMA Journal, Vol. 8/1, pp. 141-150, https://doi.org/10.31662/jmaj.2024-0180.
[24] HSE (2025), Galway virtual initiative brings cardiac care home for patients.
[66] InfoCruia (2025), “Judgment of the Court”, https://curia.europa.eu/juris/document/document.jsf?text=&docid=303863&pageIndex=0&doclang=EN&mode=lst&dir=&occ=first&part=1&cid=16531016.
[14] Kao, D. (2022), “Electronic Health Records and Heart Failure”, Heart Failure Clinics, Vol. 18/2, pp. 201-211, https://doi.org/10.1016/j.hfc.2021.12.004.
[27] Kuan, P. et al. (2022), “Efficacy of telemedicine for the management of cardiovascular disease: a systematic review and meta-analysis”, The Lancet Digital Health, Vol. 4/9, pp. e676-e691, https://doi.org/10.1016/s2589-7500(22)00124-8.
[32] Kumar, N. (2025), Smartwatch Statistics (2025) – Users & Market Share, https://www.demandsage.com/smartwatch-statistics/.
[38] Liu, L. (2024), “The effect of digital therapeutics intervention on improving hypertension management in adults: a meta-analysis of randomized controlled trial”, Hypertension Research, Vol. 48, pp. 456–469, https://doi.org/10.1038/s41440-024-01892-4.
[54] Maciejewski, C. et al. (2022), “Administrative Data in Cardiovascular Research—A Comparison of Polish National Health Fund and CRAFT Registry Data”, International Journal of Environmental Research and Public Health, Vol. 19/19, p. 11964, https://doi.org/10.3390/ijerph191911964.
[5] McClellan, M. et al. (2019), “Call to Action: Urgent Challenges in Cardiovascular Disease: A Presidential Advisory From the American Heart Association”, Circulation, Vol. 139/9, https://doi.org/10.1161/cir.0000000000000652.
[3] Mohan, K. et al. (2011), “Risk and Cumulative Risk of Stroke Recurrence”, Stroke, Vol. 42/5, pp. 1489-1494, https://doi.org/10.1161/strokeaha.110.602615.
[19] Nadarajah, R. et al. (2023), “Future Innovations in Novel Detection for Atrial Fibrillation (FIND-AF): pilot study of an electronic health record machine learning algorithm-guided intervention to identify undiagnosed atrial fibrillation”, Open Heart, Vol. 10/2, p. e002447, https://doi.org/10.1136/openhrt-2023-002447.
[62] NIPH (2024), About the Norwegian Cardiovascular Disease Registry - NIPH, https://www.fhi.no/en/nc/cardiovascular-disease-registry/about-the-norwegian-cardiovascular-disease-registry/ (accessed on 11 July 2025).
[6] Njie, G. et al. (2015), “Clinical Decision Support Systems and Prevention”, American Journal of Preventive Medicine, Vol. 49/5, pp. 784-795, https://doi.org/10.1016/j.amepre.2015.04.006.
[20] OECD (2023), Integrating Care to Prevent and Manage Chronic Diseases: Best Practices in Public Health, OECD Publishing, Paris, https://doi.org/10.1787/9acc1b1d-en.
[21] OECD (2023), “The future of telemedicine after COVID-19”, OECD Policy Responses to Coronavirus (COVID-19), OECD Publishing, Paris, https://doi.org/10.1787/d46e9a02-en.
[57] OECD (2022), Health Data Governance for the Digital Age: Implementing the OECD Recommendation on Health Data Governance, OECD Publishing, Paris, https://doi.org/10.1787/68b60796-en.
[17] OECD (2019), Health in the 21st Century: Putting Data to Work for Stronger Health Systems, OECD Health Policy Studies, OECD Publishing, Paris, https://doi.org/10.1787/e3b23f8e-en.
[56] OECD/European Commission (2024), Health at a Glance: Europe 2024: State of Health in the EU Cycle, OECD Publishing, Paris, https://doi.org/10.1787/b3704e14-en.
[52] Ole Fröbert, M. (2013), “Thrombus Aspiration during ST-Segment Elevation Myocardial Infarction”, N Engl J Med, Vol. 369, pp. 1587-1597, https://doi.org/10.1056/NEJMoa1308789.
[45] Oliveira Hashiguchi, T., L. Slawomirski and J. Oderkirk (2021), “Laying the foundations for artificial intelligence in health”, OECD Health Working Papers, No. 128, OECD Publishing, Paris, https://doi.org/10.1787/3f62817d-en.
[55] Palmieri, L. et al. (2018), “Cardiovascular diseases monitoring: lessons from population-based registries to address future opportunities and challenges in Europe”, Archives of Public Health, Vol. 76/1, https://doi.org/10.1186/s13690-018-0283-3.
[35] Perez, M. et al. (2019), “Large-Scale Assessment of a Smartwatch to Identify Atrial Fibrillation”, New England Journal of Medicine, Vol. 381/20, pp. 1909-1917, https://doi.org/10.1056/nejmoa1901183.
[42] Poterucha, T. et al. (2025), “Detecting structural heart disease from electrocardiograms using AI”, Nature, Vol. 644/8075, pp. 221-230, https://doi.org/10.1038/s41586-025-09227-0.
[4] Qi, Y. et al. (2025), “Digital Health Solutions for Cardiovascular Disease Prevention: Systematic Review”, Journal of Medical Internet Research, Vol. 27, p. e64981, https://doi.org/10.2196/64981.
[61] Raffaelli, F. et al. (2025), “Unleashing the potential of health data: from vision to reality in the HealthData@EU Pilot Project”, European Journal of Public Health, Vol. 35/Supplement_3, pp. iii1-iii2, https://doi.org/10.1093/eurpub/ckaf003.
[43] Singh, M. et al. (2024), “Artificial intelligence for cardiovascular disease risk assessment in personalised framework: a scoping review”, eClinicalMedicine, Vol. 73, p. 102660, https://doi.org/10.1016/j.eclinm.2024.102660.
[12] Singh, R. et al. (2025), “How AI is used in FDA-authorized medical devices: a taxonomy across 1,016 authorizations”, npj Digital Medicine, Vol. 8/1, https://doi.org/10.1038/s41746-025-01800-1.
[15] Slawomirski, L. et al. (2023), “Progress on implementing and using electronic health record systems: Developments in OECD countries as of 2021”, OECD Health Working Papers, No. 160, OECD Publishing, Paris, https://doi.org/10.1787/4f4ce846-en.
[9] Viceconti, M. (2016), “In silico clinical trials: how computer simulation will transform the biomedical industry”, International Journal of Clinical Trials, Vol. 3/2, https://doi.org/10.18203/2349-3259.ijct20161408.
[41] Williams, B. et al. (2022), “Establishing a National Cardiovascular Disease Surveillance System in the United States Using Electronic Health Record Data: Key Strengths and Limitations”, Journal of the American Heart Association, Vol. 11/8, https://doi.org/10.1161/jaha.121.024409.
[34] Zhu, Y., Y. Zhao and Y. Wu (2024), “Effectiveness of mobile health applications on clinical outcomes and health behaviors in patients with coronary heart disease: A systematic review and meta-analysis”, International Journal of Nursing Sciences, Vol. 11/2, pp. 258-275, https://doi.org/10.1016/j.ijnss.2024.03.012.
← 1. Comparable data on the number of approved AI-enabled medical devices in the EU context not currently available.