Unfortunately, your browser is too old to work on this website. Please upgrade your browser
Skip to main content
Analysis

Snakes and ladders: how NHS league tables work and how they can be improved

Published 19 November 2026
Time to read clock icon About 18 mins
Authors

Key points

  • League tables have long been a controversial way of assessing NHS performance in England. The latest version – the NHS Oversight Framework – ranks NHS trusts using a range of performance indicators, such as on waiting times and finances. A trust’s ranking influences the level of support, scrutiny and intervention they receive from NHS England. 
  • But there are fundamental problems with the current approach to league tables in the NHS. In this analysis, we look closely at the most recent acute trust results, highlight five of these problems and suggest potential solutions. 
  • Trusts are assessed against different indicators depending on the services they provide. This makes comparisons between organisations difficult and unfair. Grouping similar trusts and assessing them against consistent indicators would provide a more meaningful comparison. 
  • Although all indicators are given equal weight in calculating league table rankings, differences in how indicators are scored cause some to have a disproportionate impact on the results. Explicitly weighting indicators would make the relative importance of different aspects of performance more transparent. 
  • The ‘financial override’ – applied where trusts have a planned or in-year budget deficit or receive deficit support funding – can distort rankings. Trusts can fall substantially down the league table because of relatively small deficits, even where performance in other areas has improved. Removing the override, or assessing financial performance separately, would provide a clearer picture of how trusts are performing. 
  • Annual changes to the methodology used by policymakers make it difficult to judge whether ranking changes reflect genuine improvement or deterioration in performance or just changes in how performance is measured. A smaller, more consistent set of indicators would help address this and allow like-for-like comparisons with previous results. 
  • The statistical approaches used to compare the rankings do not adequately account for differences in the indicators used, potentially over- or understating performance changes. Greater attention to changes in individual indicators and trusts' wider circumstances may provide a more interpretable assessment. 
  • Evidence on the impact of using league tables for NHS organisations is distinctly mixed. Aggregate ratings may improve performance in the areas measured but may also have negative effects and unintended consequences – and their impact ultimately depends on the wider context and policies used to improve performance based on the data. 
  • If league tables are here to stay, the way they are constructed needs to be simpler and more consistent to provide a meaningful picture of performance. The current approach makes it difficult for patients, the public and NHS leaders to understand how trusts are performing. More importantly, the ratings need to be used as a way to support struggling organisations to improve rather than another tool for top-down control and performance management. 
 

Introduction

League tables have long been a controversial tool for monitoring and managing the performance of the NHS in England. The latest version – the NHS Oversight Framework, introduced by the Labour government in 2025 in an attempt to ‘drive up standards’ in the NHS – continues this tradition.  

Evidence points to fundamental challenges in making league tables work in the NHS. Ratings can improve performance in the areas measured but can also have negative effects, such as gaming the system or the distortion of local priorities. League tables are complex to design and risk concealing more than they reveal. Some issues with the latest league tables have been highlighted by the Nuffield Trust.  

Despite these concerns, league tables look set to stay. Ahead of the upcoming publication of the latest rankings for NHS organisations in England, we look closely at the most recent results – focusing on acute trusts – to examine how these tables are currently constructed and where it could be improved. We highlight five problems with the initiative and suggest potential solutions.  

 

How the current acute trust league table is constructed

League tables depend on making a judgement about who is performing best, worst and somewhere in-between. But evaluating the wide range of activities performed by NHS trusts with a single score is challenging at best. The current acute trust league table is a composite index, meaning a trust’s overall score is based on its performance across a range of indicators. These include data on a trust’s finances, waiting times, clinical safety, staff wellbeing, patient experience, innovation and population health. The index is recalculated and published each quarter and the methodology revised yearly. The Q1 2026/27 publication was the first time the methods have been revised. 

Each indicator is scored from 1 (best performance) to 4 (worst performance). However, different approaches are used to determine these scores across indicators. Some indicators are scored based solely on a trust’s rank relative to other trusts, while other indicators assign scores based on whether a trust meets a specified threshold or criterion. Each indicator score is added together and averaged to get a trust’s overall score.  

Trusts are ranked according to their overall scores and then divided into four segments. A trust’s segment is used to inform the level of support, scrutiny and intervention they receive from NHS England. For example, as of Q1 2026/27, all trusts in an ‘intensive recovery programme’ – designed to provide enhanced NHS England oversight and targeted support to help trusts address longstanding financial and operational challenges – are in segment 4 and all trusts involved in the advanced foundation trust programme are in segment 1 or segment 2.  

In the first version of the league tables, trusts were divided into four equally sized segments, with the highest scoring quarter of trusts in segment 1 and the lowest scoring quarter in segment 4. For 2026/27, however, trusts are assigned to segments based on their relation to the average score of all scored organisations from 2025/26. After this, any trusts in segments 1 or 2 meeting criteria for a ‘financial override’ – relating to having a planned budget deficit, being in receipt of deficit support funding or being behind by more than any planned surplus in the year to date – are moved down to segment 3.  

Use of the financial override, coupled with the new segmentation rules in 2026/27, produces a distribution where segment 4 includes around 14% of acute trusts, while segment 3 is disproportionately large (62%), having received financial override trusts from segments 1 and 2, as well as trusts that would have been in segment 4 under previous rules (Figure 1).  

Figure 1

 

Problems and solutions

Problem 1: Trusts are not judged on the same indicators 

The segment a trust falls into is based, in significant part, on how it scores in relation to other trusts. But trusts are scored only on indicators relevant to the services they deliver, which varies considerably from trust to trust.  

For example, Somerset NHS Foundation Trust – which provides acute, community, mental health and learning disability services – is scored on 36 indicators. In contrast, the Royal Marsden NHS Foundation Trust, The Christie NHS Foundation Trust and the Clatterbridge Cancer Centre are each scored on 17 indicators, reflecting their narrower scope of activities as specialist cancer hospitals.  

Most acute trusts are scored on between 26 and 29 indicators, but which indicators are used varies substantially between trusts (see Figure 2). Most trusts in segment 1 – the better performing trusts – are judged on fewer indicators. 

Figure 2

Being scored against more indicators means each indicator counts less towards a trust’s overall score. Both the Royal Marsden and Somerset trusts scored 4 on the percentage of trials set up within 90 days indicator. Removing this indicator causes the average score for Somerset (against 36 indicators) to improve by just 0.04 points (from 2.33 to 2.29), while the Royal Marsden’s score (against 17 indicators) improves by three times as much (0.14 points, from 1.74 to 1.60).  

Combining all acute trusts into a single league table when they are judged on different indicators, with different distributions, is unfair and makes comparisons more difficult. 

How this could be fixed 

There is no way to fairly rank trusts that provide different services and are assessed on different metrics. One approach could be to group trusts according to their scope of services and rank them using consistent indicators for each group – for example, comparing specialist acute trusts with other specialist trusts rather than organisations providing a much wider range of services. But this would ignore the significant differences between the services specialist hospitals provide.  

Trusts delivering a mix of services could be assigned to whichever group best reflects their main activities, and only assessed using a core set of indicators for that group. This would inevitably mean some services provided by these trusts wouldn’t be reflected in their ranking, but that may be a reasonable compromise to ensure organisations are being compared on a broadly consistent basis. Alternatively, a trust could be featured in different league tables for each of the acute, community or ambulance services it provides.

Problem 2: Some indicators are much more influential than others 

Index scores are unweighted – a trust’s overall score is simply the mean of its indicator scores. This may reflect a decision on the part of NHS England to avoid making judgements about whether to attach more or less importance to different indicators when determining a trust’s overall score. But this does not mean all indicators contribute equally to a trust’s overall score – and, in either case, an equal ranking between indicators is in itself a judgement about their relative importance.  

Indicators are scored in a variety of ways (see problem 1). In practice, this means some indicators have scores evenly distributed across values between 1 and 4, while others have scores notably concentrated at either end of the scale. As a result, each indicator exercises different power on the relative overall scores of trusts. This means that when indicator scores are combined, those with more variation between trusts will exert more influence on each trust’s overall score and ranking.  

If all trusts have the same score for an indicator, including it will have little effect on the rankings. If half of trusts score 1 and half score 4, this will create bigger differences between the overall scores of trusts. Of indicators applied to more than 100 trusts, the percentage of patients readmitted within 30 days has the most polarised scores (58% of trusts scored 1, 31% scored 4, and the remaining 11% scored either 2 or 3), while the Summary Hospital-level Mortality Indicator has the least polarised (trusts scored 2 unless statistically above or below expectations, which was uncommon – 84% of trusts scored 2). 

Some indicators are applied to very few trusts. Two indicators only apply to Somerset NHS Foundation Trust (Care Quality Commission community mental health survey satisfaction rate and mean length of stay for adult acute and psychiatric intensive care unit patients), while three only apply to the Isle of Wight NHS Trust (average response time to 999 calls, percentage of all ambulance incidents with an outcome of ‘see and treat’ or ‘hear and treat’ and mean response time to category 2 calls).  

In practice, this could mean that a trust looking to boost its ranking has an incentive to focus on certain indicators rather than aim for improvement across the board.  

A simple way to test which indicators make the most difference to overall rankings is to look at how much trust rankings change if an indicator is removed. We calculated the average ranking change using the sum of the absolute ranking changes divided by the number of trusts. Applying this ‘leave-one-out’ method, without applying the financial override, we find that three indicators have the greatest influence on trust rankings: the percentage of patients readmitted within 30 days of discharge, the combined finance score and the percentage of clinical trials set up within 90 days.  

As shown in Table 1, the average change in ranks from removing an indicator is strongly correlated with the spread of scores for that indicator between trusts (the standard deviation) and the number of trusts that indicator is applied to (that is, how much opportunity there is for the indicator to influence a trust’s score). Differences between trusts are largely the result of variations in how an indicator is scored. Indicator scores calculated based on rankings alone have a more consistent spread than those based on whether a trust falls above or below a threshold. This shows that despite nominally equal weights, some indicators have unequal empirical influence on the final ranking.  

How this could be fixed 

NHS England could regain control over the relative importance of different aspects of trust performance by explicitly weighting indicators. An unweighted index is not a neutral choice – differences in how indicator scores are distributed and how widely certain indicators are used mean that some have considerably more influence over rankings than others. Assigning weights would make these judgements more transparent, allowing the contribution of each indicator to reflect its relative importance. Financial performance, which already substantially influences rankings, could be given greater or lesser weight in this way, rather than relying on the much blunter financial override (see problem 3). 

Table 1

Problem 3: The financial override is a blunt instrument and too strong a factor 

The combined finance indicator is already one of the more powerful in determining the rank of a trust. A simplified version of this metric – the financial override – applies the same principle to trust rankings in an even more powerful way.  

Encouraging good financial planning and management is a legitimate and important policy objective. But assigning trusts to segments based on their average scores and then applying a financial override creates cliff edges where trusts can see large changes in their ranks between quarters despite their scores remaining similar or even improving.  

For example, between Q4 2025/26 and Q1 2026/27, scores worsened by 0.36 points for both West Hertfordshire Teaching Hospitals NHS Trust and the Princess Alexandra Hospital NHS Trust. Yet West Hertfordshire’s ranking improved by 15 places after it moved out of deficit and the financial override was disapplied, while Princess Alexandra, which remains in deficit, fell 32 places. 24 trusts saw their ranking improve despite their overall score worsening. 

The financial override is a blunt instrument applied to trusts in deficit and receiving deficit support, regardless of whether the deficit was planned or due to a poorer than expected in-year financial performance. No distinction is made between trusts with:  

  • no planned deficit and a small negative variance 
  • a positive variance but smaller than their planned deficit 
  • a large planned deficit 
  • a substantial planned deficit and further large negative variance. 

Through application of the financial override, some otherwise high performing trusts dropped down many rankings and two segments as a result of small deficits. Meanwhile, trusts with large deficits already in segment 3 or 4 based on their overall score alone saw no change in their ranking – effectively no punishment for being in deficit. Once a trust is in deficit, the override does not further penalise trusts with additional negative variance, nor does it reward trusts making difficult choices to balance their budgets. 

Conversely, some trusts operating with small surpluses, or that broke even, saw large ranking jumps as all the better performing trusts with deficits slid below them into segment 3. This was most clear for trusts at the bottom of segment 2 with small planned surpluses or that broke even – for example, Royal Devon University Healthcare NHS Foundation Trust moved up 16 places from 48th to 32nd based on its overall score alone as a result of thus far sticking to its plan to break even. 

How this could be fixed 

Perhaps the single most important improvement to the league table system would be to remove the financial override. As currently implemented, it penalises trusts in deficit regardless of the size of that deficit or whether they are making progress to reduce it. This can cancel out hard-won improvements in other areas through a relatively small change in financial position. It can also make changes in rankings difficult to explain and potentially encourage trusts to prioritise financial performance at the expense of patient care.  

One alternative would be to publish separate league tables for trust finances and the quality of services they provide, ensuring both remain prominent without allowing financial performance to obscure the wider picture of how well a trust is doing. 

Problem 4: Changing methods between years prevents like-for-like comparison 

Changes to the indicators used and how segments are defined between Q4 2025/26 to Q1 2026/27 make it harder to discern whether changes in a trust’s ranking are due to performance, how league tables are constructed, or both. In the latest Oversight Framework, the number of indicators acute trusts are scored against has risen from a maximum of 22 in 2025/26 to up to 36.1 As league tables become more complicated, any improvements contained within them become more opaque, and potentially harder for patients and the public to understand.  

New indicators for acute trusts this year include those related to data quality, spending on temporary staffing, the percentage of clinical trials recruiting their first patient within 90 days of approval, the percentage of inpatients referred to and seen by an in-house tobacco treatment service who make a supported attempt to quit. While the overall number of indicators has increased, some have also been removed, including the 2-hour urgent response performance of community providers and the Care Quality Commission safe inspection score.  

Accounting for these changes, we can still examine changes to the rankings between quarters to see which trusts have risen or fallen – though in some cases, these changes seem to result from differences in the methods used rather than trust performance.  

  • The Newcastle Upon Tyne Hospitals NHS Foundation Trust improved 37 places between quarters despite maintaining a very similar score (2.32 in Q4 2025/26, 2.28 in Q1 2026/27, 0.04 points difference) as it slips under the score threshold for segment 2, which means it is ranked ahead of higher scoring trusts in financial deficit.  
  • Dartford and Gravesham NHS Trust was one of the big movers between the old method and the new, with its score improving from 2.47 to 2.17, taking it up 49 places from 90 to 41, though remaining in segment 3 due to its financial deficit.  
  • Three trusts improved by two segments – West Hertfordshire Teaching Hospitals NHS Trust, Frimley Health NHS Foundation Trust and Queen Victoria Hospital NHS Foundation Trust. In each case, these trusts would have been in segment 1 in Q4 2025/26 but for the financial override, and moved out of financial deficit while maintaining a score sufficient to enter segment 1.  
  • Two trusts fell from segment 1 to segment 3 (London North West University Healthcare NHS Trust (1.84 to 2.43) and Barts Health NHS Trust (1.99 to 2.6)). Though the deterioration in scores would be enough to move both trusts into segment 3, Barts Health NHS Trust also saw the financial override applied again, having been lifted in Q4 2025/26.  
  • Of the 26 trusts that improved by one segment, 19 (73%) were the result of trusts moving from segment 4 to segment 3, with only four trusts moving in the opposite direction. If the previous quartile-based approach to segment assignment had been applied, another 15 trusts would have found themselves in segment 4, eight of which had improved segment from Q4 2025/26 to Q1 2026/27. 
  • 12 trusts in financial deficit in Q4 2025/26 are no longer in deficit in Q4 2026/27, while four trusts entered into financial deficit over the same period. Financial status for 118 trusts did not change – 83 that were in financial deficit stayed in financial deficit, while 35 not in financial deficit remained so.

The impact of methodological changes is particularly apparent among the best performing trusts. The top three trusts in Q4 2025/26 – Royal Papworth Hospital NHS Foundation Trust, The Christie and the Royal Marsden – saw their scores worsen by around 0.5 points in Q1 2026/27. For the Royal Marsden and The Christie – specialist cancer hospitals – the worsening scores can be largely attributed to poor performance in only three new indicators (the previously mentioned clinical trials and smoking cessation metrics, and relative cost difference – a new measure of productivity).  

The 2025/26 Oversight Framework measured productivity using an implied productivity indicator (a measure of year-on-year change in the health care activity delivered relative to the resources used). This has been replaced by a relative cost difference indicator in Q1 2026/27. While both metrics are intended as measures of productivity, there is very little relationship between how trusts score on one versus another. For The Christie and the Royal Marsden, the impact of the change of measure is substantial – without it and two other newly introduced indicators covering smoking cessation programmes and clinical trial recruitment, their scores would have remained largely stable from quarter to quarter. Their change in scores therefore seems driven by methodology rather than performance. 

Taken together, these examples show how changes to methodology can make it difficult to judge whether trusts are genuinely improving or deteriorating over time. 

How this could be fixed 

A smaller, more consistent set of indicators is needed to make the league tables meaningful – focusing on those that best reflect the most important aspects of trust performance. Before introducing or replacing indicators, NHS England should assess whether they adequately measure the intended aspects of performance and test their impact on different types of trusts, particularly where differences in services may make comparisons less meaningful. 

Where changes are made, previous results should be recalculated using the new methodology wherever possible (or current results recalculated using the old methodology and indicators). This would help distinguish genuine changes in performance from results impacted by revised methods. Where like-for-like comparisons cannot be made, this should be made clear. 

Comparisons should also take greater account of the wider context in which trusts operate. This includes the populations they serve and their infrastructure, geography and local economies. Comparing trusts with similar peers may help account for some of these differences in rankings, although this will not provide a complete solution. 

Problem 5: Statistical comparisons may give a misleading picture of progress 

It is challenging but important to determine whether changes in scores between quarters are meaningful or not. The Oversight Framework uses a statistical tool called a funnel plot to assess whether trusts have genuinely improved their scores from one quarter to the next. Funnel plots are commonly used to compare hospital performance. For example, they can help identify whether a hospital has an unusually high rate of complications following surgery, taking account of the number of patients treated. The aim is to distinguish meaningful differences in performance from normal variation.  

The 2025/26 Oversight Framework used a similar approach to assess changes in trusts' overall scores between quarters (Figure 4). Though not a feature of the Q1 2026/27 league tables, the segmentation summary for 2026/27 indicates funnel plots may return in Q2 when comparable data are available.  

Figure 3

There are two important problems with applying funnel plots in this way. 

  • Trusts are being compared using different measures. As we explored in problem 1, in Q4 2025/26, some trusts were assessed using just eight indicators, while others had 22. The range has now increased to between 17 and 36 indicators for 2026/27. These indicators cover different types of services, including acute hospitals, community services and mental health care. Some measures may be easier to improve between quarters than others. Applying the same statistical thresholds to trusts assessed on different sets of indicators may therefore produce unfair comparisons. 
  • Some trusts cannot achieve the improvement the plot expects. Scores cannot fall below 1 (the best possible score), so trusts that already have good scores have less room to improve. All three trusts assessed on eight indicators in Q4 2025/26 scored less than 1.70 in Q3, yet the plot sets a threshold of around 0.7 points for identifying particularly strong improvement. These trusts could never meet that threshold, however well they performed. 

Similar concerns apply to the use of confidence intervals published alongside trusts' scores and ranks. These are calculated using 'bootstrapping', repeatedly sampling the indicators used to assess each trust to estimate a range of possible scores and ranks. However, given that the indicators measure different aspects of performance and are not necessarily comparable, it is unclear how useful these confidence intervals are in assessing uncertainty about a trust's actual performance or relative position. They can be very broad – Norfolk and Norwich University Hospitals NHS Foundation Trust’s ‘likely range of rank’ spans all the way from 25th to 126th –  and so do little to resolve whether differences between trusts are meaningful. 

As a result, these methods may give an unjustified impression of statistical confidence. They risk praising or criticising trusts for score changes without properly accounting for their starting positions or the different measures used to assess them. Rather than provide a fair assessment of progress, the approach could result in misleading conclusions about which trusts are genuinely improving. 

How this could be fixed 

As the league tables are the result of using varying indicators to assess diverse organisations, any efforts to statistically examine differences between trusts, or changes over time, are fraught with difficulty. Rather than attempting to determine whether changes in a trust's overall score between quarters are statistically significant, the Oversight Framework should focus on changes in the underlying measures of performance. Looking at how performance against individual indicators has changed and taking account of a trust’s wider circumstances and starting position would provide a more useful basis for monitoring and managing performance. Funnel plots may be a helpful tool for NHS England to avoid inappropriate focus on trivial changes over time, but may not identify trusts making meaningful improvement within the bounds of statistical significance. The same caution should apply when using rankings or segments to determine the level of oversight or support trusts receive. 

 

Conclusions

When introduced, the new trust league tables were heralded by Wes Streeting, then Secretary of State for Health and Social Care, as a way for ‘patients and taxpayers… to know how their local NHS services are doing compared to the rest of the country’. Looking closely at the numbers, it is hard to say this has been achieved. 

Instead, trusts are being judged by changing and varying indicators, with shifting rules for who falls into different performance segments. The use of a blunt financial override is causing some trusts to fall down the league table despite improvement in care, while other trusts are climbing the rankings through financial restraint while overall performance declines. Patients, the public, health care staff and trust leaders may, understandably, be left wondering what it all means.  

League tables are not only problematic as a tool for the communication of NHS performance to the public – their growing use to determine trust oversight, autonomy and access to opportunities such as advanced foundation trust status raises the stakes further.  

The arrival of a new secretary of state and new prime minister has coincided with some rethinking of health and social care policy. For example, the government has reversed course on its controversial decision to scrap Healthwatch, while boosting the ambition of its plans to reform social care. This could also offer an opportunity to reflect on the approach to overseeing and managing the performance of NHS organisations – including the design of, or need for, league tables. The fundamental challenges with composite indicators to judge NHS organisations are not going away. 

If league tables are here to stay, greater consistency and simplicity would make them more useful to those inside and outside the NHS trying to understand how local services are performing. But better rankings alone will not improve care – what matters is whether the approach to oversight helps identify the challenges trusts face and provides them with the support and resources needed to address them. This will depend on a broader change in approach to overseeing and managing NHS performance – away from an overreliance on top-down control and performance management and towards a greater focus on developing the skills and capabilities needed for organisations to improve care.  

Further reading

Related content