Kyongjun Kwak
OECD Economic Surveys: New Zealand 2026
Abstract
New Zealand’s health system faces growing pressure from population ageing, workforce shortages and persistent inequities in access to care. Digitalisation, including electronic health records, telemedicine and artificial intelligence, offers opportunities to improve efficiency, service quality and system resilience. Evidence across OECD countries shows that well-designed digital tools can enhance diagnostic accuracy, expand access and reduce administrative burdens. Although adoption remains uneven, recent reforms have strengthened New Zealand’s efforts to digitalise the health system. These include the 10-Year Health Digital Investment Plan (HDIP), supporting implementation structures such as the Centre for Digital Modernisation of Health, and key assets such as the National Health Index and patient portals, which provide a strong foundation for future progress. To build a more integrated and people-centred health system, priorities include clearer governance arrangements, updated data and AI regulations, sustained investment in interoperable infrastructure, and stronger digital skills. Co-design with Māori communities and clinicians will be essential to ensure inclusion and trust.
Most core health services in New Zealand are publicly funded and universally available, within a system that includes public sector agencies, Health New Zealand (Te Whatu Ora) and a mix of community and private providers. Primary and community care, including general practice, are often the first point of access to the health system, making care continuity and coordination especially important as demand rises. Like many OECD economies, New Zealand faces growing structural pressures from population ageing, with implications for both health outcomes and system capacity. The share of people aged 65 and over has increased steadily, and older adults are expected to live longer than the OECD average (Figure 3.1). This means that a larger proportion of the population will require sustained, coordinated management of multi-chronic and age-related conditions over a longer period of life, with lifestyle-related factors such as the high prevalence of obesity further contributing to long-term care needs and system demand. Meeting this need effectively will require strengthening care continuity, primary care capacity, and system-wide coordination.
Note: For non-European countries, life expectancy at 65 has been calculated as the unweighted average of life expectancy of men and women.
Source: OECD Health Statistics (database) and Eurostat.
However, compared with other OECD countries, New Zealand has lower levels of some health system inputs. Health spending per capita and as a share of GDP is slightly above the OECD average, but the availability of key inputs such as the health workforce and hospital beds remains below that of many peers (Figure 3.2 and Figure 3.3). New Zealand’s number of practising doctors per capita remains below the OECD average, and there are regional disparities in access to medical professionals (Figure 3.4). New Zealand faces high competition for medical professionals from Australia and elsewhere, where wages are higher. These workforce constraints place sustained pressure on primary and specialist care services, particularly in rural and high-needs communities.
2023 or latest, the lowest/highest value and the OECD average
Note: The figure shows New Zealand, the OECD country with the lowest/highest value and the OECD average, Canada, the Netherlands, and the United Kingdom. 2024 value for New Zealand.
Source: OECD (2025a), Health at a Glance 2025.
2024 or latest
Note: Data on intensive care beds for Iceland, Mexico and New Zealand include neonatal and paediatric ICU beds.
Source: OECD Health Statistics 2025.
1. 2019 data for New Zealand.
Note: In Panel B, the value for District of Columbia (12.13) is not shown for the United States.
Source: OECD Health Statistics and OECD Regional database.
Pressure on front-line services has also increased as capacity constraints interact with rising demand. Waiting times for specialist assessment and elective treatment have risen in recent years (Figure 3.5), and timely access to general practice has become more constrained. Assuming a continuation of past cost trends, these pressures are projected to lead to an increase in health and long-term care spending by around 2.6 percentage points of GDP by 2060, with larger increases possible over a longer horizon (Chapter 1). The authorities, cognisant of this pressure are aiming policy efforts at moderating cost growth over time.
Source: Policywise and OECD Health Statistics (database).
Digital technologies offer an opportunity to ease these pressures on the health system within existing system and fiscal limits by raising productivity and improving outcomes. This will require some well-targeted upfront investments to scale up these technologies, alongside ongoing support for systems maintenance and upgrades. Historically, technological progress in health has tended to expand the range and volume of treatments, which has contributed to rising public health expenditure (OECD, 2017; Marino and Lorenzoni, 2019).
However, digitalisation, and especially artificial intelligence (AI), may differ from earlier waves of health technology. Digital tools and AI, unlike existing technologies, have the potential to increase productivity by automating labour-intensive tasks in health services delivery such as writing clinical notes or triaging patient information. While their benefits are still emerging, these tools could play a critical role in improving efficiency, service delivery and prevention, thereby enabling fiscal cost savings or the provision of more health services within existing resources. Beyond cost pressures, AI also has the potential to ease workforce constraints by supporting clinicians facing persistent shortages, which is increasingly important for maintaining access and improving health outcomes. By easing work pressure and pushing the system closer to the technical frontier, digital and AI tools can also help attract and retain medical professionals against strong competition internationally. However, robust system-wide economic evaluations remain limited, as many tools are still evolving rapidly, and realised gains depend on effective implementation, workflow redesign and adoption at scale.
Digitalisation is reshaping health systems across OECD countries. The growing use of electronic health records (EHRs), telemedicine and AI offers opportunities to expand access, improve quality, and make care more affordable. International evidence shows that well integrated digital tools in existing workflows can expand access to services, improve the quality and safety of care, and help reduce costs by streamlining clinical and administrative workflows.
AI applications, for example, are increasingly used to enhance diagnostic accuracy, detect errors and reduce the burden of documentation, though their uptake also raises new questions about data protection, transparency and accountability. Diagnostic errors alone are estimated to account for up to 17.5% of health spending in OECD countries (Slawomirski et al., 2025), suggesting large potential efficiency and safety gains from more systematic use of digital and AI-enabled solutions. International experience provides early evidence of such gains. In Denmark’s population-based breast cancer (mammography) screening programme, the introduction of AI-supported reading has been associated with improved or non-inferior cancer detection performance, alongside reductions in recall rates and false positives (Lauritzen et al., 2024).
Beyond diagnostics, AI tools can also help reduce administrative workload by automating documentation and routine workflow tasks. Recent estimates suggest that up to 30% of these tasks could be partially automated through digital solutions, allowing clinicians to devote more time to direct patient care (OECD/European Commission, 2024). In the United Kingdom, National Health Service (NHS)-sponsored trials of AI-assisted clinical documentation report meaningful reductions in time spent on clinical notetaking, with modelling suggesting that scaling such tools could generate material capacity gains and help ease pressure on scarce clinical staff (GOSH/NHS England, 2025). Together, these applications can support new models of preventive and community-based care, enable remote consultations and data-driven planning, and improve system performance within existing resource constraints.
Accelerating the uptake of digital technologies can strengthen the resilience, performance, quality and equity of New Zealand’s health system. Interoperable EHRs can improve continuity of care and reduce duplication, telehealth can expand service reach and convenience, and AI can support more timely and accurate diagnosis while automating administrative tasks. Strategic investment in secure and interoperable infrastructure, driven by a people-centred design, could help relieve the burden on New Zealand’s health workforce and reduce persistent disparities in health outcomes. If deployed inclusively, these tools can also help address long-standing health inequities for Māori and Pasifika populations by lowering cultural and geographic barriers to access.
New Zealand has made progress in developing the foundations of a digitalised health system. The creation of Health New Zealand | Te Whatu Ora and its National Digital Services Directorate in 2022 established a single structure to coordinate investment and set system-wide priorities. The 2025 Budget introduced funding for digital related services, which supports integration including the roll-out of a 24/7 virtual-care service and the expansion of e-prescribing across general practice. In October 2025, Health New Zealand also initiated a national rollout of an AI-assisted clinical documentation tool in emergency departments, with around 1 000 licences procured following successful pilots in Hawke’s Bay and Whanganui. These steps mark an important shift towards a more integrated and patient-centred digital ecosystem. In November 2025, the government released a 10-Year Health Digital Investment Plan (HDIP), providing clearer long-term priorities for interoperability (i.e., the capacity to combine and use data from disparate digital tools with ease, coherence and efficiency), core infrastructure, cybersecurity and digital equity. The Plan is complemented by the establishment of the Centre for Digital Modernisation of Health (CDMH), which provides an implementation mechanism and capability-building support for delivery, including through the Digital Academy programme to support a sustained increase in digital capability among health professionals, while strengthening digital literacy across the wider Health NZ workforce. The plan offers a clearer strategic direction for digital transformation and signals a shift toward more coordinated, sequenced investment over the coming decade.
New Zealand’s core digital assets provide a solid base for further progress. The National Health Index (NHI) covers around 98% of the population and provides a unique system-wide patient identifier. Patient portals and telehealth platforms expanded rapidly during COVID-19 and remain in regular use, supporting functions such as viewing test results and managing prescriptions. Telehealth consultations account for around 7–8% of outpatient appointments and are increasing towards the government target of 10% (Figure 3.6, Health New Zealand, 2025). However, uptake varies across regions and population groups, with lower telehealth use observed among Māori, Pacific and low-income communities. New Zealand’s national pharmaceutical data collections also offer a comprehensive source of prescribing and utilisation information, supporting both clinical care and system planning.
Share of medical appointments via telehealth, by ethnicity and region, 2025Q1
Note: The four regions of New Zealand are Te Waipounamu (South Island), Te Manawa Taki (Midland), Central, and Northern.
Source: Health New Zealand (2025), Quarterly Performance Report Quarter Three 2024/25.
Some regional and sector-specific initiatives provide the building blocks for a fully digitally enabled system. The South Island (Te Waipounamu) has developed a shared record system (HealthOne) that allows clinicians to access patient information across district boundaries. In the North Island, several regions also operate shared clinical portals or integrated care record programmes, although approaches differ by locality. These initiatives demonstrate the potential for locally-driven solutions to improve continuity of care and clinical decision-making. In the private sector, a range of technology firms are active in developing digital solutions, and several hospitals are experimenting with the use of digital assistants, robotic process automation and integrated mental health models.
Despite these initiatives, the system is fragmented and uptake across the health sector continues to vary. Data systems are not yet fully aligned, a national shared-care plan is still under development, and digital maturity differs across providers. The Hira programme, launched in 2021 to connect health data through a federated model (i.e., a model in which data are held in distributed locations, such as hospitals, primary care practices or regional systems, and are accessed through common interoperability standards rather than being copied into a single central repository), was paused in 2024 after implementation challenges and limited buy-in.
While its aims of national data standards and shared-care capabilities remain valid, future strategies are taking a modular, phased approach, in which specific capabilities are prioritised and deployed incrementally to reduce implementation risk rather than through a single large-scale programme as attempted under Hira. As part of this transition, Health New Zealand has begun national implementation of a Shared Digital Health Record (SDHR) data service, providing clinicians with a secure, consolidated view of patient information drawn from multiple systems. The SDHR is being rolled out incrementally, beginning with core clinical data and expanding region by region, with full national coverage expected by mid-2026.
Hospital adoption has also progressed at a variable pace. Stakeholders noted that variability in hospital IT systems limit continuity of care and may make it harder to attract younger clinicians who expect digital-first workplaces.
Funding volatility further constrains progress. New Zealand spends a slightly higher share of GDP on health than the OECD average, but capital investment in health infrastructure remains slightly below peers and is volatile, suggesting scope for more sustained investment to modernise clinical systems and enable data-driven care (Figure 3.7). Recent reprioritisations of digital funding, including the reallocation of NZD 330 million back to the central budget in 2024, have been associated with implementation uncertainty. The 2025 Budget introduced targeted investments in digital health. More stable funding would provide certainty for procurement and support increase in digital and AI adoption.
Source: OECD Health Statistics, Statistics New Zealand and OECD National Accounts database.
Regulatory and cultural factors also play a role. Overly rigid workforce-planning frameworks can limit flexibility for deploying new digital tools. As in other countries, some clinicians exhibit caution towards AI adoption, citing legal liability and professional responsibility concerns. In radiology, for example, questions remain about who bears responsibility if AI findings diverge from a radiologist’s judgement. Such uncertainty risks discouraging innovation. At the same time, training and institutional support for the use of new AI-enabled technologies across both clinical and supporting functions remain uneven, and younger clinicians increasingly expect digital assistants and automation to reduce administrative workload, revealing a generational tension between caution and enthusiasm.
Responsibly scaling digital health solutions remains constrained by limited access to capital and fragmented procurement pathways, making it difficult for promising tools to move beyond pilot stages or achieve national uptake. In response to these structural challenges, New Zealand has adopted a more deliberate, system‑level approach through the HDIP and the establishment of the CDMH, emphasising the need to address long‑standing infrastructure deficits and legacy system constraints as a foundation for sustainable scaling. While the private sector is increasingly active, including start-ups, and some regional hospital pilots are experimenting with, for example, mental health and automation tools, these innovations often remain localised rather than scaled system wide. This highlights the need for more coherent and compatible mechanisms to support growth-stage investment and enable broader adoption across the health system, such as aligned procurement frameworks and evaluation standards.
Equity considerations are important in the expansion of digital health. Telehealth uptake has improved access overall but varies widely across communities. According to Health NZ, while Community Services Card holders, who are eligible for income-tested health subsidies, account for around one quarter of telehealth consultations, Pacific peoples, despite higher health needs, represent only around 6%. High GP enrolment (around 94–97%) provides a strong foundation for continuity of care, yet broadband gaps and persistent digital literacy barriers continue to create a digital divide. Smaller clinics, including some Māori and Pacific providers, face difficulties in adoption due to procurement and financial constraints.
Overall, New Zealand has put in place important foundations for digital adoption, but the absence of a comprehensive, nationally integrated EHR, among other factors, has slowed progress. Establishing secure and interoperable infrastructure, supported by clear national standards, will be a necessary first step, while stronger governance and multi-year planning frameworks will also be critical enablers. Recent initiatives such as the HDIP and the CDMH reflect recognition of their importance. Addressing structural and cultural barriers, investing in workforce upskilling and training, and ensuring digital inclusion for Māori and Pacific populations will also be essential for long-term success.
International experience suggests that successful digitalisation of health systems requires robust and compatible governance, interoperable and secure infrastructure, and sustained investment in organisational change (including workforce capacity building). The Global Digital Health Partnership (GDHP) and OECD Policy Repository maps these enabling conditions across countries, providing an overview of national approaches to areas, such as data-sharing frameworks, cloud and cybersecurity strategies (GDHP/OECD, 2024). New Zealand, like other OECD countries, has progressed across all four broad domains including analytics and use, integrated health data, technical infrastructure, and workforce and human factors (Figure 3.8) including through a unified cross-agency digital framework through the Government Chief Digital Officer (GCDO). Recent OECD research on scaling AI in health also suggests that New Zealand has put in place some important elements for readiness, including a national oversight mechanism and a national approach to strengthening workforce capability, although the strategy is incomplete and some policy tools could be further developed (OECD, 2026; Box 3.1).
Total number of policies, 2023/24
1. Average of 12 OECD countries.
2. Semantic interoperability means ensuring that the structure and meaning of data is preserved as the data flow across separate systems. Among the key tools for achieving semantic interoperability are metadata repositories and logical data models (i.e., conceptual structures that catalogue the full set of data entities and data elements while specifying the relationships between them). For more information, see OECD Digital Education Outlook 2023.
Source: GDHP/OECD Policy Repository Tool.
AI adoption in health is advancing rapidly, with applications emerging in clinical administrative and operational automation, clinical decision support, and diagnostics and predictive analytics. However, progress remains localised and uneven across OECD health systems, reflecting differences in data foundations, governance arrangements, workforce readiness and levels of public trust. To support responsible scale, the OECD’s recent report, Scaling Artificial Intelligence in Health (OECD, 2026), presents a sector-specific policy checklist to assess countries’ readiness to adopt and scale AI in health systems. The checklist emphasises that consistency within countries and compatibility across countries are essential to enable safe scaling, reduce duplication and avoid reinforcing fragmentation.
The checklist is organised across four pillars that guide countries’ readiness for AI-enabled health systems:
Enablers (e.g., high-quality and linkable data, interoperability, secure cloud infrastructure, workforce capability),
Guardrails (e.g., clarity on liability, oversight responsibilities and post-market monitoring),
Engagement (e.g., public participation, provider involvement, trusted communication), and
Ethics (e.g., fairness, transparency, proportionality, person-centred design).
Rather than prescribing a single model, the framework provides guiding questions that help countries assess whether foundational conditions for responsible AI are in place, while preserving national autonomy.
New Zealand appears to have elements of the checklist in place, including strong health data-protection legislation, nationally coordinated digital governance structures, and an established oversight mechanism through the National AI & Algorithm Expert Advisory Group. Recent OECD analysis also suggests that New Zealand is among a smaller group of countries that have established a national oversight mechanism for AI in health and a national approach to strengthening workforce capability. National public engagement initiatives and public assemblies for the design of AI are also in place. At the same time, interoperability implementation, formal ongoing monitoring of AI tools once deployed (for example, to detect performance drift, safety incidents and unintended bias), and the scaling of proven use cases remain areas where further system-wide coordination could strengthen readiness.
Source: OECD (2026), Scaling Artificial Intelligence in Health.
New Zealand has laid important foundations for digital health, but fragmentation, short budget cycles and uneven adoption continue to limit impact. Progress now hinges on five mutually reinforcing pillars: (1) clearer national governance, (2) modernised regulation and trust frameworks, (3) sustained investment in interoperable infrastructure and scalable deployment, (4) workforce capability and clinical adoption, and (5) equity and cultural inclusion. This section sets out practical directions under each pillar, drawing on domestic experience and international practice.
3.3.1. Strengthening national governance and strategic coherence
Over the past three decades, regional autonomy contributed to the proliferation of disparate systems, with more than 6 000 separate servers and locally tailored workflows. The key constraint is not technological capability, but rather the lack of sustained strategic direction and continuity over time. Frequent organisational restructures and shifting political priorities have made it difficult to complete essential foundational work, slowing progress toward shared data infrastructure and consistent national standards.
The establishment of Health New Zealand in 2022 was intended to consolidate responsibility for planning, commissioning and data oversight, and to achieve greater system-wide coordination and economies of scale. Replacing the 20 district health boards (DHBs) that had overseen public health services since 2001 with two national entities required significant operational and governance adjustments, including changes in leadership structures, funding pathways, and lines of accountability. While centralisation has improved visibility over system performance and investment priorities, it can also entail trade-offs, including reduced regional flexibility and slowed local innovation. International evidence suggests that national direction is most effective when complemented by some operational flexibility, allowing local services to adapt and iterate as new digital and AI-enabled tools are introduced (Scheiber et al., 2021).
Delays in procurement and uncertainty over the respective responsibilities of the Ministry of Health, Health NZ and other central digital authorities for digital standards, operational delivery, and data stewardship in some cases have contributed to delays in implementation, highlighting the need for governance arrangements that combine national direction with practical flexibility for local clinical contexts. The reform has also involved substantial one-off restructuring and system-integration costs, which placed additional pressure on health budgets during the transition period. Budget allocations in 2022/23 and 2023/24 supported system stabilisation and the transition to Health NZ (New Zealand Government, 2022).
The decision to pause the Hira programme in June 2024 reflected a change in direction for the organisational and governance arrangements for national digital health initiatives in New Zealand. The pause followed a period in which Hira faced practical challenges, including securing sustained provider engagement, confirming long-term funding settings and establishing scalable delivery arrangements. While Hira did not proceed as originally designed, its core objectives, including interoperability, shared care and patient-centred access to health information, continue to be supported through subsequent initiatives and system-wide reform settings. Following the pause of Hira, responsibility for planning and delivery of the digital health modernisation agenda has transitioned to updated system-level arrangements. The HDIP and the establishment of the CDMH now provide the primary framework for coordinating national digital health investment, setting priorities, and overseeing delivery, including initiatives that had previously been within Hira’s scope.
New Zealand has continued to favour a federated approach to health data sharing, reflecting privacy considerations and public expectations relating to centralised repositories. International experience indicates that both centralised and federated models can be effective when supported by strong governance, clear accountability and enforceable interoperability standards. For example, Denmark has progressively moved from regionally distinct electronic health record systems toward more consolidated national platforms, supported by shared procurement frameworks and common clinical data standards. Estonia has retained a federated approach while enabling seamless data exchange through a national unique identifier and a secure interoperability layer (X-Road), supporting consistency and transparency without centralising data storage.
Comparable developments are emerging in Australia, where the establishment of the Australian Centre for Disease Control aims to strengthen national coordination and data-sharing capabilities across jurisdictions, supported by legislation defining its analytical and public-health data functions. Similarly, in the United Kingdom, the UK Health Security Agency is consolidating key public-health datasets into an Enterprise Data and Analytics Platform to enable secure, standards-based reuse of data for surveillance, planning and service improvement. Both cases illustrate that clarifying stewardship roles and shared data responsibilities can improve system-wide coherence, regardless of underlying technical architecture (Fellner, Sutherland and Vujovic, 2025).
The HDIP adopts a modular, phased implementation model, enabling incremental deployment aligned with organisational capacity and budget cycles. It reflects a revised approach, described as “federated but governed”, that prioritises common data standards and clear stewardship roles, rather than large, single-stage platform builds. This model aims to preserve local flexibility while ensuring national interoperability and continuity over time.
The Plan sets out a structured and sequenced roadmap for digital transformation over 2025–35. It outlines three investment phases: stabilise through strengthening digital foundations (2025–28), modernise through expanding interoperable services and data-sharing capabilities (2028–32), and transform by scaling advanced analytics and AI-enabled tools (2032–35) (New Zealand Government, 2025). Priority areas include national interoperability standards, cloud-based core infrastructure, secure API-enabled exchange frameworks, improved identity and access management, and targeted measures to reduce digital inequities. The plan also introduces clearer national governance arrangements and a consolidated view of system-wide capital requirements, aiming to reduce duplication and improve alignment across regional and sector initiatives. Durable political and institutional commitment, along with clearer implementation pathways, staged action plans and clearer expectations of returns, will be essential to sustaining progress.
Strengthening execution capability will also be critical. High staff turnover, particularly within Health NZ’s Digital Services Directorate, have limited the development and retention of internal delivery expertise, affecting continuity in implementation and potentially slowing the uptake of digital tools. In the current fiscal context, demonstrating clear delivery capability and value for money will be increasingly important. Sustained investment in programme management, engineering and clinical informatics skills, alongside stronger co-design processes with frontline clinicians, would help build a more stable execution base.
Clarifying and empowering the respective roles of the Ministry of Health (policy, standards and oversight) and Health NZ (delivery and operational data stewardship) would support more predictable planning and reduce coordination delays. Responsibilities for different elements of health system digitalisation have historically been distributed across the Ministry of Health, Health NZ and the Government Chief Digital Officer under the Public Service Commission, which in some cases contributed to slow decision-making and diffuse accountability. Strengthening joint governance arrangements, such as establishing a permanent national steering group with clinical, Māori and community representation, could support more consistent decision-making and build trust across the system. Importantly, once governance settings are established, avoiding further large-scale structural reorganisations will be essential, as repeated restructuring has disrupted continuity and slowed progress in digital health reform.
3.3.2. Updating regulation and enabling frameworks
Effective regulation can accelerate the adoption of digital health and AI technologies by reducing uncertainty and strengthening trust. New Zealand currently lacks a unified data-governance and AI framework for health, leading to variable standards and inconsistent interpretations of privacy obligations across providers and regions. While national interoperability standards have been developed, implementation has been uneven, and legacy systems and vendor lock-in make it difficult to operationalise these standards in practice. Current legislative frameworks for health data, including the Health Information Privacy Code (2020), were not designed with cloud-enabled platforms and AI-enabled clinical tools in mind, leaving gaps in areas, such as data localisation, cross-border storage and third-party algorithmic processing. Work under the HDIP is expected to clarify authorising environments for secure access and reuse of health data, alongside stronger governance and trust settings.
Data security considerations also shape policy settings. In 2021, a ransomware cyberattack on the Waikato DHB (the former regional health authority serving the Waikato region) disrupted clinical and administrative systems across five hospitals for several weeks, severely affecting service delivery. The incident prompted strengthened cybersecurity requirements, reflecting the increasing economic value and vulnerability of health data. In 2022, a national Cybersecurity Uplift Programme was launched to enhance threat-detection, incident-response and recovery capabilities across hospital networks. While these measures have improved system resilience, they have also contributed to a cautious approach to data-sharing and digital experimentation. Strict security protocols coexist with hospital IT systems that are slow to update, creating operational trade-offs between security compliance and timely system upgrades and, in some cases, delaying the adoption of new digital tools. More recently, a ransomware breach affecting a privately operated patient portal used by some general practices highlighted ongoing vulnerabilities beyond the public hospital sector. In early 2026, the government commissioned a review to assess the causes of the incident, the adequacy of data protections and lessons for strengthening system-wide cybersecurity and incident response.
Clearer and adaptive regulatory guidance on permitted data uses, cloud environments and conditions for AI-assisted clinical decision-making will be critical to enabling responsible innovation. Ensuring that security standards protect privacy without preventing safe experimentation will be essential to balancing trust and progress. For example, this may include establishing a national baseline upgrade and maintenance schedule for hospital IT systems, expanding secure testing environments (sandboxes) for controlled evaluation of digital and AI tools, and introducing targeted funding mechanisms to support upgrades, particularly for smaller hospitals and primary care providers that face greater resource constraints. The introduction of a national provider accreditation scheme for telehealth services is expected to contribute to more consistent clinical governance and data interoperability. Further updating the Health Information Privacy Code (2020) and clarifying regulatory expectations for high-risk clinical AI, potentially drawing on international frameworks, such as the European Union’s Artificial Intelligence Act (2024), could further strengthen patient protection, transparency and cross-border compatibility while supporting safe innovation (Box 3.2). Legislative reforms under the forthcoming Medical Products Bill are also expected to introduce regulation for certain AI tools classified as Software as a Medical Device.
The European Union (EU) adopted the EU AI Act in 2024, establishing the first comprehensive regulatory framework for AI. The Act applies a four-level risk-based classification, with AI systems used in diagnosis, clinical decision support or care prioritisation designated as “high-risk” and therefore subject to strengthened oversight. These range from “unacceptable risk” systems that are prohibited (e.g., social scoring), through high-risk systems subject to strict safeguards (e.g., clinical AI tools), to “limited-risk” (e.g., chatbots) and “minimal-risk” applications (e.g., spam filters) with lighter requirements.
Key safeguards include representative and traceable training data, clear human oversight requirements to ensure clinicians retain decision-making authority, and continuous post-market monitoring. To enable responsible innovation, the Act also introduces regulatory sandboxes, allowing supervised real-world testing before wider scale-up.
The EU framework provides a structured model for balancing innovation with public trust. In particular, it highlights the importance of clinical oversight, transparent data governance, equity safeguards and continuous monitoring over the full lifecycle of AI-enabled tools. While New Zealand’s regulatory environment aligns with these principles in broad terms, it remains less formalised and less standardised (Table 3.1). Clarifying the classification of high-risk clinical AI, establishing structured post-market monitoring, and developing a nationally coordinated sandbox environment would support safe experimentation while maintaining trust and equity. However, adapting large-scale regulatory models to smaller health systems may require proportional approaches that reflect institutional capacity and resource constraints.
In addition to clarifying permitted data uses, recent research on the secondary use of health data highlights the importance of clear processes and safeguards for enabling data access when it serves well-defined public-interest purposes (OECD, 2025b). Updating the regulatory framework to clarify how, and under what protections, health data can be made available in contexts such as public health emergencies or system-performance monitoring would help support responsible and trusted data use.
Rigid institutional settings can also slow the adoption of digital tools. Workforce-planning rules, such as Care Capacity Demand Management (CCDM), which prescribe minimum staffing ratios to protect patient safety and manage workloads, can limit flexibility when new technologies change task distribution or reduce administrative burden. In practice, providers report limited scope to reassign staff time or redesign workflows, even where digital tools have demonstrated efficiency gains. Introducing more adaptive workforce provisions, for example, allowing temporary adjustments to staffing configurations within defined safety parameters or linking staffing models to observed outcomes rather than fixed ratios, where clinically appropriate, could enable safe experimentation while maintaining accountability and the quality of care.
Liability frameworks for AI-enabled decision support are still evolving internationally. The Health and Disability Commissioner framework regulates clinical professionals rather than technology developers, leaving uncertainty in cases where AI outputs influence or shape clinical decisions. Although New Zealand’s Accident Compensation Corporation scheme limits personal liability for clinical harm, ambiguity persists over how responsibility is shared between clinicians, organisations and vendors when AI recommendations are misinterpreted or followed without adequate review. In practice, this has contributed to cautious adoption, particularly where clinicians feel uncertain about when and how to challenge AI-generated outputs. Introducing clear disclosure requirements for when AI is used in clinical processes, along with guidance on human oversight, shared decision-making and post-implementation monitoring, including clearer processes around reporting AI-related harms and clarifying associated responsibilities, could help reduce uncertainty while more comprehensive regulatory models are developed.
3.3.3. Investing in interoperable infrastructure and scaling-up deployment
Sustained and predictable investment is required to support system-wide digital adoption. Short budgeting cycles and fragmented capital planning have contributed to stop–start progress, with infrastructure upgrades and digital platform deployments often initiated but not completed. These conditions make it challenging to scale successful pilots, achieve interoperability across providers, or build the secure cloud environments needed to support data-driven care. Because digital investment typically requires multi-year returns rather than year-to-year performance gains, annual funding models that prioritise short-term outputs can limit the scale and continuity of progress.
Moving toward multi-year, outcome-based funding frameworks would provide a more stable basis for long-term planning and implementation. While the 10-Year HDIP represents an important step toward more coordinated and sequenced investment, the plan will require consistent governance, transparent prioritisation criteria and alignment between annual budgeting decisions and long-term system objectives. Complementing this, a more consistent cost–benefit assessment approach would help ensure that investment and procurement decisions are guided by long-term value, scalability and system-wide benefit, supporting sustained digital transformation and avoiding repeated stop–start cycles linked to annual funding constraints. Although digital and AI investments may involve upfront public expenditure, particularly in the early stages of implementation, these costs are often temporary and are expected to decline over time. Over the medium to longer term, if well implemented, these investments can improve productivity and contribute to more efficient care delivery, more than offsetting initial spending pressures and supporting better value for money.
Procurement fragmentation also continues to constrain efficiency. Stakeholders noted health service providers often purchase digital tools individually, leading to duplication, inconsistent technical standards and higher per-unit costs than a centralised system. Hospital-level adoption of core systems, such as electronic prescribing, remains limited and many facilities continue to rely on manual or partially digitised workflows. Legacy hospital information systems vary in maturity and interoperability, and capital budgets are often directed toward maintenance rather than system renewal. As a result, digital maturity differs substantially across regions and care settings, weakening continuity of care and the ability to share data efficiently. While recent pooled purchasing initiatives aim to reduce costs for smaller organisations, legacy contractual arrangements and vendor lock-in still restrict system-wide integration.
Multiple patient management systems operate in parallel across primary, community and hospital settings, with limited ability to exchange information. While GP-to-GP transfer systems function for patient enrolment, data often degrade when moving across platforms, and hospitals frequently rely on manual letters to communicate discharge summaries. Adoption of national interoperability standards, such as FHIR, has progressed unevenly, and an estimated 60% of existing records remain in legacy formats. This results in fragmented patient information across hospitals, GPs, pharmacies, and community providers, limiting continuity of care. A legally enforceable national interoperability framework, supported by technical certification requirements for vendors and phased transition timelines, would help ensure that new investments are compatible across the health system and avoid further divergence. Ensuring consistency with widely adopted international standards, where appropriate, would further support interoperability and reduce duplication of effort.
Incentive structures will also shape the pace of adoption. Primary care practices and smaller providers often face resource and capability constraints but currently receive limited direct support or incentives to upgrade systems or participate in shared-care models. International experience shows that targeted subsidies for system upgrades, guaranteed support services during transition, and outcome-based reimbursement models can accelerate uptake, particularly in primary and community care. Structured pilot programmes, accompanied by shared evaluation frameworks, could help identify scalable implementation models while reducing risk for early adopters. Korea’s recent reforms illustrate how coordinated investment, strengthened data infrastructure and practical incentives can accelerate digital and AI adoption at scale (Box 3.3).
Korea has pursued a coordinated, system-wide approach to digital and AI-enabled health, combining sustained investment, structured data infrastructures and incentives to support adoption. An OECD review conducted with Korea in 2021–22 highlighted the country’s early progress in integrating data systems, establishing governance mechanisms and preparing for AI-enabled health applications (OECD, 2022). In the 2026 Budget, health and medical R&D funding will reach around KRW 1.1 trillion (USD 0.8 billion), and investment in medical AI research is planned to increase more than four-fold between 2024 and 2026, from KRW 28.6 billion (USD 22 million) in 2024 to KRW 131.8 billion (USD 100 million). These multi-year commitments are aligned across 11 ministries, led by the Ministry of Health and Welfare, providing consistent priority-setting and reducing fragmentation across research, regulation and data policy.
To enable safe development and adoption, Korea has established national data and AI testbeds that provide developers and clinicians with structured environments for algorithm validation, workflow integration and iterative refinement. As part of its AI-bio strategy, Korea is also developing AI-driven drug discovery platforms, including AI- and robotics-enabled “self-driving labs” that aim to automate experimentation and significantly increase throughput. Complementary data-governance mechanisms—including data review boards, secure analysis environments and standardised data-sharing agreements—offer clear safeguards while lowering uncertainty for both providers and innovators. These infrastructures help reduce evaluation costs, accelerate cycle times and support responsible secondary use of health data.
Large-scale data initiatives incorporate additional incentives. Korea’s medical AI data voucher scheme provides AI SMEs and start-ups with support of up to KRW 400 million (USD 0.3 million) for data preprocessing and analysis, enabling them to work with real clinical datasets from a network of accredited medical data-centric hospitals. This demand–supply matching model lowers entry barriers for innovators by covering data-related costs, while participating institutions have administrative and labour-cost burdens partially offset. Together with tailored engagement for participants, these mechanisms help strengthen public trust, encourage participation and support sustained data quality. Korea has also strengthened institutional incentives for digital adoption through standards-based regulation. From 2025, the second cycle of its EMR certification framework has placed greater emphasis on interoperability and standardised information exchange, helping align vendor products and provider systems with national digital health objectives. In practice, Korea’s earlier shift toward electronic insurance claims—which were processed far more efficiently than paper submissions— also created strong operational incentives for providers to adopt EMRs and maintain alignment with evolving national data standards.
Korea’s approach illustrates how coordinated investment, clear data-use safeguards and structured incentives, supported by strong public–private collaboration, can accelerate the development and scale-up of AI-driven health solutions.
Source: Korea’s Ministry of Health and Welfare.
Scaling up digital health solutions will also depend on the strength of the domestic innovation ecosystem. While the private sector plays a central role in developing new tools and care models, stakeholder consultations revealed that scale-up is constrained in a similar way to many technology driven start-ups in New Zealand by limited access to growth-stage capital (Chapter 4). This constrains the availability of mature, locally developed tools that could be adopted across the public health system. Greater coordination between digital health investment planning and capital-market development strategies, including co-investment mechanisms, procurement pathways that recognise proven domestic products, and structured evaluation environments, could help strengthen the innovation ecosystem and support wider deployment of scalable digital solutions.
New Zealand’s national medicines data infrastructure, including the Pharmaceutical Collection maintained by Health New Zealand, represents an important asset for digital health development. The national medicines formulary (the list of publicly funded medicines and the conditions under which they are prescribed) and medicines utilisation datasets already support evidence-based decision-making and could enable better prescribing if linked more seamlessly with primary and hospital electronic systems. Improving interoperability would strengthen medicines management across the continuum of care and provide a practical early use case for supervised AI adoption. In July 2025, the government encouraged Medsafe and Pharmac, New Zealand’s national medicines purchasing and reimbursement agency, to explore the use of AI and other digital tools to streamline assessment and review processes to improve the timeliness of access to medicines. The government has also asked Pharmac to strengthen performance monitoring and consumer engagement while maintaining value for money, reflecting its strong operational capability and delivery record. This positions medicines management as a credible and well-governed pilot domain for responsible AI deployment as broader system capabilities mature.
3.3.4. Building workforce capability and supporting clinical adoption
Technology adoption depends not only on technical infrastructure but also on workforce capability and organisational culture. In New Zealand, digital uptake remains uneven due to variation in digital confidence across the workforce and limited structured support for change management. Younger clinicians also report few opportunities to develop practical digital and AI competencies, and there is concern that automation may reduce early-career exposure to case-based learning. These dynamics underscore the importance of integrating digital tools in ways that augment clinical judgment and strengthen foundational skill development. Ensuring consistent, career-long access to digital and AI training, from initial education through continuing professional development, will be essential to support confident and sustained adoption. This should also include strengthening capability among supporting functions and emerging hybrid roles that work alongside clinical teams in implementing and using digital and AI-enabled tools. Recent initiatives such as the Centre for Digital Modernisation of Health and its Digital Academy programme aim to strengthen digital skills among both technical specialists and non-technical staff, support digital career pathways, and build system-wide digital literacy across the health workforce.
Building on these workforce capability needs, embedding digital literacy, data-governance awareness and AI competencies across medical, nursing and allied-health education pathways will be essential. At present, structured training on digital health remains limited in both undergraduate and postgraduate curricula, while some OECD countries have begun integrating digital and AI competencies into core medical education frameworks (Box 3.4). In New Zealand, digital and AI content is often delivered informally or as optional modules rather than embedded in core training, leading to variation across institutions and regions. This limits systematic capability development and places reliance on individual interest or local champions. Strengthening curriculum standards and accreditation expectations could help establish a consistent baseline of competencies, while allowing flexibility in how programmes are delivered across universities and clinical training settings.
Several OECD countries have begun to embed digital and data competencies within medical and health professional education (Table 3.2). Common elements across these approaches include: (1) a baseline competency framework defining minimum digital literacy and data governance skills for all clinicians; (2) practical clinical application, including supervised use of digital and AI-assisted tools during training; (3) ethical, legal and governance training, particularly relating to patient data stewardship; and (4) leadership and change capability, recognising the role of clinicians in digital transformation.
Source: Australian Medical Council (AMC); Medical Schools Council (MSC); Association of American Medical Colleges (AAMC).
In Australia, the Australian Medical Council announced the Digital Health Capability Framework in 2021, which defines required competencies across the full training pathway (undergraduate, residency, and specialist), supported by a national guidance for supervised implementation in clinical environments (AMC, 2021). In the United Kingdom, the Medical Schools Council and Health Data Research UK have introduced a national competency framework (2024–25 rollout), which provides structured expectations for digital and data-science skills within undergraduate medical curricula, with emphasis on data governance, informatics and foundational AI literacy (MSC, 2025). In the United States and Canada, curriculum reform has been more decentralised, but leading medical schools have begun to incorporate structured training in clinical AI, data stewardship and responsible use of digital systems (AAMC, 2025). Several other OECD countries, including France and Korea, have also introduced nationwide requirements for digital and AI-related competencies in medical education, supported by structured programmes for practising clinicians.
New Zealand currently lacks a unified digital capability framework, and training in digital and AI competencies is largely optional or locally driven, leading to variation in clinical confidence and adoption across regions and institutions. This suggests that a more coordinated approach could help support consistent capability development and sustained adoption in New Zealand, while recognising that implementation pathways may vary across tertiary providers. Recent initiatives, including the development of a Digital Academy through the Centre for Digital Modernisation of Health and the establishment of a new medical school with a strong focus on digitally enabled learning, indicate growing momentum toward strengthening national capability.
Professional bodies will play an important role in translating capability expectations into practice. Colleges and regulatory councils can support consistent skill development by accrediting digital and AI training modules, recognising digital competencies within continuing professional development (CPD) and revalidation processes, and supporting peer-learning and clinical informatics networks. These mechanisms can strengthen adoption while ensuring that training remains clinically relevant and grounded in professional standards. International approaches, such as those of the UK National Health Service Digital Academy, which combines technical up-skilling with leadership development, and Australia’s national Digital Health Capability Framework, demonstrate how structured support and clear career recognition can encourage sustained engagement. Adapting similar approaches in New Zealand, for example by formalising digital and AI skill recognition within CPD pathways and supporting accredited training modules, could help strengthen capability development and encourage broader uptake.
Clinician engagement will be central to successful adoption in practice. Co-design processes that involve clinicians in tool selection, workflow redesign and evaluation can help ensure that digital solutions are usable and clinically relevant. Recognising digital leadership in career pathways, for example, through formal clinical informatics roles or protected time for digital improvement work, can further strengthen engagement. Strengthening pilot programmes that enable safe experimentation, and iterative refinement can also support effective adoption. For example, feedback from the clinicians noted that the clinician-led development of the Cortex mobile care coordination system in Canterbury illustrates how frontline involvement can improve usability and increase uptake. Iterative refinement based on ward workflow feedback supported smoother task handovers and reduced duplication, contributing to more sustained use.
3.3.5. Promoting equity and inclusion
Digital transformation has the potential to expand access and improve continuity of care, but differences in current uptake patterns could risk reinforcing existing inequities if not addressed. Adoption has progressed fastest in tertiary hospitals (i.e., major specialist referral hospitals) and large urban providers, while smaller regional and community-based services have experienced slower digital expansion. These differences reflect underlying variation in infrastructure, workforce capacity and funding flexibility across the system. Māori and Pacific communities, as well as rural populations, continue to face barriers related to connectivity, affordability, cultural relevance and trust. Without targeted measures to address these structural factors, digital health could widen service gaps rather than close them. Indeed, digital and AI-enabled tools can amplify inequities if foundational access gaps are not addressed (OECD, 2024).
The design of digital and AI-enabled services also affects equity in access and quality of care. If AI tools are deployed primarily as lower-cost substitutes for in-person clinical care, they may risk contributing to a two-tier system in which some populations receive clinician-augmented care while others receive digital-only pathways. There is evidence that AI yields the greatest benefit when used to augment clinical judgment – for example, as an additional reader in cancer screening – with human oversight remaining integral to final decision-making (Houssami and Marinovich, 2023; van Winkel et al., 2025). Embedding AI within comprehensive care models, with clear review and escalation pathways, will therefore be important to maintaining safety, quality and trust.
Addressing underlying access barriers will require approaches that are responsive to community contexts. Māori and Pacific communities experience higher health needs and lower levels of digital access, underscoring the importance of co-designed and culturally safe digital services. Tailored interfaces, bilingual communication tools and community-based digital literacy programmes can strengthen usability and sustained engagement. The Government also supports a range of community-based digital inclusion initiatives aimed at strengthening digital skills and confidence among priority populations, an encouraging step toward reducing access barriers.
Applying data sovereignty principles, such as shared governance, transparent data stewardship and clear benefit-sharing, will also be essential to supporting trust in digital and AI-enabled services. New Zealand has already taken steps in this direction (Fellner et al., 2025). Initiatives, such as Te Mana Raraunga and the Whakamaua Māori Health Action Plan (2020–2025), have called for frameworks for co-governance and culturally grounded data stewardship (Ministry of Health, 2020). The establishment of Iwi Māori Partnership Boards in 2022, alongside measures such as zero-data access for vaccination and test information, reflects ongoing efforts to reduce structural barriers to participation. Sustaining and deepening these partnerships will be important to ensure that digital transformation leads to more equitable outcomes. In Canada, the First Nations Health Authority applies Indigenous data sovereignty principles to the design and governance of digital health services. The First Nations Principles for Health Legislation emphasise community control over data access and use, shared governance arrangements, and benefit-sharing, ensuring that digital tools strengthen rather than displace community-led models of care (FNHC, 2025).
Targeted affordability measures can help mitigate digital exclusion, particularly for communities with higher health needs. Options, such as reducing user fees for telehealth in priority populations, providing data-free mobile access for patient portals, and supporting rural broadband improvements can enhance access for low-income and geographically isolated households. Some of these measures are already in place. For example, the Zero Data initiative provides zero-rated mobile data access for key health websites and patient portals, helping reduce cost-related barriers for mobile users. To ensure value for money, these measures are likely to be most effective when targeted based on need and delivered in partnership with existing digital inclusion initiatives, such as device-access and affordable-connectivity programmes (e.g., Spark NZ), rather than through broad universal subsidies.
Integrating equity impact assessments into the design, implementation and evaluation of digital-health initiatives would also help ensure that benefits are shared across population groups. In OECD countries, few health systems currently monitor whether the benefits of AI-enabled tools are distributed equitably across genders, regions or disadvantaged populations, making it difficult to identify or address emerging disparities (OECD, 2024). Additionally, strengthening channels for patient and public engagement in the development of digital and data policies can help reinforce trust and ensure that digital transformation aligns with community expectations.
References
AAMC (2025), Artificial Intelligence Curricula in U.S. and Canadian Medical Schools: Data Snapshot, Association of American Medical Colleges, Washington DC.
AMC (2021), Digital Health in Medicine Capability Framework, Australian Medical Council, Sydney.
European Parliament (2023), EU AI Act: First Regulation on Artificial Intelligence, European Parliament, Brussels.
Fellner, R., E. Sutherland and K. Vujovic (2025), “Digitalisation of public health: Leading practices in immunisation reporting and respiratory disease surveillance”, OECD Health Working Papers, No. 185, OECD Publishing, Paris.
FNHC (2025), First Nations Principles for Health Legislation, First Nations Health Council, British Columbia.
Global Digital Health Partnership / OECD (2024), Launch of the GDHP/OECD Policy Repository Tool, Global Digital Health Partnership, London.
Great Ormond Street Hospital (GOSH) & Natioinal Health Service (NHS) England (2025), “GOSH-led trial of AI-scribe technology shows transformative benefits for patients and clinicians across London”, 4 September 2025.
Health New Zealand | Te Whatu Ora (2025), Quarterly Performance Report Quarter Three 2024/25, Health New Zealand, Wellington.
Houssami, N. and M. L. Marinovich (2023), “AI for mammography screening: enter evidence from prospective trials”, The Lancet Digital Health, 5(10), e641-e642.
Lauritzen, Andreas D., et al. "Early indicators of the impact of using AI in mammography screening for breast cancer", Radiology 311.3 (2024): e232479.
Marino, A. and L. Lorenzoni (2019), “The impact of technological advancements on health spending: A literature review”, OECD Health Working Papers, No. 113, OECD Publishing, Paris.
MSC (2025), Data Science in the Medical Curriculum – Equipping Medical Students for the Digital Age, Medical Schools Council, London.
Ministry of Health (2020), Whakamaua: Māori Health Action Plan 2020–2025, Ministry of Health, Wellington.
New Zealand Government (2022), Budget 2022: Vote Health – Health System Reform Funding Decisions, New Zealand Treasury, Wellington.
New Zealand Government (2025), “Digitising healthcare to put patients first”, Wellington.
OECD (2017), New Health Technologies: Managing Access, Value and Sustainability, OECD Publishing, Paris.
OECD (2022), Towards an Integrated Health Information System in Korea, OECD Publishing, Paris.
OECD (2023), Health at a Glance 2023: OECD Indicators, OECD Publishing, Paris.
OECD (2024), AI in Health: Huge Potential, Huge Risks, OECD Publishing, Paris.
OECD/European Commission (2024), Health at a Glance: Europe 2024: State of Health in the EU Cycle, OECD Publishing, Paris.
OECD (2025a), Health at a Glance 2025: OECD Indicators, OECD Publishing, Paris.
OECD (2025b), “Facilitating the secondary use of health data for public interest purposes across borders”, OECD Digital Economy Papers, No. 376, OECD Publishing, Paris.
OECD (2026), Scaling Artificial Intelligence in Health, OECD Publishing, Paris.
Policywise (2025), “Long waiting lists in New Zealand’s public health system: Causes and impacts on patients”, Policywise, Mount Maunganui.
Slawomirski, L. et al. (2025), “The economics of diagnostic safety”, OECD Health Working Papers, No. 176, OECD Publishing, Paris.
Sutherland, E. et al. (2023), “Fast-Track on digital security in health”, OECD Health Working Papers, No. 164, OECD Publishing, Paris.
van Winkel, S. L. et al. (2025), “AI as an independent second reader in detection of clinically relevant breast cancers within a population-based screening programme in the Netherlands: a retrospective cohort study”, The Lancet Digital Health, 7(8).