Global State of Patient Safety 2025

Institute of Global Health Innovation, Imperial College London
NIHR North West London Patient Safety Research Collaboration
Patient Safety Watch

John Illingworth, Sophia Batchelor, Ishan Khalsa, Melanie Leis, Peter Howitt, James Titcombe, Mike Durkin, Ara Darzi, Bryony Dean Franklin

A nurse with her right hand on a male patient's shoulder, showing empathy

Foreword

This second report on the global state of patient safety will inevitably be compared to the first, published two years ago. At first glance, the headline findings seem very similar, with Norway retaining top spot in our patient safety ranking and the UK remaining 21st out of 38 Organisation for Economic Co-operation and Development (OECD) countries.

But stability in the rankings should not obscure the improvements taking place beneath the surface. Across the world, inspirational work is happening at local and national levels – work reflected in the case studies we present, and work that is ripe for wider adoption too.

Updated data on indicators such as neonatal mortality and treatable mortality shows continued long-term downward trends. Patient safety worldwide is, therefore, on some measures improving, albeit with the caveats that available data from lower- and middle-income countries (LMICs) remains woefully incomplete.

The report also underlines the scale of the opportunity for improvement in the UK. Our analysis shows that if the UK were to achieve the same treatable mortality rate as Switzerland, around 22,000 lives could be saved every year – equivalent to roughly 60 people every day. This is not an abstract aspiration, but a reminder that meaningful progress is both possible and measurable.

We have strengthened this report’s analysis by expanding the indicators covered by our patient safety dashboard from 89 to 108.  It includes more diverse indicators, such as healthcare workers’ perceptions of the safety culture within their organisation. These measures help us move towards a richer understanding of safe, high-quality care.

Yet, the challenges remain stark. Data limitations and time lags mean our view of global safety is partial. What data we have predates the notable 2025 reductions in health aid spending made by the US administration and others. The data is not sufficiently real time to draw any conclusions on progress made by the current UK government, for instance.  

To supplement the important contribution of drawing together available patient safety data in a meaningful way, we have sought qualitative insights from experts across the world. We therefore present four case studies of countries that are all in the overall top 12 of our ranking.

Australia, Ireland, the Netherlands and Norway all have important perspectives to share. These case studies are combined with insights from leaders in Canada, Morocco, Singapore and Thailand. Whilst their health systems differ in some ways from the NHS, we remain of the view that in patient safety there is more that unites us, than divides us. 

That is why we consolidate our learning into 16 ambitions across the four domains of strategy and governance, implementation and learning, involvement and capability, and data and measurement. Drawn from our research, the ambitions bring together the features of a high-performing patient safety system for all countries to aspire to. We know these ambitions can be achieved, because we have seen evidence of them already being realised somewhere in the world.

We want all countries to learn from the successes and failure of others so that we can improve global patient safety. Now is the time for action, to have a more systematic approach to improving safety, backed up by improved metrics. That is why we will continue to advocate for more measurement and more collaboration, so that healthcare globally is safer.

Professor the Lord Ara Darzi, Director
Institute of Global Health Innovation
Imperial College London 

James Titcombe OBE, CEO
Patient Safety Watch

Professor Ara Darzi and James Titcombe OBE

Professor Ara Darzi and James Titcombe OBE

Professor Ara Darzi and James Titcombe OBE

Executive summary

This report from the Institute of Global Health Innovation at Imperial College London, and Patient Safety Watch, presents The Global State of Patient Safety 2025.

The report is accompanied by our updated patient safety dashboard. Comprising 108 indicators across 209 countries and spanning 25 years, the dashboard provides a single point of access to the latest global patient safety data.

Overall findings

As in 2023, Norway again ranks first out of OECD countries in our patient safety ranking. The UK remains in 21st place out of 38. Few countries showed significant movement since 2023, suggesting that meaningful change in patient safety takes time.

Analysis of the data highlights several global trends and opportunities for improvement. We found that upstream deficiencies and inequities in patient care must be tackled if safety risks are to be reduced later. This report highlights how such a preventative approach to patient safety in areas such as mental health and neonatal care can be adopted.

While some aspects of mental health care have improved in recent years, excess mortality for people with severe mental illness remains a major concern worldwide. Since 2015, excess mortality has risen for people with schizophrenia and bipolar disorder by 21 per cent and 41 per cent respectively.

Research has consistently shown that people with severe mental illness die 10-20 years earlier than the general population, and benefit less from treatments for cardiovascular, respiratory and metabolic diseases. Our analysis suggests that this inequity is worsening.

Since 2000, neonatal mortality rates (deaths of babies under 28 days old) globally have fallen 46 per cent, reflecting continued progress in newborn survival worldwide. Countries such as China, Belarus and Kazakhstan have seen remarkable reductions in neonatal mortality rates, illustrating that progress can be made irrespective of geography and wealth.

The broader conclusions we can draw are, however, hindered by the quality and coverage of the available data. Only around a third of the indicators include data from LMICs, compared to 100 per cent of high-income countries (HICs). Most data represents performance in acute care, leaving notable gaps in our knowledge in primary and community settings. These gaps mean it is impossible to provide a truly global picture of patient safety, and why much of our focus remains on OECD and HICs.

Combined with other data challenges, the patterns presented in this report are strong enough to guide learning and policy discussion, but not precise enough to support detailed performance judgements or causal claims. The value of this data lies, therefore, not in ranking for its own sake, but in revealing where further inquiry and action on patient safety is required.

Insights for the UK

It is our hope that countries use our patient safety dashboard to better understand their performance on patient safety indicators, celebrate progress, and identify opportunities for improvement. One way to visualise these opportunities is by comparing performance to countries performing particularly well on key indicators:

  • Treatable mortality refers to deaths that can be mainly avoided through timely and effective healthcare, in areas such as sepsis and venous thromboembolism. The UK’s most recent data for treatable mortality (deaths that can be mainly avoided through timely and effective healthcare interventions) was 2021. That year, the best performing OECD country on treatable mortality was Switzerland. If the UK had matched the rate of treatable mortality of Switzerland, the UK could have had 22,789 fewer deaths in 2021.
  • The neonatal mortality rate in the UK has fallen since 2000 but plateaued since 2017 while the OECD average rate has continued to fall. If the UK matched the neonatal mortality rate of Japan (which ranks first out of OECD countries for this measure), the UK could have had 1,123 fewer neonatal deaths in 2023.

The leading cause of neonatal mortality in the UK is preterm birth. Since 2003, the UK has consistently performed worse than the OECD average on this measure. Improving care for women at risk of preterm birth is urgently required to reduce neonatal mortality, and reduce the disparity in outcomes for babies of Black and Asian ethnicity, and for those born in the most deprived areas.

While the data can tell us the ‘what’, it cannot answer the ‘how’. To better understand how patient safety systems have improved across the world, we conducted interviews with patient safety leaders from Australia, Canada, Ireland, Morocco, Netherlands, Norway, Singapore and Thailand, and produced detailed case studies for four of these countries.

We found that the countries with well-developed patient safety systems balance compliance-based activities, such as checking adherence to standards, with work to build the wider conditions and staff capabilities for safe care.

As a result, in these systems it is healthcare providers – not regulators or oversight bodies – that are empowered with ultimate responsibility for delivering safe care.

Regulation remains essential when it is expert-led, adaptive, and uses data to identify emerging risks before they lead to harm. Simply adding more rules and regulations, however, is rarely the answer.

Actions for safer care worldwide

The insights from our interviews and data analysis have been consolidated into 16 ambitions for safer systems. Across four domains (strategy and governance, implementation and learning, involvement and capability, and data and measurement), these ambitions describe how countries can strengthen their patient safety systems. They are intended to be used in tandem with the existing Framework for Action of the WHO Patient Safety Action Plan 2021-2030. 

In our 2023 report, we recommended action on three core areas of patient safety. We stand by these recommendations, updating them below to provide further detail on how they can become a reality:

  1. To create a more comprehensive set of global patient safety indicators, we encourage international organisations focused on safety and quality to develop a roadmap to improve data coverage in LMICs. The example of maternal and neonatal safety data shows how collective action can lead to global coverage, further supporting safety improvement efforts.
  2. To support improved adoption of best practice in patient safety, we advise countries to use our patient safety dashboard, international insights, and 16 ambitions to learn from best practice in core aspects of patient safety. These complement emerging resources, including the WHO Global Knowledge Sharing Platform for Patient Safety.
  3. To help ensure patients, families and carers become active partners in the delivery of safe care, we advocate national and international action to address the inequities in safe care identified in this report. Our case studies demonstrate how patients and the public can become part of the solution, helping to design and inform national policies and interventions for safer care.

3. Our approach

A man types on a laptop keyboard. The image is overlaid with transparent medical symbols including a heart, a stethoscope and a hospital.

Summary

To produce this report, we undertook the following initiatives:

  • Updated our patient safety dashboard of global patient safety indicators, and analysed the data for key trends and insights over the past 25 years.
  • Updated our patient safety ranking of OECD countries, based on four core patient safety indicators.
  • Interviewed patient safety leaders in eight countries, and created four case studies, to generate insights about improving safety.
  • Created 16 ambitions to support the future development of patient safety systems.

3.1 Data dashboard

The patient safety dashboard is an interactive platform, displaying data on 108 patient safety indicators in 206 countries, areas and territories. Users can organise and view the data in different ways. For example, they can compare the performance of countries on specific indicators over time or they can view a country overview on key indicators using the latest data. In each tab, we have included an information icon, providing details about how to navigate each section.

We designed the dashboard to provide a consolidated picture of global patient safety. This work began in 2023 and included 89 patient safety indicators in the first version. We selected the indicators after reviewing the relevant literature and public databases, and after talking to experts. We only included data that were publicly available, standardised and collected from multiple countries.

For the 2025 version, we reviewed the original indicators and databases to check for updates, and to introduce or remove  indicators. This resulted in a total of 108 indicators in the new dashboard (see Section 5.1.1 for more information). Instead of including data from all years available, as we did in 2023, this updated dashboard focuses on the period 2000-2024.

The sources for the data are:

Due to limited data coverage in some regions of the world, the data dashboard and our resulting analysis is skewed towards European, OECD and HICs. Further information on these data challenges, and our analysis of key trends and insights, can be found in Section 5.

3.2 Patient safety country ranking

The patient safety country ranking uses a methodology derived from the Commonwealth Fund’s ranking of health systems. Countries are scored based on how their performance in four key patient safety indicators compares with the OECD mean. Each indicator is standardised relative to the OECD distribution. Scores are expressed as distances from the OECD mean in units of the OECD standard deviation. So, a score of 0 represents the OECD average, and +1 represents one standard deviation above.

For each country, the composite score is calculated as the mean of its indicator-level standardised scores, with all indicators weighted equally. Countries are then ranked from highest to lowest composite score, meaning the top country performs furthest above the OECD average. This standardisation allows for fair comparison across indicators that are measured on different scales. Only OECD countries are included for comparison.

In our 2023 report, we noted the challenge of comparing health system performance. The small number of indicators used, data lags and coding differences, mean that our ranking is indicative only. It is intended to highlight areas for further inquiry, not provide definitive judgements on performance.

Despite this, international comparisons remain a helpful way to highlight examples of excellence, and opportunities for improvement. To do so effectively requires a commitment to transparency, and the use of a few, carefully selected indicators to galvanise action – these themes also emerged in our interviews with patient safety leaders.

For consistency, our 2025 ranking uses the same four patient safety measures and data sources (OECD and GBD) used in the 2023 version. They are:

  • Maternal mortality (deaths per 100,000 live births): comprising all causes of maternal deaths.
  • Treatable mortality (deaths per 100,000 population): comprising causes of death that can be mainly avoided through timely and effective health care interventions. For example, sepsis, venous thromboembolism, and adverse events.
  • Adverse effects of medical treatment (deaths per 100,000 population): comprising deaths sustained as the result of undergoing a procedure, treatment, or other exposure to the health-care system. For example, inpatient admission, outpatient facilities, emergency care, or during home treatment.
  • Neonatal disorders (deaths per 100,000 live births): this aggregate indicator comprises five main causes of death and disability – preterm birth complications, neonatal encephalopathy due to birth asphyxia and trauma, neonatal sepsis and other neonatal infections, haemolytic disease and other neonatal jaundice, and other neonatal disorders.

    • These measures were included in our ranking methodology because:

    • They are well-established and understood measures of patient safety.
    • They are relatively well defined, which reduces the likelihood of differences in reporting between countries.
    • They are available for all OECD countries, allowing for comparisons.
    • They are predominantly rates of death, which tend to be more consistently and reliably reported.

    The inclusion of the treatable mortality measure intentionally broadens the scope of what is traditionally considered patient safety. In an England poll conducted in 2024 with YouGov, we found that waiting for urgent care was consistently ranked as the number one patient safety priority for the public and for health and social care workers. In this context, we consider patient safety to include harm experienced while people wait for their care, not only as a result of their care, as captured in the treatable mortality measure. The updated ranking can be found in Section 5.2.

    3.3 Case studies and interviews

    To better understand the development of national patient safety systems, we conducted interviews with 12 patient safety leaders from eight countries. The selection of countries was informed by discussions with experts, a scan of the relevant literature, and the result of our updated country ranking. Interviewees spanned government, regulators, regional and national agencies, and academia.

    We sought to gain first-hand perspectives on the journeys that countries have been on to develop their patient safety systems, and what they have learned through the process. The interview topic guide was informed by findings from our previous reports and aligned to the strategic objectives within the Framework for Action of the WHO Patient Safety Action Plan 2021-2030.

    Interview questions focused on:

    • Strategies for improving patient safety
    • Implementation of strategies and interventions, and methods for measuring impact
    • Methods for involving health workers, patients and the public in safety improvement efforts, and perspectives on their effectiveness
    • Efforts to improve the culture of safety, and perspectives on the leadership attributes that support it.

    After the interviews, we selected four countries as case studies, based on their well-established patient safety systems and their potential for transferrable learning. Those countries were:

    Australia – adapting systems of regulation and building the conditions for safe care
    Ireland – using safety failings to accelerate change and create a system built around openness and transparency
    The Netherlands – improving safety through national improvement programmes and by rethinking the role of regulation
    Norway     – committing to systematic improvement of patient safety using multiple mechanisms of change.

    The other countries represented in our interviews were Canada, Morocco, Singapore and Thailand. Insights from the interviews can be found in Section 6. The full list of interviewees can be found in the Acknowledgements.

    3.4 Ambitions for national patient safety systems

    To consolidate the learning from our interviews and case studies, we created 16 ambitions across four domains to support the strengthening of national patient safety systems. The ambitions are based on the systems and processes that countries with well-developed approaches to patient safety are putting in place.

    They also incorporate ideas from existing frameworks on safety culture, safety measurement, patient involvement and reducing inequalities.

    The ambitions can be used in tandem with the Framework for Action of the WHO Patient Safety Action Plan 2021-2030, which sets out the detailed infrastructure required to continuously improve patient safety at a national level.

    We welcome feedback on how the ambitions can be used by countries in a range of different contexts. Feedback can be sent to the authors here.

    4. Global patient safety movement:
    where are we now?

    A male patient in a hospital bed is shown medical data on a screen by a female doctor

    The challenges to continuously improving patient safety have been well documented, and debated.

    A recent study of the prevalence of harm identified at least one adverse event in 23.6 per cent of hospital admissions.

    Of these adverse events, 22.7 per cent were judged to be preventable and 32.3 per cent were judged to be serious (requiring substantial intervention of prolonged recovery) or worse. Although based on a sample of 11 hospitals, the findings illustrate the ongoing challenges in reducing healthcare-related harm.

    In 2021, the WHO took stock of progress, recognising the low impact of patient safety improvement efforts relative to the resource invested.

    This resulted in the Global Patient Safety Action Plan 2021–2030, which set seven strategic objectives for member states to address (Figure 1).

    In 2024, the WHO reviewed progress against these objectives in its Global Patient Safety Report 2024. The report found that, while many patient safety objectives were well recognised by member states, progress on implementing specific actions towards them was patchy.

    Relevant to the themes identified later in this report, it found:

    • A quarter of countries had made efforts to develop a culture of safety across their facilities and services.
    • A quarter of countries had implemented human factors principles into practice.
    • A quarter of countries reported systemic and proactive management of safety risks.
    • 17 per cent of countries systematically included safety in primary care programmes.
    • While three-quarters of countries had identified patient safety indicators, only a minority had integrated them into health information systems.

    The 2024 report also outlined progress against 10 core patient safety indicators to enable ongoing monitoring of key actions (Figure 2). These indicators measure the progress member states are making on the infrastructure for patient safety, such as national plans, patient safety curricula, and incident reporting and learning systems.

    Despite the gaps in implementation, it is important to recognise the progress that has been made. For example, the WHO reported notable use of patients’ rights charters and open disclosure policies in the event of patient harm. Our case study of Ireland illustrates how approaches to patient engagement and open disclosure can not only be implemented but can become the foundational principles of a patient safety system.

    The WHO report also highlighted many examples of excellence in developing national safety improvement programmes. To support this, our other case studies (Australia, the Netherlands, and Norway) provide accounts of how they are taking systematic and organised approaches to building capability, spreading improvement interventions, and using data transparency for improvement.

    5. Data insights

    A stethoscope on top of printed pie and bar charts

    Summary

    • Our updated patient safety dashboard includes 108 patient safety indicators from 209 countries. There remain, however, substantial data gaps outside of high-income countries.
    • Other challenges include the recency of data, changes in definitions over time, and differences in measurement practices, which affect comparability between countries and over time.
    • Most of the indicators relate to maternity and neonatal health, workforce, and harm caused by treatment. Fewer indicators cover other important aspects of safety, including those directly reported by patients.
    • As in 2023, Norway ranks first in our updated patient safety ranking of OECD countries. The UK remains in 21st place of 38. We encourage use of the ranking to prompt further inquiry and learning, rather than to draw definitive judgements on performance.
    • In 2023 there were approximately 103,000 deaths due to adverse effects of medical treatment worldwide.
    • If the UK had matched the rate of treatable mortality of Switzerland, the UK could have had 22,789 fewer deaths in 2021.
    • Some aspects of mental health care have improved in recent years. However, excess mortality for people with severe mental illness remains a major concern and source of inequity.
    • Average rates of maternal deaths, stillbirths and neonatal deaths continue to fall worldwide. However, the rates and causes of neonatal mortality in the UK, particularly related to preterm births, warrant further investigation and action.
    • If the UK matched the neonatal mortality rate of Japan, the UK could have had 1,123 fewer neonatal deaths in 2023.
    • OECD average waits for nearly all planned procedures fell before and after the pandemic. The UK experiences higher than average waits for more complex procedures, but further investigation is needed due to the lack of consistently reported data.
    • OECD average rates for four of five indicators for surgical complications have fallen since 2009. The UK recorded the highest complication rates for three of five indicators in the most recent data, but further investigation is needed due to the lack of consistently reported data.
    • We conclude that, by addressing certain risk factors and upstream deficiencies in care, harm can be reduced later in key areas of mental health, maternal and neonatal health, elective surgery and post-operative care. Our patient safety dashboard can be used to identify countries doing well in these areas to support improvements.

    5.1 Overview of the data

    This year’s Global State of Patient Safety 2025 Dashboard includes 108 patient safety indicators and data from 209 nations for the years 2000-2024. A subset of 23 core indicators has been selected for more in-depth review in this report.

    5.1.1 Data updates

    Since the publication of our 2023 data dashboard, two measures have been removed by the OECD and are no longer available for analysis. Additional indicators have been marked for deprecation by their respective reporting bodies, meaning that the collection and maintenance of the data for those indicators will be discontinued.

    The two removed measures are:

    • post-operative wound dehiscence (a surgical complication where a wound reopens)
    • in-patient suicide among patients diagnosed with a mental disorder.

    We added 13 measures to our data dashboard, reflecting insights generated through our research. We heard, for example, about the importance of recognising all types of harm, such as loss of dignity and respect (see Netherlands case study), and the increased risk of untreated physical illness for mental health patients (see Norway case study).

    The dashboard also includes new indicators on patient safety culture from the OECD Healthcare Quality and Outcomes Database. They provide data on how health workers perceive the organisational culture around patient safety. The full list of new indicators is set out here:

    Indicators included
    • care providers involving mental health patient in decisions about care and treatment (inpatient care)
    • care providers treating mental health patients with courtesy and respect (community-based care)
    • care providers treating mental health patients with courtesy and respect (inpatient care)
    • excess mortality for patients diagnosed with bipolar disorder
    • excess mortality for patients diagnosed with schizophrenia
    • generalist medical practitioners, per 1,000 inhabitants
    • maternal deaths due to ectopic pregnancy
    • maternal disorders
    • medical tests, treatment or follow-up skipped due to costs
    • midwives, licensed to practise, per 1,000 inhabitants
    • midwives, practising, per 1,000 inhabitants
    • obstetric trauma vaginal delivery with instrument
    • obstetric trauma vaginal delivery without instrument
    • extent to which staff work together as an effective team
    • extent to which staff experience a safe and sustainable workload
    • extent to which staff are treated fairly when they make mistakes
    • extent to which supervisors, managers, or clinical leaders consider staff suggestions for improving patient safety
    • extent to which staff are informed when errors occur
    • extent to which staff speak up if they see something unsafe
    • extent to which mistakes are reported by staff
    • extent to which hospital management shows that patient safety is a top priority
    • extent to which processes are regularly reviewed to prevent repeated mistakes from happening
    • extent to which important patient care information is transferred across hospital units and during shift changes
    • percentage of positive overall ratings on patient safety given by staff
    • percentage of staff that work more than 40 hours a week in the hospital where they are employed

    Several other indicators were adjusted, to align with global definitions and data standards. For the full list of indicators, please visit the patient safety dashboard.

    5.1.2 Data classification

    We organised the data using Avedis Donabedian’s classic approach to measuring quality in healthcare (see Figure 3), and classified the measures as relating to healthcare structures (31.5 per cent of the 108 indicators), processes (31.5 per cent) and outcomes (37 per cent):

    • Structure measures reflect the context in which care is delivered, such as the make-up of the workforce or the availability of hospitals and beds.
    • Process measures reflect the way that the systems and processes work to deliver an outcome for patients, such as whether patients are involved in their care, or how long they wait for treatment.
    • Outcome measures reflect the impact of all activities on patients, such as whether they experience safe care when having a baby, or whether they develop an infection following surgery.

    The indicators were sub-divided into the areas of care listed below (see Figure 4). Indicators relating to the healthcare workforce (30 indicators) were the most common, followed by those relating to maternal and neonatal health (24).

    Some of the indicators included in the dashboard go beyond what is traditionally considered to be directly linked to patient safety, such as the timeliness of care or patient-centredness. These have been included to illustrate how people’s access to, and experience of, care can have safety-critical consequences, or impact on how safe people feel.

    With the addition of the OECD’s 12 new patient safety culture indicators to this year’s report, 21 indicators (19 per cent) used data that was directly reported from patients and health workers. This indicates an increase in awareness of the role that hospital culture has on patient safety. However, data for these indicators is sparse, with only Belgium providing data before 2020 for any of the 12 indicators.

    5.1.3 Data completeness

    Data completeness has improved substantially over the past 25 years, reflecting a worldwide focus on the recording and reporting of patient safety. However, data availability remains unevenly distributed across income groups and regions, with substantial gaps outside high-income settings. Even across OECD countries, there is marked variation (Figure 5).

    All indicators (100 per cent) have data reported from HICs worldwide. Only 81 per cent of indicators are reported on by upper-middle income countries (UMICs), and less than a third of indicators are reported on by lower-middle countries (31 per cent) and low-income countries (29 per cent) . This pattern is consistent with the 2023 analysis.

    Regional coverage also reflects similar disparities. Both Europe and Central Asia, and East Asia and the Pacific regions, have coverage across all indicators (100 per cent). Data reported across indicators for the Middle East and North Africa is also good (90 per cent), followed by North America (88 per cent), and Latin America and the Caribbean (80 per cent). In contrast, data is available for approximately only a third of the indicators for Sub-Saharan Africa (34 per cent) and South Asia (31 per cent).

    These patterns suggest that, while the infrastructure for patient safety data collection has become more established in many regions, sustained investment and capacity building is still needed to build a more complete picture of patient safety globally. This is a particular challenge in fragile, conflict-affected and vulnerable settings, the number of which continues to increase.

    At the country level, indicator coverage also varies. No country has data available for all the included indicators. Italy reports on the highest number of indicators (99 out of 108, representing 92 per cent), followed by the Netherlands (89 per cent). These are followed by Israel, New Zealand, Norway, and Sweden, who each report across at least 80 per cent of the indicators. The UK has data for 74 of the available 108 indicators (69 per cent).

    It is important to note that indicator coverage does not directly reflect data completeness reported at the country level (Figure 6):

    • data completeness refers to the share of all possible data points across indicators and years that contain a reported value for a country
    • indicator coverage refers to how an indicator is ‘covered’ if a country reports data for that indicator in at least one year, regardless of how many years it was reported.

    A country can therefore have high indicator coverage but low completeness if reporting is sporadic across years.

    To investigate this further, we analysed a subset of 23 indicators collected from OECD countries (Figure 7). We selected these indicators to give a diverse range and to allow for comparison across contexts. We use these indicators to analyse key trends and observations in Section 5.3.

    Only six indicators are reported by all 38 OECD countries. The most reported were neonatal preterm births, neonatal mortality rate, and adverse effects of medical treatment (deaths and disability-adjusted life years, DALYs), with 96 per cent data completeness (Figure 8). Only six countries currently report on measures ‘care providers involving mental health patient in decisions about care and treatment’ and ‘treating mental health patients with courtesy and respect’.

    These 23 indicators are sorted into five themes according to their area of care. Across these themes, data completeness varies significantly:

    • Maternal and neonatal health and harm from treatment indicators demonstrate the highest coverage: data is available for 93 per cent of all possible countries and years, reflecting long-standing international efforts to report avoidable harm.
    • Indicators focused on access to treatment and post-operative care have significantly lower coverage: data completeness is 20 per cent, showing considerable variation in reporting frequency and availability.
    • Mental health indicators exhibit the lowest completeness overall: just 14 per cent across the five indicators.

    5.1.4 Data challenges

    Analysis of the latest available data reinforces one of the conclusions in our 2023 report – that there remains a lack of timely, relevant and complete data at a global level. These challenges significantly affect comparability across health systems, which is compounded by the lack of data availability in fragmented health systems.

    We have found that:

    • Data completeness remains a significant challenge: both data coverage and data completeness continue to be clustered in wealthier regions and, for many indicators, updates are still infrequent or delayed. The typical country reports on around 30 indicators of the 108 we have included.
    • The data shows a persistent publication lag: international datasets are updated on an annual or multi-year cycle, meaning that the most recent figures often describe the recent past rather than current conditions. For around a third of the indicators, the most recent figures date back to 2023 or earlier.
    • The available indicators do not provide a full picture of safety: of the 108 indicators included in our analysis, 30 relate to mortality. While this remains a key measure of harm, many other important aspects of safety – such as psychological harm, diagnostic delay, loss of dignity, and access to safe care –remain largely absent from routine reporting.
    • Data updates create challenges with consistency and comparability: apparent rises or falls in the data may reflect a change in what is being measured (and how) rather than a true shift in performance. For small-number events, such as maternal deaths, minor revisions can create large year-to-year swings, particularly for reporting countries with small populations.
    • Differences in measurement practices weaken insights: although lower reported rates of harm may reflect genuine improvement, they can also stem from under-reporting, the loss of information due to rounding, or narrower inclusion criteria. Differences in data encoding processes between providers can also lead to differences in data for the same named indicators provided by different sources.

    Taken together, these challenges set the context in which this data should be interpreted. The patterns presented in the dashboard and this report are strong enough to guide learning and policy discussion, but not precise enough to support fine-grained performance judgments or causal claims. Variation between countries is expected, reflecting context and reporting maturity. The value of this data lies, therefore, not in ranking for its own sake, but in revealing where measurement can be improved, where further enquiry is needed, and where the patient safety community still lacks the data it needs to understand and prevent harm.

    5.2 Patient safety country ranking 2025

    As in 2023, Norway ranks first in our updated patient safety ranking of OECD countries. The UK remains in 21st place out of the 38 countries included (Figure 9). Few countries showed significant movement since 2023 (Figures 10 and 11), with the largest changes being:

    • Germany moved up seven places, from 23rd to 16th
    • Belgium moved down six places, from 20th to 26th
    • Poland moved down six places, from 22nd to 28th
    • New Zealand moved down five places, from 13th to 18th

    It is helpful to consider the following when interpreting the ranking:

    • Higher than average: in 2025, 20 countries performed better than the OECD average. The differences across these countries are marginal – they are all within one standard deviation of the mean. This suggests that performance among the leading countries continues to be closely clustered. The relative position of a country within this group should therefore be considered with caution.
    • Lower than average: the remaining 18 countries perform worse than the OECD average, with two countries (Colombia and Mexico) beyond one standard deviation of the mean. Among the other 16 countries below the OECD average, differences in scores are comparatively small. This suggests that most OECD countries remain concentrated around the average, rather than being separated by large performance gaps. The relative position of a country within this group should also therefore be considered with caution.

    The overall spread of scores has remained stable, with a slight narrowing in 2025 compared to 2023 (standard deviation of 0.55 vs 0.58), despite a marginally wider overall range (–2.32 to 0.74 in 2025 vs –1.97 to 0.74 in 2023). This trend indicates that while the extreme lower tail has extended slightly, most countries have moved closer together around the OECD mean.

    Analysis also showed a high Pearson correlation (r = 0.979) between 2023 and 2025, confirming strong stability in relative positions across the two years. Analysis of adjacent-ranks reinforces this finding and indicates that differences between neighbouring countries are generally small across most of the distribution, with only the lowest-ranking countries showing significant separation from the mean.

    Additionally, among the top 20 OECD nations, the gaps between adjacent ranks have narrowed in 2025 compared to 2023, reflecting a smaller dispersion in performance among leading systems (Figure 11).

    This analysis, combined with the lack of significant movement for most countries over the past two years, highlights that significant change in patient safety takes time. The group of best-performing and worst-performing countries have remained largely the same, with the most notable movements happening for countries in the middle. For the top performing countries, this may indicate established patient safety systems that others would do well to learn from. For the countries towards the bottom, this highlights the need for redoubled efforts to improve patient safety.

    However, it is important to note that this methodology has important limitations:

    • The four included indicators only represent a narrow aspect of health system performance, with overrepresentation of maternal and neonatal care. A more meaningful ranking would look at a broader range of indicators, but currently this is not possible because the data is not consistently reported.
    • The indicators primarily look at rates of deaths. The global patient safety movement has tried to move towards a broader understanding of harm including morbidity and psychological harm, but data across these other areas is still limited and inconsistently reported.
    • Research has shown that countries from a lower socio-demographic context will have lower service provision, and therefore utilisation. This results in lower reported adverse effects of medical treatment. Comparisons among countries of similar levels of development are, therefore, more meaningful.

    The Global Burden of Disease (GBD) Study reported that, in 2023, adverse effects of medical treatment led to approximately 103,000 deaths worldwide (upper estimated limit: 128,596, lower estimated limit 85,269). This is a very slight decrease from 2019, as reported in the Global State of Patient Safety 2023 Report.

    The OECD reported that avoidable mortality from treatable causes comprised 77 deaths per 100,000 population OECD countries in 2023. The UK’s most recent data for treatable mortality was 2021. That year, the best performing OECD country on treatable mortality was Switzerland. If the UK had matched the rate of treatable mortality of Switzerland, the UK could have had 22,789 fewer deaths in 2021.

    5.3 Key trends and observations

    We set out below four notable findings based on analysis of the most recent trends using the selected set of 23 indicators in our updated data dashboard. This is not an exhaustive analysis and does not cover all countries. For comparability purposes, insights are largely drawn from the pool of OECD countries that have reported data for each of the four areas explored.

    These metrics are a subset of the full list of 108 metric included in the dashboard, which can be explored for all countries that report data (not just those in the OECD). We encourage users to interrogate the dashboard for insights relevant to their own context and areas of interest.

    5.3.1 Mental health

    Some aspects of mental health care have improved in recent years. However, excess mortality for people with severe mental illness remains a major concern and source of inequity.

    This year, we have included two patient-reported indicators that provide further insight into aspects of mental health care – being involved in decisions about people’s care and treatment and being treated with courtesy and respect. These are factors that can help people feel emotionally safe, and that they have a sense of control over important care and treatment decisions.

    The countries in the report show an overall improvement in these two indicators since data collection started in 2017. However, data coverage for these indicators is extremely poor, making it difficult to draw definitive conclusions. The continued adoption and reporting of these indicators should be encouraged. Clearer, and more worrying, is the worsening trend in excess mortality for patients diagnosed with two severe mental illnesses – bipolar disorder (Figure 12) and schizophrenia (Figure 13).

    On average, across reporting OECD countries, excess mortality for bipolar disorder rose from 2.7 in 2000 to 3.17 in 2024, representing an increase of 17 per cent. Alarmingly, the rate has been steeper in recent years, with a 21 per cent rise since 2015 alone.

    Of the 14 countries with available data, only five – Denmark, Finland, Lithuania, Norway and Sweden – report a decrease between the first and most recent year of data. In contrast, several countries experienced notable rises. These patterns, taken across the past 25 years, suggest that the increase in excess mortality for bipolar disorder reflects a long-term structural issue.

    The situation is worryingly similar for patients with schizophrenia. The OECD average increased from 3.0 in 2000 to 5.1 in 2024, representing an overall increase of around 70 per cent, and a 41 per cent rise since 2015.

    Research has consistently shown that people with severe mental illness die around 10-20 years earlier than the general population. Moreover, the data suggests that people with severe mental illness benefit less than the general population from broader improvements in healthcare.

    Studies have shown that this reduced life expectancy is not only explained by causes such as suicide, but also by physical illnesses including cardiovascular, respiratory and metabolic diseases. One of our interviewees considered this to be the most pressing patient safety concern in their country (see Norway case study).

    A recent investigation (2025) in England into mental health inpatient care found several gaps in the provision of physical health care for people with severe mental illness. It recommended improving the consistency of physical health checks, boosting emergency responses, and addressing the common mis-attribution of physical symptoms to mental illness.

    5.3.2 Maternal and neonatal health

    Average rates of maternal and neonatal deaths continue to fall worldwide. However, the rates and causes of neonatal mortality in the UK, particularly related to preterm births, warrant further investigation and action.

    In our 2023 report, we found that overall rates of maternal deaths (deaths of women related to their pregnancy) had fallen by nearly 42 per cent in OECD countries, representing a major global achievement. The latest OECD data we analysed shows that, overall, this trend has continued (Figure 14).

    Norway, Australia, Ireland and Iceland have maternal mortality rates significantly lower than the OECD average. Countries such as Denmark, the Netherlands, and Ireland show statistically significant decreasing rates over the past 25 years.

    The UK experienced a significant rise in maternal mortality between 2020 and 2022. This was notably higher than most OECD countries. By 2023, maternal mortality rates in the UK fell to just below the OECD mean. However, this improvement relative to the OECD mean should be interpreted with caution.

    As can be seen in in Figure 14, the OECD average is skewed upward by a small number of outlier countries (Colombia and Mexico) which report maternal mortality rates more than two standard deviations above the OECD mean. These outliers inflate the OECD average and therefore make countries such as the UK appear closer to the mean. When performance is  assessed on a standardised scale, the UK ranks 21st of 38 OECD countries.

    The WHO reports that neonatal mortality rates (deaths of babies under 28 days old) have fallen steadily globally from around 30 per 1,000 live births in 2000 to around 17 per 1,000 in 2023. This reflects continued progress in newborn survival worldwide. The OECD average rate of neonatal mortality has fallen steadily since 2000, even accounting for the COVID-19 pandemic (Figure 15), representing a 46 per cent reduction.

    Our dashboard highlights some remarkable reductions in neonatal mortality rates in many non-OECD countries between 2000 and 2023. For example, rates fell from 21.4 to 2.8 in China, from 5.2 to 0.8 in Belarus, and from 23.3 to 4.2 in Kazakhstan. Improvement across such a diverse group of countries illustrates that progress can be made, irrespective of geography and wealth. However, increases in neonatal mortality rates in a small number of countries such as Botswana and Venezuela, highlight the urgent need for more targeted support.

    Although neonatal mortality rates for the UK have decreased from 3.8 to 2.7 per 1,000 live births over the same period, the UK has plateaued since 2017, lagging the pace of improvement seen in Ireland and Norway, and moving further away from the OECD average. Across OECD countries, Japan has the lowest rate of neonatal mortality. If the UK matched the neonatal mortality rate of Japan, the UK could have had 1,123 fewer neonatal deaths in 2023.

    The global rate of deaths due to preterm births (a baby born before 37 weeks of pregnancy) has also declined markedly, from 17.4 per 1,000 live births in 2000 to 7.7 in 2023. Preterm birth is the leading cause of neonatal mortality in the UK, with 1,212 deaths reported in 2023. Although average rates across OECD countries fell by 53 per cent, the figure for the UK was 29 per cent (Figure 16). The data shows that the UK has consistently seen an above-average rate of preterm births since 2003 and has plateaued since 2019.

    There are many biological, environmental and social causes of preterm birth. Recommended preventative action focuses on three main areas: improving women’s health during pregnancy, universal screening for risk factors, and referral to specialist clinics for women identified as at risk. Evidence presented to a House of Lord’s Committee reported disparities in preterm birth rates and outcomes between different socio-economic and ethnic groups, and a lack of progress to meet the UK’s target of reducing the rate of preterm births to 6 per cent by 2025.

    While many of the recommended preventative actions do not necessarily fall within patient safety, inadequate attention to these factors can place greater strain on maternity and neonatal services already experiencing significant quality and safety concerns.

    5.3.3 Waiting times

    OECD average waits for most planned procedures fell prior to, and following, the COVID-19 pandemic. The UK experiences higher than average waits for more complex procedures, but further investigation is needed due to the lack of consistently reported data.

    In the National State of Patient Safety 2022 Report, we argued that the “the lack of timely and equitable access to care should be considered an urgent patient safety issue”. This followed the challenges of increasing waiting lists for planned care, falls in the diagnosis of some long-term conditions, and significant pressures on emergency services, exacerbated by the COVID-19 pandemic.

    Our dashboard contains indicators on waiting times, from specialist assessment to treatment, for seven procedures. Figures 17-23 show the data trends for available OECD countries and compute OECD average trend lines for three distinct time periods: 2000-2019, 2020-2022, and post 2023. This allows us to look at changes before, during and after the effect of the COVID-19 pandemic.

    We see that OECD average waiting times for these procedures fell between 2000 and 2019, or remained relatively stable in the case of knee replacements. Understandably, following the advent of the pandemic, waiting times for these (usually) planned procedures increased.

    The latest available data suggests a downward trend towards pre-pandemic levels for these procedures, except in the case of coronary artery bypass grafts (a procedure to improve blood flow to the heart using a healthy blood vessel, Figure 20) and trans-urethral prostatectomy (a procedure to remove a section of the prostate gland that is blocking the urethra, Figure 23).

    The UK is broadly in line with the OECD average for waiting times for cataract surgery, hip replacement and knee replacement (Figures 17-19). These procedures are standardised and less likely to result in emergency situations. As NHS capacity has diminished, these procedures have been increasingly performed by private providers – either outsourced by the NHS or self-funded. For example, around half of all hip and knee replacements were performed privately in 2021-22, half of which were funded by the NHS.

    In contrast, the UK is a notable outlier (above or well above –  the OECD average) for waits for coronary artery bypass graft, hysterectomy, trans-luminal coronary angioplasty (a procedure to widen blocked or narrowed coronary arteries using a balloon catheter) and transurethral prostatectomy (Figures 20-23). These procedures remain largely NHS-provided given their greater degree of complexity.

    It is important to note that the UK has not consistently reported data across these indicators, making it impossible to monitor trends over time. This data should, instead, be used to trigger further inquiry, rather than to draw definitive conclusions.

    5.3.4 Surgical safety

    The average rates for of five indicators for surgical complications have fallen since 2009. The UK recorded the highest complication rates for three of five indicators in the most recent data, but further investigation is needed due to the lack of consistently reported data.

    Surgery and anaesthesia have been areas of major focus for patient safety improvement efforts over the past 25 years, particularly due to standardised processes before, during and after surgery. Five of the indicators within the dashboard relate to safety during and after surgery. Since 2009, there were falls in the average rates for the following indicators across countries reporting data to the OECD even accounting for the impact of the COVID-19 pandemic:

    • Deep vein thrombosis (DVT, Figure 24) – a blood clot that forms in a deep vein, often in the leg, which can later become life threatening.
    • Pulmonary embolism – a blockage in the artery in the lungs, often caused by DVT – following hip and knee replacement (Figure 25).
    • Retained surgical item or unretrieved device fragment during surgical hospital admission (Figure 26).
    • Sepsis – a life-threatening reaction to an infection – following abdomino-pelvic surgery, within 30 days of surgery (Figure 27a).

    The average rates of sepsis following abdomino-pelvic surgery during surgical hospital admission (Figure 27b) was the only indicator to show an increase across countries reporting data to the OECD.

    Based on the latest available data points, the UK recorded the highest rates for three indicators (Figures 24, 26 and 27a) and recorded the third highest rate for a fourth indicator (Figure 25) among the countries that supplied data. However, for three of the five indicators, the UK provided only one data point, meaning that the data should be used to trigger further investigation.

    For pulmonary embolism following hip and knee replacement (during surgical hospital admission, Figure 25), the UK experienced an upward trend during and after the Covid-19 pandemic, the causes of which also warrant further investigation.

    5.4 Conclusions from the data

    The data presented in this section offers useful signals of patient safety performance to help determine where improvement efforts should be focused. Although not intended to be a comprehensive analysis, these findings point to an important conclusion: upstream deficiencies in care can manifest as, or exacerbate, patient safety risks later. For example, inattention to the preventative measures during pregnancy that reduce the risk of preterm birth can place greater strain on maternity and neonatal services and can increase the risk of poor outcomes for mothers and babies.

    Based on the four themes identified in Section 5.3, we consider below what implications this could have for future patient safety work:

    • Safety in mental health: to ensure the physical health of people with severe mental health is attended to in a timely and preventative way, to the same standard as for the general population.
    • Maternal and neonatal safety: to take effective action to reduce rates of preterm birth, which would alleviate pressures on strained acute services and, ultimately, reduce neonatal mortality rates.
    • Safety of people waiting for treatment: to proactively monitor people waiting for their procedure for signs of deterioration in order to offset the risks presented by extended waits, particularly for more complex procedures.
    • Surgical and post-operative safety: to adhere to recommended practices around enhanced recovery for people preparing for surgery in order to reduce the risks of complications and poor outcomes.

    One of the key ways to achieve these improvements is to use the data dashboard to identify countries that are performing well and to learn from what they are doing. Using the example of neonatal mortality, we know that preterm birth rates can be reduced through improving women’s health during pregnancy, universal screening for risk factors, and referring women identified as at risk to specialist clinics.

    6. International insights

    Two hands holding a small globe

    Summary

    • Mature patient safety systems balance compliance-based activities, such as checking adherence to standards, with work to build the wider conditions and staff competencies for safe care.
    • The introduction of new rules and regulations should be treated with caution and monitored for unintended consequences – effective and sustainable change is achieved through a combination of levers.
    • Investment in data-led systems improves the identification of emerging safety risks and can support more targeted and proportionate approaches to regulation and monitoring of system performance.
    • Structured and systematic methods should be used to identify, scale and disseminate proven safety interventions, using the learning from the case study countries presented in this report.
    • Involving patients in patient safety must become mainstream: in policy and service design, through action on all types of patient harm, and by targeting improvements right across care pathways.
    • Our 16 ambitions can support effective national action in these areas, in line with the Framework for Action of the WHO Patient Safety Action Plan 2021–2030.

    This section includes four country-level case studies that describe the patient safety systems they have implemented. The countries were selected for their potential for transferable learning. Further information on our approach, including the selection process, is described in Section 3.3. The countries are:

    The case studies are not intended to provide a comprehensive account of patient safety systems in each country. Instead, they offer first-hand and real-world perspectives on the approach that was taken and on the lessons learned.

    Generalisable findings from the case studies can be found in Section 6.5, which includes insights from interviews with patient safety leaders in four other countries: Canada, Morocco, Singapore and Thailand.

    An overview of the healthcare systems in the four case study countries is provided below. The UK numbers are included for reference.

    Healthcare System Overview

    UK

    Australia

    Ireland

    The Netherlands

    Norway

    Population (millions)

    67.6

    24.4

    5.2

    18

    5.5

    Life expectancy (years)

    81

    83

    82

    81

    83

    Health system

    Publicly funded, universal

    Publicly funded, universal

    Public and private

    Compulsory basic private insurance

    Publicly funded, universal

    % GDP spent on health

    11.1

    10.5

    6.7

    11.3

    9.9

    Hospital beds per 1,000 inhabitants

    2.4

    3.8

    2.9

    2.3

    3.3

    GPs per 1,000 inhabitants

    0.8

    1.8

    2.1

    1.8

    1.1

    2025 patient safety ranking

    21

    12

    4

    10

    1

    2023 patient safety ranking

    21

    10

    7

    8

    1

    Country case studies

    6.5 Insights from international perspectives

    Informed by the case studies above, and interviews with 12 patient safety leaders in eight countries, some generalisable lessons and emerging themes on the development of patient safety systems are presented below (and summarised at the beginning of Section 5). Interviewees represented Australia, Canada, Ireland, Morocco, the Netherlands, Norway, Singapore and Thailand.

    6.5.1 The evolution of national patient safety systems

    National patient safety systems have evolved, and continue to evolve, over time. This can be characterised as a shift from a compliance-based approach to one that also builds the wider conditions for safe care. These two approaches do not follow a linear path, nor are they incompatible. Instead, they reflect the growing realisation that the next phase of safety improvement requires a different approach to focusing solely on compliance.

    Many of the patient safety systems began by establishing safety and quality standards. Accreditation, licensing and regulatory models were put in place to assess and monitor healthcare providers’ compliance with the standards. Many of these regimes remain in place, but work is underway to adapt them to keep pace with evolving technology, models of care, and patient expectations.

    In Australia, a major revision of the national standards will reflect these developments. The standards will focus more on outcomes, thereby reducing organisations’ preoccupation with process and compliance tasks. Similarly, SingHealth – a large public healthcare provider in Singapore – has replaced its long association with external accreditation with a commitment to continuous improvement using locally developed and contextualised standards.

    The advantages from further accreditation weren’t high compared to the leverage gained from the initial rounds. And it was creating a lot of stress. At the same time, organisations were going for accreditation, getting the gold seal, rather than being genuinely focused on wanting to improve care.”
    Kok Hian Tan, SingHealth

    Most systems are now paying attention to building the conditions that support safe care. This includes building psychological safety and systems thinking, based on learning from other safety-critical industries. Safer Care Victoria, for example, has shifted its focus in recent years from reactive work – like responding to safety events – to improvement, learning and capability-building work, on the basis that “front end engineering around quality and safety will avoid unnecessary harm and costs down the track”.

    Alongside this shift is the recognition that new levels of safety will not be achieved by adding more rules and regulations. Developments such as the ‘duty of candour’ in Victoria, Australia, and the patients’ rights legislation in Norway have been welcomed. However, the experience of implementing them highlights the risks of fixating on the process rather than the purpose, and adds complexity and confusion for organisations and patients alike.

    The Dutch Health and Youth Care Inspectorate acknowledges the challenge of monitoring compliance with an ever-increasing list of requirements. It is, therefore, increasing its work to build capability in the system, rather than exercising its regulatory powers. This includes, for example, delivering safety science training to inspectors and applying systems thinking to organisations’ safety investigations.

    Norway created the first-ever independent safety investigation body. Ukom performs systemic safety investigations into serious adverse events and safety issues. They do this without attributing blame to individuals, instead they focus on the processes and conditions that contributed to the risk. A similar organisation was later established in England (the HSSIB).

    These examples highlight how proportionate and expert-led models of regulation can support the evolution of patient safety systems. By shifting the balance of work away from punitive action towards building safety capability, and from embracing ideas from Safety-I and Safety-II perspectives, regulators can help to create, and model, psychological safety across the system.

    6.5.2 Making safety improvements a reality

    The example of Norway shows the value of having a clear vision for improving safety, accompanied by simple regulations – “easy to understand, and easy to follow up” – that outline the responsibilities for leaders at all levels to improve safety systematically.

    The publication of an annual white paper made public the progress being made – including a consolidated set of safety indicators – and the priorities for the following year.

    The regulations were accompanied by sustained work on cultural change – “we fall short if we believe that we can improve things only through laws and regulations.”

    Work by Healthcare Excellence Canada reaffirms the importance of using a variety of levers to achieve sustained changes in safety and quality. It highlighted that positive change results from a shift in mindset, the dedication of frontline teams, removing barriers or adding enablers at the policy level, and occasionally making legislative or structural change:

    Almost always it was a combination of all of the above. It wasn't just one thing. […] You need structural levers, not just the commitment of strongly committed and wonderful individuals or individual organisations who are willing to move forward.
    Jennifer Zelmer, Healthcare Excellence Canada

    Important learning has been generated from large-scale safety improvement programmes that can be applied to a wide range of contexts. For example, the Netherlands experience shows how ideas from safety-critical industries (for example, SMSs), social movement theory (for example,. change ambassadors and storytelling), and Safety-II (for example, appreciative inquiry) can be applied to build and sustain engagement on a wide range of patient safety challenges.

    Other countries have also led the way in overcoming challenges in replicating safety improvements. Work in Norway demonstrated how to truly understand the core components of an intervention – in this case, the co-design of cancer pathways with patients and professionals – to replicate and improve upon the successes achieved in another country. Healthcare Excellence Canada has developed tools for teams to assess readiness to both spread and receive interventions, then work with early adopters to understand the core and adaptable elements, before deciding to scale.

    6.5.3 Partnering with patients to develop national patient safety systems

    High-profile safety failings acted as an accelerant to building the necessary infrastructure for some patient safety systems – from establishing a dedicated government office or directorate to initiating improvement programmes or enacting legislation. In the case of Ireland, it led to the principle of open disclosure – being honest with patients and families when harm occurs – becoming an important underpinning principle for its patient safety system.

    Partnering with patients in patient safety can take many forms. In Ireland, it includes an ongoing relationship with patient safety advocates to inform policies and interventions designed to embed a culture of openness. In Norway, it included a commitment to transparency in publishing progress on patient safety in annual reports to its Parliament. In Australia and Canada, it includes working with First Nations communities to better understand and meet the patient safety needs of groups with different cultural and historical contexts.

    Healthcare Excellence Canada has worked with First Nations, Métis and Inuit individuals and groups to understand their views on what safer, high-quality care means. In addition to highlighting the importance of cultural safety, it has led to broadening person-centred care perspectives to be more community-focused, reflecting “not just the individuals receiving care, but the wellness of the community in which those individuals live or with which they identify".

    6.5.4 The future development of national patient safety systems

    The leaders interviewed for this research are engaged in a process of understanding how their work in patient safety should adapt and evolve to meet evermore complex patient needs and system pressures. There is no easy answer or simple fix. However, there is an emerging consensus on where efforts should be focused.

    First, all areas of the system – policy, practice, research and regulation – need to focus on improving safety in the settings where people receive care, many of which still receive little attention from a patient safety perspective. This includes care provided in people’s homes, and in primary and community care settings. It also relates to people’s multiple and long-term care needs in different care settings.

    Addressing patient safety in these contexts will have implications for how accompanying work is defined, measured, improved and evaluated. Critically, it will ensure that harm from healthcare is understood to take many different forms, including dignity and respect, and emotional and psychological harm – not just physical harm. This is patient safety from the patient, family and community perspective.

    Second, many systems are re-imagining the role that regulation plays in  patient safety. Ensuring compliance with fundamental standards will remain its cornerstone, as long as the standards – and approaches to seeking assurance – remain agile in the face of changing expectations and technological developments. System-focused investigations, using safety science principles, should become mainstream.

    However, there are concerns that the introduction of new rules and regulations can inadvertently create confusion and complexity for patients and providers, and an unmanageable workload for regulators. A streamlined approach to regulatory oversight – which makes better use of data to identify emerging risks, not just to react to episodes of harm – remains an ambition for many health systems.

    This desire to make better use of data is the third area of consensus. This relates to existing data, as well as addressing knowledge gaps . Efforts are already focused on collecting more leading indicators of safety, not just past indicators of harm, with patient insight and feedback increasingly central to this.

    For data to be meaningful they need to be used as part of a ‘living’ or ‘learning’ system. This means using them to understand the performance of the system, to identify emerging risks, to improve efforts, and to provide constant feedback in order to close the learning loop.

    The system can have many forms, such as safety and quality management or learning health systems. There is  emerging evidence and case studies on how they are being applied in practice in healthcare. Regardless of the system chosen, the people delivering and receiving care within it must be partners in its development and ongoing use.

    7. Ambitions for national
    patient safety systems

    A nurse sanitises a smiling, male patient's left arm in preparation for drawing blood

    Imperial College London asset library/Dave Guttridge

    Imperial College London asset library/Dave Guttridge

    7.   Ambitions for national patient safety systems

    Our interviews with system leaders, and accompanying case studies, illustrate the range of journeys that countries have been on to improve patient safety. No two journeys are the same. No one system is necessarily better than another. However, the themes identified in Section 6.5 highlight many areas of learning, and how patient safety systems are improving.

    We have combined this learning into 16 ambitions, across four domains of safe care (see figure 30 below), for countries to work towards. The ambitions are based on the systems and processes that countries with well-developed approaches to patient safety are implementing. They also incorporate ideas from existing frameworks on safety culture, safety measurement, patient involvement, and on reducing inequalities.

    These ambitions are for countries to check whether the actions they are taking, and goals they are aspiring to, are in line with best practice. They can be used in tandem with the Framework for Action of the WHO Patient Safety Action Plan 2021-2030, which sets out the detailed infrastructure required to keep improving patient safety at a national level.

    Figure 28. IGHI’s ambitions for national patient safety systems

    Domain

    Ambition

    Strategy and Governance

    A national strategy for patient safety is complemented by the inclusion of patient safety in key workforce, financial and demand planning processes.

    Safety standards focus on the outcomes of care and are periodically updated to reflect changes in patient expectations, service developments, and technological advancements.

    Regulation takes a proportionate and expert-led approach, balancing enforcement work with capability-building activity.

    Capability exists to conduct independent, expert investigations into sources of systemic risk, including those that span care sectors.

    Implementation

    and Learning

    Reactive work in response to safety issues is balanced with proactive work to build the conditions to support safe care.

    Systematic methods to spread and scale improvement interventions are deployed to address prioritised safety challenges.

    Systems are in place to anticipate, monitor, and respond to any unintended consequences of new safety rules and regulations.

    Feedback loops ensure that learning from all types of safety investigations, analysis and improvement work is accessible, and routinely applied system-wide.

    Involvement

    and Capability

    Patients, families and advocates are partners in the design and implementation of safety initiatives and are supported to be active partners in their own care and safety.

    Principles of open disclosure, restorative practice, and just culture are applied routinely following an incident, to prioritise patient and family wellbeing and maximise learning.

    Patient safety science is integrated into education and continuous professional development in the healthcare and healthcare management professions.

    Proactive workforce planning ensures that safe staffing is in place to meet existing needs, and the system is resilient enough to respond to unexpected demand.

    Data and

    Measurement

    There is a commitment to the principle of transparency, through public reporting, to demonstrate progress on patient safety and highlight priority areas.

    A balance of leading and lagging indicators are used to identify emerging risks, to track system performance, and to understand the impact of safety improvement interventions.

    Patient feedback is sought and used routinely to monitor safety along patient pathways, and to monitor for inequities in safe care across protected demographic characteristics.

    Data are fully integrated into learning and improvement cycles, forming part of an appropriate management system for safety and quality.

    8. Conclusions

    Two wooden blocks printed with an image of hands holding a heart with a medical cross on it

    8.   Conclusions

    Three important conclusions can be drawn from our analysis of the data.

    First, it is impossible to present a truly global picture of patient safety given the gaps in the data, particularly from LMICs. While this deficit is concentrated on certain parts of the world, it is our collective responsibility to address it if we wish to see sustained improvements in patient safety.

    Second, the data provide only a narrow, largely hospital-based view of safety. There are still patient safety blind spots in primary, community and long-term care, with no routinely collected data on people’s safety while they experience extended waits for their care.

    Third, we must tackle care deficiencies if we are to prevent patient safety risks later. This report provides examples from mental health, maternal and neonatal care, elective care, and post-operative care, of what action can be taken when adopting this perspective.

    If we view patient safety in this preventative way it can make the task of improving safety appear more daunting. It suggests that safe outcomes are determined by an array of complex factors, requiring input from a range of health and care teams, and in some cases, throughout a person’s life. Despite these challenges, we should see this as an opportunity.

    The report makes clear how patient safety is everyone’s responsibility, and demonstrates how healthcare teams in one part of the system can support safety in another. It fosters multi-disciplinary teamworking and cross-sector collaboration, which we know to improve patient safety. It means that patient safety improvements are developed in line with the latest thinking – that is, from the patient’s perspective, as the management of risk over time, and by using a range of tailored safety strategies.

    Insights from our interviews and case studies highlight some of the key ingredients for effective national action on patient safety. They include investment in data-led approaches as part of a learning or management system, structured implementation of proven safety interventions, and partnerships with patient advocates to design safety systems.

    We heard about the limitations of adding more rules and regulations to improve safety. Sustained change is, instead, achieved through a combination of levers, where providers and commissioners of care are empowered with ultimate responsibility for delivering safe care. This finding supports a conclusion of the recent Dash Review of the patient safety landscape in England.

    To galvanise action in these areas, we have updated and expanded the three recommendations we made in our 2023 report:

    1. To create a more comprehensive set of global patient safety indicators, we encourage international organisations focused on safety and quality to develop a roadmap to improve data coverage in LMICs. The example of maternal and neonatal safety data shows how collective action can lead to global coverage, supporting safety improvement efforts.

    2. To support improved adoption of best practice in patient safety, we advise countries to use our patient safety dashboard, international insights, and ambitions for national patient safety systems to learn from best practice in core aspects of patient safety. These tools complement emerging resources, including the WHO Global Knowledge Sharing Platform for Patient Safety.

    3. To help ensure patients, families and carers become active partners in the delivery of safe care, we advocate national and international action to address the inequities in safe care identified in this report. Our case studies demonstrate how patients and the public can become part of the solution, helping to design and inform national policies and interventions for safer care.

    Acknowledgements and suggested citation

    This report was produced by Imperial College London’s Institute of Global Health Innovation and the charity Patient Safety Watch. It was made possible by the generosity of Patient Safety Watch, founded by the Rt Hon Sir Jeremy Hunt MP, chaired by James Titcombe OBE.

    The authors would like to thank the following people who have provided input and advice in the production of this report: Georgia Butterworth, Jeffrey Braithwaite, Kate de Bienassis, Ezequiel García Elorrio and Roberto Fernández Crespo.

    Special thanks to the system leaders and experts who kindly gave their time to contribute to this work: Anne Duggan, Gillian Giles, Kok Hian Tan, Orla Healy, Bent Høie, Catherine Katz, Ian Leistikow, Piyawan Limpanyalert, Louise McKinlay, Kate O’Flaherty, Anne Marie Weggelaar, Amal Yassine and Jennifer Zelmer.

    Thank you to the team within Imperial College London who supported the development of the report: Dara O’Hare, Emma Parnham, Wiktoria Tunska and Emma Watson.

    The authors received infrastructure support from the NIHR North West London Patient Safety Research Collaboration. Patient Safety Watch would like to thank Joe Kiani for his dedication and international patient safety leadership, and the Masimo Foundation for their generous financial support.

    Suggested citation: Illingworth J, Batchelor S, Khalsa I, Leis M, Howitt P, Titcombe J, Durkin M, Darzi A, Franklin BD. Global State of Patient Safety 2025. Imperial College London (2026).

    TopBuilt with Shorthand