Further supplementary written evidence submitted by the Department of Health and Social Care, Qs 1 - 4 (NLR0073)

Contents

Introduction

Responses to the Committee’s requests for information

Q1.              Further analysis of why the cost per capita of clinical negligence claims in other jurisdictions is lower than in England

Q2.              Estimated proportion of spend on legal costs across different countries

Q3.              How neonatal mortality rates compare between jurisdictions

Q4.              Data on neonatal mortality rates by ethnicity and socioeconomic background, and neonatal disability rates, for this and other countries

Annex: WHO Neonatal Mortality data for OCED countries, 2000 and 2018 data

Introduction

1.      On 1st February 2022 Maria Caulfield MP, Parliamentary Under-Secretary of State (Minister for Patient Safety and Primary Care) at the Department of Health and Social Care, gave evidence to the Committee along with Helen Vernon, Chief Executive of NHS Resolution, and Matthew Style, Director General for NHS Policy and Performance at the Department of Health and Social Care.

2.      This followed written evidence provided in December 2021 and amended international data provided in January 2022.

3.      During the oral evidence session, the Committee requested further evidence on the following:

a)      Analysis on the issue of why we think the cost of clinical negligence claims per capita in other tort systems are substantially lower than in England

b)      Further information to be added to the international data table included in the written evidence (Table A: Clinical negligence claims and costs data across jurisdictions in NLR 0072), including, for those countries:

• Estimated proportion of spend on legal fees
• Neonatal mortality rate and disability rate for other countries.
• Data on ethnicity/background for neonatal mortality and disability rate.

c)      Numbers of bereavement midwives and plans for future training.

d)      Analysis of impact of reduction of legal aid and introduction of CFAs on legal costs.

4.      Responses to requests a) and b) are set out below. Responses to requests c) and d) will follow shortly, as arranged with the Committee. We have done our best to locate all the relevant data and explain how and whether it can be compared across jurisdictions. Where data is not available for or comparable between jurisdictions this has been highlighted.

5.      As part of our additional research, we undertook an online literature review of published statistics for each of the countries referenced. This was in addition to researching data sources from the World Health Organisation and World Bank, and looking at the annual reports or official clinical negligence publications for each of the following organisations:

• NHS Resolution
• NHS Wales Shared Services Partnership
• Scotland’s Clinical Negligence and Other Risks Indemnity Scheme (CNORIS)
• Canadian Medical Protective Association
• Australian Institute of Health and Welfare
• Löf (Swedish Patient Insurer)
• Patienterstatningen (Danish Patient Compensation Association)
• Accident Compensation Corporation (New Zealand)
• Japan Council for Quality Healthcare.

6.      The additional data requested by the Committee has been presented in several tables in this document, to allow clear presentation and explanation of any data gaps or limitations, for each of the areas investigated.

7.      As we stated in our previous evidence, making comparisons between countries is difficult: factors such as differences in healthcare systems, state provision of services, legal systems, coverage of claims data between organisations and various inclusions and exclusions between the different data sources mean that the international comparisons information should be viewed as illustrative; therefore the figures cannot be compared on a straightforward like-for-like basis.

Responses to the Committee’s requests for information

Q1.              Further analysis of why the cost per capita of clinical negligence claims in other jurisdictions is lower than in England

8.      The Committee asked for an analysis of why costs in other tort systems are substantially lower than in England. The data suggests the cost per capita for claims in England (£42.1) and Wales (£29.2) are much higher than in the other countries listed, which range from £4.1 to £18.7.  For completeness, we have also included the other countries discussed in our supplementary evidence.

Context

9.      First of all, the rises we have seen in this country and the various drivers of cost that underly them are an important backdrop to differences in overall cost between this country and other jurisdictions. As set out in the Department’s written evidence:

• Claimant legal costs have increased substantially in recent years and remain high. Between 2006-07 and 2016-17, claimant legal costs rose fivefold from £98million to £496million[i].  Since 2016-17, claimant legal costs have levelled out, reducing slightly to £433million in 2020/21[ii]. Claimant legal costs per claim may have been impacted by the introduction of the reforms under the Legal Aid, Sentencing and Punishment of Offenders Act (LASPO) 2012. However, in 2020/21 they remained at a high level relative to 2006/07, accounting for 20% of annual cash payments for claims[iii].
• Following an earlier rise in claim volume and high claimant legal costs, over the last few years the overall cost increase has largely been driven by compensation awards for the highest-value claims. Most of these high-value claims are in maternity, and compensation awards in these claims have been growing at rates far above inflation. Compensation awards accounted for 73% of cash payments by 2020-21 (including the impact of the Personal Injury Discount Rate (PIDR) change)[iv].
• The Government remains committed to the continuous improvement of patient safety and embedding a learning culture across the NHS. Following high-profile failings, the NHS in England has made real progress towards meeting the maternity safety ambition, which is to halve the 2010 rates of stillbirths, neonatal and maternal deaths, and brain injuries in babies by 2025.
• However, we do not believe any measurable decline in safety is driving the long-term rise in the cost of clinical negligence claims – we have found no evidence to suggest this, and the NAO found this “unlikely”.[v] Similarly, we have not found evidence that differences in health outcomes have any particular relationship to claims patterns between countries.

Differences in per capita costs of claims – overview

10.  As set out in our previous evidence, making comparisons between countries is difficult. We included Table A in our evidence to the Committee to draw attention to the apparent differences between claims costs and volumes between jurisdictions, and to invite further evidence and debate. However, it should be viewed as illustrative.

11.  There are several potential reasons for the differences in overall costs between England and other jurisdictions: 

• As we said in our supplementary evidence, the data sources for some jurisdictions reflect different costs as compared to England.

-          In some cases the data sources cover a lower proportion of the country’s claims (such as Australia and Canada), may include or exclude primary care or secondary care claims, or exclude certain costs (such as Sweden and New Zealand) compared to England. We explain these differences in more detail below.

-          Continued access to the courts: Some administrative schemes such as those in Sweden and Japan offer an alternative to the court route, which we understand has higher barriers to entry than in this country. In those countries, bringing a legal claim is still permitted and the costs of associated legal advice or additional compensation obtained through those legal claims are not shown in the administrative scheme data. One exception is New Zealand, which has disallowed access to the courts for treatment injury claims.

• Other factors include differences in how countries’ legal or compensation systems operate, in any barriers or enablers to access compensation and behavioural factors including propensity to claim. Differences in drivers of cost include:

-          Claim volumes: Compared to the other jurisdictions with a tort system, claim volume per 100,000 population in England appears higher than in Wales, Scotland and Canada; only Australia appears higher. 

-          Schemes with broader eligibility (such as those in Sweden and New Zealand) tend to have higher claim volume than in the countries with tort systems; volumes are substantially higher than those in England even though claim volume has doubled in this country since 2006.

-          Compensation levels: We understand that some or all of those administrative schemes tend to offer lower compensation per claim, and we set out some key differences under each country in paragraph 14 below.  It is possible that other jurisdictions with tort systems may also vary in how they approach compensation awards within their legal frameworks. We have not found data sources comparing compensation levels between different tort systems for like-for-like injuries.

-          Legal costs: Administrative systems such as in Sweden and Japan tend to involve low or minimal legal representation (but see the point below about legal claims in these jurisdictions). It is possible that the other jurisdictions with tort systems have lower legal costs than for England. We set out any available data on legal costs in response to Question 2, however the published data is limited.

• Further differences between England, Scotland and Wales, which we would expect to have the most comparable systems, include:

-          Legal advice arrangements: Between the three countries, funding arrangements for claimant legal fees differ, possibly introducing differences in claim behaviour as well as implications for how data is constructed. There may also be some differences for defendant legal advice as to how much is undertaken “in-house” and/or the fee structure for such advice so the split between defendant legal costs and administrative costs in available data may vary.

-          Potentially lower compensation levels: Lower property costs and average earnings may lead to lower levels for some aspects of compensation in Scotland and Wales than in England.

-          Deductions: Both Wales and Scotland operate a scheme where local health boards pay the first £25,000 of any claim. The costs of this are included within Wales’ total costs of clinical negligence, but not within Scotland’s (see paragraph 8 for detail). However, this exclusion from Scotland’s figures is likely only to account for modest proportion of the difference in costs per capita.

-          Different Periodic Payment Orders (PPOs): Although all three jurisdictions award PPOs routinely for high-value claims, detailed practices between the countries may vary, leading to potential differences in costs.

-          Behavioural factors: As mentioned above, there are also likely to be behavioural factors (such as the propensity to claim) that may also drive differences in claims volumes and, subsequently, costs.

Specific issues in data from different jurisdictions

12.  There are also specific differences in data which may explain some of the difference between cost/capita in England and in other jurisdictions.

13.  Jurisdictions with tort systems:

• Wales[vi] – There may be some differences in the approach to inflation adjustments, as this can be assessed on a case-by-case basis in Wales due to the smaller number of claims.
• Scotland[vii] – as indicated above, £25,000 is deducted from the total cost of each claim when it is paid for by the Scottish Government via the CNORIS scheme. This £25,000 is instead paid for by NHS Scotland Health Boards. Funding routes for claims also differ.
• Canada[viii] – Legal costs are excluded from these figures. Also, the Canadian scheme covers an estimated 90% of claims against doctors, as that is the proportion of doctors covered by the scheme. It is unclear whether nurses and other health practitioners are also covered by this scheme.
• Australia[ix] – These figures represent clinical negligence in private hospitals only, as they exclude claims arising from public hospitals and other public services. Australia operates under a hybrid model for healthcare (a mix between private and public hospitals), where private hospitals account for over 40% of admissions in 2017/18[x], so a significant proportion of claims (those in public hospitals) are omitted from these costs.

14.  Jurisdictions with alternative compensation systems (which replace or sit alongside tort systems):

• Sweden[xi] – Under their avoidable harm administration scheme (Löf, the Swedish Patient Insurer), an administrative body decides on eligibility and compensation for claims, so have much-reduced use of the courts and legal costs as a result. Costs arising from claims that are subsequently taken to court are also not included in the figures (we understand that around 20-25 cases go into the court system per year compared to roughly 18,000 claims that go through the scheme[xii]). Medical care and social care costs are excluded from compensation awards because they are covered entirely by the social security system. In contrast, in England, 51% of total compensation costs are attributed to social care[xiii]. In addition to this, around 75% of loss of income is covered by the social welfare system in Sweden and the rest is covered by the patient insurance[xiv], whereas ‘Loss of Earnings’ accounts for around 7% of total compensation costs in EnglandError! Bookmark not defined.. These provide some reasons why the size of Swedish compensation awards is considerably lower than those in England.
• New Zealand[xv] – These total costs exclude legal costs. There are also various caps on heads of loss, such as impairment payments for lifelong injuries and loss of income, which is awarded at up to 80% of pre-accident income and is capped[xvi]. These costs also do not cover the entirety of the claimant’s medical and/or care costs. Therefore, like Sweden, the size of the compensation awards is considerably lower than those in England as they do not cover the same types of spend.
• Japan has a scheme offering a single rate of fixed compensation (a lump sum of c. £195,000 and annual instalments of c. £8,000 for 20 years[xvii]) for certain cerebral palsy cases only, on a “no-fault” basis, paid until the child is aged 20 when certain state benefits become payable. 

15.  The differences set out above – both in terms of the coverage of costs by the available data sources, and visible differences in the drivers of costs – go some way to explain the difference in the costs presented for different jurisdictions. However, as stated in our previous evidence, we cannot be sure that these differences alone can account for the large variations observed.

16.  As the Minister indicated during the hearing, we do not think it is possible to draw firmer conclusions as to reasons for the differences than the above without further extensive research.

Table A: Clinical negligence claims and costs data across jurisdictions. This is the same table as appeared in our supplementary evidence dated January 2022.

Table A: International Comparisons table previously presented in Supplementary written evidence submitted by the Department of Health and Social Care to the Committee (NLR 0072).

Q2.              Estimated proportion of spend on legal costs across different countries

17.  As noted in the Introduction above, claimant legal costs in England have increased substantially in recent years and remain much higher than in earlier decades. However, there is very little published data available for other countries so only limited comparisons and conclusions can be drawn as to differences and any reasons for them.

18.  In England, total legal costs include both claimant and defendant legal costs. Claimant legal costs accounted for 74% of these in 2020-21. Whilst average claimant legal costs per claim were double the average defendant costs in 2006-07, they have since grown to be quadruple the size of defendant costs in 2019-20. Overall, this proportion has been growing[xviii], although last year it decreased from a high point of 77% in 2019-20. Claimant legal costs are also disproportionately high relative to the damages received by claimants in low-value claims.

19.  A consultation on fixed recoverable costs for these types of claims (up to £25,000 in damages) was published on 31st January 2022. This sets out proposals to address this issue for low-value claims.

20.  There is very limited published data available on legal costs for the countries included in the international comparison table above, aside from these slight references:

• NHS Wales Shared Services Partnership reports that the legal costs claimed by claimant’s solicitors are one of the most worrying areas in clinical negligence, but do not provide proportions of spend[xix].
• In Canada, average legal costs per civil case (those that might lead to compensation for patients) have increased from around C$16,000 to over C$20,000 between 2010 and 2019. There has been an increase of 7.4% in legal-civil costs between 2018 and 2019[xx].

21.  This suggests that there have been increases in legal costs over time in some other jurisdictions, but it is not possible to draw any stronger conclusions or make comparisons. 

22.  It should be noted that differences in systems will impact on legal costs. The compensation schemes in e.g. Sweden and New Zealand will involve much-reduced use of legal representation relative to tort systems. Legal costs may not therefore figure in data for these schemes, or may not be relevant. However, in Sweden (not New Zealand) it is possible for claimants to subsequently take their claims to court; there may be costs of legal representation for such claims and those costs are not included in data produced for the scheme. We understand legal claims happen infrequently – only around 20-25 cases go into the court system per year as compared to c. 18,000 claims that go through the administrative scheme[xxi].

Q3.              How neonatal mortality rates compare between jurisdictions

Neonatal mortality rates in context

23.  The Committee has indicated that it wants to look at neonatal mortality rates, potentially as a proxy for maternity safety, in considering the overall costs of clinical negligence claims.

24.  The NHS in England has made progress on maternity safety. The Committee will recall from its Maternity Safety Inquiry and our written evidence for this Inquiry that in England there has been a 25% reduction in the stillbirth rate since 2010, and a 29% reduction in the neonatal mortality rate for babies born over the 24-week gestational age of viability. Data on these outcomes in 2020 was published by the Office for National Statistics on 17 February 2022 which show that the neonatal mortality rate for babies born over 24-week gestation has decreased by 36% since 2020. It means there are around 750 fewer stillbirths and hundreds more mothers and families going home with a live, healthy baby each year.

25.  In our response to this question we set out 2018 data comparing rates with other countries, as this was the most recent comparable international data. In England, rates have since improved further to 1.3 deaths per 1,000 live births in 2020[xxii]. 

How neonatal mortality rates compare between jurisdictions

26.  The Committee will be aware that the high-value claims associated with around half of total costs of clinical negligence in England are for obstetric brain injury rather than neonatal mortality. Whilst progress in reducing obstetric brain injuries in England has been less consistent than for some other indicators, there has been a decline in the number of settled high-value maternity claims (specifically, obstetric cerebral palsy and brain injury claims) in recent years. There is not yet a source of internationally comparable data for brain injury rates.

27.  Neonatal mortality rates refer to deaths within the first 28 days of life per 1,000 live births. The below table presents World Health Organisation estimated neonatal mortality rates (for the United Kingdom, Australia, Canada, Sweden, Denmark, New Zealand and Japan). There are some data limitations for what is presented below in the table, so the figures may not be fully comparable (see paragraph 34). Additional data sources (ONS and NRS) have been used to show data for England, Wales and Scotland.

28.  Data for OECD member countries for 2018 has been provided in the annex as these are the most comparable countries to those presented in the table below. The neonatal mortality rate amongst OECD countries ranges from 0.87 deaths per 1,000 live births in Japan to 8.68 in Mexico. The United Kingdom’s rate was 2.78, similar to the rates for Switzerland (2.85), Denmark (2.73) and Poland (2.79).

29.  The data presented in the Annex includes data from the year 2000, as well as the 2018 data. All OECD countries saw a reduction in their neonatal mortality rates over this period; the UK reduction was 26%. Between 2000 and 2018, Estonia and Turkey saw the largest reduction (80% and 70%, respectively), while France and Canada saw the smallest reduction (6% and 11%, respectively). 

Neonatal mortality and clinical negligence claims patterns

30.  Attempting to make links between compensation systems and neonatal mortality rates is challenging because a range of factors can lead to different rates for maternity outcomes between countries. These include multiple pregnancy rates, average maternal age, average body-mass index, smoking and substance misuse, and socioeconomic factors

31.  Further, the table below presents a very small sample of countries. Whilst the table represents a sample, we are also limited by the small number of countries operating systems that are different to ours, such as administrative schemes. There is also more we seek to understand about the impact of other factors such as the decline in the birth rate in Japan or changes in scheme eligibility in Japan and New Zealand.

32.  As the data shows, we cannot see any clear associations between neonatal mortality rates and differences in the type of compensation scheme. Even if there were a pattern, it could be misleading to draw any conclusions on causation due to small samples and the number of factors impacting neonatal mortality rates.

33.  As we have explained above, while it is true that there could be no successful claims without negligence, we have yet to see clear evidence of a link between safety data and claims patterns. 

34.  For these reasons we do not believe robust conclusions as to correlations or causations between compensation systems and neonatal mortality rates can be drawn from this data.

Table B: Neonatal mortality rate (per 1,000 live births) in 2018.

35.  The following limitations to the data should be noted:

• The neonatal mortality data is derived from three sources: The UK-wide and international data is from the same source (WHO). The data for England and Wales is from the Office for National Statistics and Scotland is from the National Record of Scotland.
• The figures for all gestational ages are provided, to allow comparability between the sources, although there may be minor variations as pointed out below. The UK figure will be largely driven by England’s rate because its population accounts for 84% of the UK population[xxxiii].
• The World Health Organisation data is gathered from civil registration data but where this coverage is limited, population census and household survey data has also been used. There may therefore be some variation across countries, particularly in countries where there are limits in civil registration data for neonatal deaths or where neonatal definitions have minimum gestational age and birth weight.

Q4.              Data on neonatal mortality rates by ethnicity and socioeconomic background, and neonatal disability rates, for this and other countries

36.  Data on neonatal mortality rates by ethnicity or socioeconomic background are available for England and some data is also available for other countries. However, for the majority of countries this data is not published.

37.  For the countries where data is available, those from the most deprived backgrounds tend to suffer worse outcomes with regards to neonatal mortality rates. This is also the case for ethnic minority groups.

• The World Health Organisation data is gathered from civil registration data but where this coverage is limited, population census and household survey data has also been used. There may therefore be some variation across countries, particularly in countries where there are limits in civil registration data for neonatal deaths or where neonatal definitions have minimum gestational age and birth weight.

38.  The table below presents the disaggregated neonatal mortality rates for the countries that provide data. Some data is provided at the UK level only, rather than at the devolved administration level. Where data is provided, the key limitations are presented in the table as these may mean data for different countries are not directly comparable. For neonatal disability, and despite extensive searching (see note 1), data on neonatal disability rates were not found to be reported for any of the below countries, including in England. This may be due to the time lag in diagnosing disabilities in babies.

39.  On 23rd February the Minister announced plans to establish a Maternity Disparities Taskforce to tackle disparities for mothers and babies and reduce maternal and neonatal deaths by improving access to effective pre-conception and maternity care for women from ethnic minorities and those living in the most deprived areas. The Taskforce will bring together experts from across the health system, government departments and the voluntary sector to address disparities in maternal and neonatal outcomes. The first meeting of the Taskforce is due to take place in early March, and we plan to hold the meetings every two months afterwards to maintain progress with delivery of actions and interventions to tackle disparities in maternal and neonatal outcomes.

Annex: WHO Neonatal Mortality data for OCED countries, 2000 and 2018 data[xxxix]

March 2022