Technology for better adult congenital heart disease care: the time is now Authors: Sarah R Blake, Michael Gatzoulis Published: 2025 ABSTRACT AbstractBackgroundThe growing population of patients with adult congenital heart disease (ACHD) present complex lifelong care needs that traditional health systems are struggling to meet. Without innovation, gaps in access, timeliness and specialist oversight will widen. Digital health technologies, including artificial intelligence (AI), telemedicine, wearable devices and interoperable platforms offer a unique opportunity to transform care, but their potential in ACHD remain underexplored.Current developmentsAI-driven prediction models show encouraging performance in mortality and event risk but require external validation and lesion-specific adaptation. Telemedicine, accelerated during the COVID-19 pandemic, has demonstrated safety and high patient satisfaction in selected cohorts, yet robust hybrid pathways are lacking. Wearables can capture rhythm, oxygen saturations and activity in real time, but consumer devices remain poorly validated for complex ACHD physiology. Data integration frameworks, such as federated learning, demonstrate global feasibility but face challenges in governance and interoperability.Future prioritiesProgress in ACHD digital health will depend on three imperatives: (1) rigorous, prospective validation of digital tools in congenital populations; (2) equitable implementation, addressing digital literacy, infrastructure and reimbursement; (3) governance frameworks that embed specialist oversight, data privacy and cybersecurity from the outset.ConclusionDigital health is no longer optional in ACHD care. The field risks falling behind broader cardiovascular innovation unless patients, clinicians, technologists and policymakers commit to specialist-led integration. A decisive shift towards validated, equitable and interoperable solutions can transform ACHD management into a more predictive, personalised and preventive discipline. The aim of this viewpoint is to highlight how digital technologies could strengthen ACHD care and define priorities for future adoption. Introduction Patients with adult congenital heart disease (ACHD) now represent one of the fastest-growing groups in cardiology; a direct result of decades of surgical and medical progress.1 This success has created a new challenge: a heterogeneous population with complex, lifelong needs that frequently exceed the capacity of the current healthcare system. Traditional models, reliant on limited specialist centres and episodic clinic visits, are increasingly unsustainable.2 At the same time, digital health technologies, from artificial intelligence (AI) to wearables and telemedicine, are transforming cardiovascular medicine more broadly.3 Recent ACHD-focused reviews, including the comprehensive analysis by Antoun et al, highlight the expanding diagnostic and prognostic capabilities of AI across imaging, ECG interpretation and risk prediction. However, these advances sit alongside persistent gaps in practical implementation, workflow integration, digital equity and real-world feasibility across diverse ACHD services.4 The risk is that ACHD patients, despite being among those most likely to benefit, may be left behind in this digital revolution. This viewpoint argues that the convergence of demographic pressures, clinical complexity and technological maturity makes this a critical moment to rethink how digital tools are deployed in ACHD care. We highlight current developments, identify gaps and outline priorities for clinicians, researchers and policy makers. The message is simple: technology is not an optional add-on but an essential enabler of specialist-led, patient-centred care for the growing ACHD population. Complex care requires specialist involvement While local healthcare providers often serve as the first point of contact for ACHD patients, the responsibility for managing complex cases must remain with ACHD specialists. Studies have shown that inappropriate management of ACHD patients by non-specialists leads to increased morbidity and mortality.57 Digital technologies can bridge geographical barriers by enabling secure, rapid access to specialist input. Remote access to specialist advice, aided by electronic health records and teleconsultations, can be crucial in emergency situations. Emerging studies also highlight the need for real-time triage systems supported by AI-driven alerts to ensure that patients flagged as at risk are escalated rapidly to specialist teams.8 However, these systems must be designed carefully to avoid overburdening clinicians with false-positive alerts, necessitating rigorous validation and clear thresholds for action. Personalised risk stratification Clinical vignette A deep-learning model trained on multicentre ACHD data identifies a 40-year-old patient with Eisenmenger physiology as high-risk for 12-month mortality despite ‘stable”’ outpatient assessments. This prediction triggers closer follow-up and discussion of transplant candidacy, illustrating how AI can complement specialist judgement. AI encompasses a range of computational techniques, including machine learning (ML), which learns patterns from data to support prediction and clinical decision-making. ML applications have expanded rapidly, with substantial advances in arrhythmia detection, ECG interpretation, risk prediction and electrophysiology mapping.9 They show promise in enhancing risk stratification, achieving predictive accuracy rivalling traditional scoring systems.10 Within ACHD specifically, a landmark single-centre study by Diller et al involving over 10 000 ACHD patients demonstrated that deep learning could accurately predict all-cause mortality.11 Across published ACHD and CHD risk-prediction models, typical performance metrics fall within an Area Under the Curve (AUC) range of 0.75–0.90, broadly comparable to or exceeding established clinical risk scores.8 10 11 Despite these advances, most models are based on retrospective data, often limited to tertiary care settings, and lack external validation across diverse populations. Furthermore, models frequently omit dynamic parameters such as physical activity levels, symptom burden or wearable-derived physiological data. Prospective, multicentre studies incorporating patient-reported outcomes and lesion-specific predictors (eg, Eisenmenger physiology, Fontan-associated liver disease markers) are urgently needed. Additionally, the LINK-HF study demonstrated that non-invasive remote monitoring using multiparametric sensors could predict heart failure decompensation up to a week before clinical events, supporting the integration of predictive analytics into ACHD care pathways.12 Future clinical practice should embed AI models within clinical workflows, ensuring human oversight, interpretability and the ability to adapt dynamically as patient conditions evolve. Telemedicine and virtual consultations Clinical vignette A 32-year-old woman with repaired tetralogy of Fallot, living 150 miles from her ACHD centre, reports increasing palpitations during a teleconsultation. A remote ECG patch arranged through her local provider reveals frequent atrial flutter. This prompts an expedited in-person review and catheter ablation, avoiding a potentially destabilising delay in diagnosis. The COVID-19 pandemic created the need for widespread adoption of telemedicine. Virtual consultations have maintained high levels of patient satisfaction and clinical safety for selected stable ACHD patients, with surveys during and after the COVID-19 pandemic reporting positive experiences and endorsement of hybrid models of care.13 14 These findings are consistent with a recent review of telemedicine in ACHD, which reported that over 85% of patients rated virtual visits as convenient and clinically useful.15 European Society of Cardiology position papers propose combining teleconsultation and remote monitoring with in-person services as part of integrated care pathways.16 Stable patients may be seen virtually for interim reviews, but physical assessments, imaging and exercise testing should still occur at defined intervals. Additional ACHD-focused analyses highlight that virtual care is most effective when embedded within structured hybrid models.17 Challenges persist, including ensuring accurate remote blood pressure monitoring, detection of subtle signs of clinical deterioration and equitable access for socioeconomically disadvantaged groups. Future research must define clear criteria for virtual eligibility and establish pathways for transitioning between virtual and in-person care based on dynamic clinical risk. Telemedicine reviews in cardiovascular populations also emphasise the need for improved interoperability, device validation and digital literacy support.18 Wearable technologies and remote monitoring Clinical vignette A 25-year-old man with a Fontan circulation uses a smartwatch integrated with a clinical dashboard. Recurrent desaturation alerts flag worsening cyanosis, prompting timely assessment and medication changes. This pre-emptive action prevents an emergency admission and demonstrates the value of continuous monitoring. Wearable devices such as smartwatches, patch-based monitors and implantable sensors offer real-time tracking of rhythm, activity and saturation.3 19 Randomised trials like the Apple Heart Study have demonstrated the ability of wearables to identify atrial fibrillation with a positive predictive value of 84%.19 20 However, ACHD patients often exhibit complex arrhythmias, residual shunts, or unique haemodynamic patterns that may not be detected reliably with consumer-grade devices. Critically, no major prospective studies have validated wearable monitoring specifically in ACHD populations. In addition, questions remain regarding data integration, thresholds for clinical alerts and patient adherence over time. Recent innovations indicate that wearable-based models have potential for structured monitoring, early detection of physiological change and improved patient engagement. A recent systematic review of 84 studies reported that wearable heart-rate monitors show low mean bias (around 1–2 beats per minute) but wide limits of agreement and variable accuracy, with only two small studies assessing oxygen saturation and demonstrating similarly broad variation.21 These findings suggest reasonable performance in controlled settings but reduced reliability at extremes of physiology, which is an important consideration in ACHD. The clinical validation of wearable devices for arrhythmia detection by Mannhart et al supports their use in continuous rhythm monitoring.22 A recent advisory from the American Heart Association outlines a broader vision for applying such technologies in congenital heart disease care, including integrated tracking of saturation, activity and rhythm.23 Future initiatives should focus on validating wearables against gold-standard diagnostic tools in ACHD, incorporating signal quality analyses and developing lesion-specific algorithms (eg, for right ventricular failure detection in tetralogy of Fallot). Lifelong patient education and engagement Lifelong education is a cornerstone of effective ACHD management. Digital platforms such as mobile apps and virtual reality have been shown to reduce procedural anxiety and enhance understanding of complex cardiac anatomy. Chiu et al demonstrated that a virtual reality-based intervention significantly reduced preoperative anxiety in adults undergoing elective surgery, supporting the role of immersive digital education in preparing patients for invasive procedures.24 Co-designing educational tools with patients, prioritising digital inclusivity and ensuring cultural sensitivity will strengthen engagement and promote health literacy. Incorporating gamification and real-time feedback may further improve adherence and foster proactive health behaviours, particularly among younger ACHD patients. System-level data integration Fragmented health records remain a significant barrier to continuity of ACHD care. Centralised national databases, such as BELgian COngenital heart disease database combining administrative and clinical data and National Congenital Heart Disease Audit in the UK, demonstrate the feasibility of linking clinical, imaging and procedural datasets for large-scale analysis.25 26 More advanced initiatives are now exploring federated learning models, which allow AI algorithms to train across decentralised datasets without moving sensitive patient data.27 Global frameworks such as the OMOP Common Data Model, developed by the Observational Health Data Sciences and Informatics network, have enabled cross-institutional data standardisation to support multicentre observational research.28 Similarly, interoperability platforms like HL7 FHIR and SMART on FHIR applications facilitate secure integration of hospital and patient-generated data streams, creating cohesive and scalable care pathways.29 System-level data strategies must also embed privacy safeguards from the outset. Secure transmission, patient consent protocols and privacy-preserving analytics are critical to maintaining trust while ensuring that sensitive ACHD data can be shared safely for research and clinical innovation. Barriers to clinical adoption of digital tools Despite promising evidence, real-world deployment of digital health tools in ACHD remains limited. Barriers are multifactorial, extending beyond the need for prospective validation. Some barriers are unique to ACHD: wide physiological heterogeneity, limited specialist centres and a relatively small but highly complex patient population. These factors complicate validation of digital tools, which must perform reliably across diverse lesions and surgical histories. Regulatory uncertainty, reimbursement gaps, fragmented information technology (IT) infrastructures and lack of digital literacy among both clinicians and patients impede widespread adoption. Trust is another critical factor: clinicians may be sceptical of ‘black box’ algorithms, while patients may worry about surveillance or data misuse. Reimbursement is a particularly pressing issue. While pilot projects have shown benefit, digital tools are rarely embedded in routine funding models, making sustainability difficult. Policy alignment and clear regulatory guidance are therefore essential to incentivise adoption. Feasibility in low- and middle-income countries (LMICs) must also be considered. Many ACHD centres in LMICs face challenges including variable affordability of digital devices and shortages of specialist clinicians.30 Nevertheless, mobile-first platforms, cloud-based ECG interpretation and regional virtual specialist networks offer scalable pathways for extending ACHD expertise to resource-limited settings, provided appropriate investment and policy support are in place. Implementation science frameworks, such as the Consolidated Framework for Implementation Research, can help structure evaluation of these barriers and guide tailored solutions for ACHD care. Future directions and recommendations To unlock the full potential of digital health in ACHD care, coordinated action is required across multiple domains (figure 1). Priorities include: Rigorous validation of AI and wearable technologies, tailored specifically to ACHD populations. Rigorous validation of AI and wearable technologies, tailored specifically to ACHD populations. Hybrid telemedicine frameworks that define eligibility criteria, escalation triggers and integration with inperson care. Hybrid telemedicine frameworks that define eligibility criteria, escalation triggers and integration with inperson care. Data standardisation and interoperability, expanding registries and incorporating patient-generated data streams into real-time monitoring systems. Data standardisation and interoperability, expanding registries and incorporating patient-generated data streams into real-time monitoring systems. Equitable access, supported through public-private partnerships and policies that address socioeconomic barriers. Equitable access, supported through public-private partnerships and policies that address socioeconomic barriers. Co-produced education and engagement tools, ensuring patient voices shape development. Co-produced education and engagement tools, ensuring patient voices shape development. Robust governance and privacy safeguards, embedding cybersecurity and ethical frameworks into digital infrastructure from the outset. Robust governance and privacy safeguards, embedding cybersecurity and ethical frameworks into digital infrastructure from the outset. Clinician training and policy reform, including reimbursement strategies to ensure digital care models are sustainable. Clinician training and policy reform, including reimbursement strategies to ensure digital care models are sustainable. Using technologies to improve care in adults with congenital heart disease. AI, artificial intelligence. Ultimately, these strategies must align with the central principle of specialist-led ACHD care, ensuring that digital health complements, rather than replaces, expert oversight. Conclusion Digital health in ACHD is no longer a future possibility but an immediate necessity. The tools exist to transform risk stratification, patient engagement and lifelong care, but the progress has been fragmented and slow. We argue that delaying integration risks worsening inequities and missing the unique opportunity to reshape outcomes for this growing population. What is needed now is specialist-led validation, equitable access and decisive policy action. We call on clinicals, ACHD networks, technologists and policymakers to act with urgency: to validate, integrate and scale digital solutions that are safe, inclusive and clinically meaningful. The time to move from pilots to practice is now.