1. Main points

  • In 2013, approximately 23% (114,740 out of 506,790) of all deaths registered in England and Wales were from causes considered avoidable through good quality healthcare or wider public health interventions
  • In England and Wales, the majority (at least 60%) of potentially avoidable deaths in each year between 2001 and 2013 were among males
  • Avoidable mortality rates for cardiovascular disease (disease of the heart or blood vessels) fell by 52% between 2001 and 2013, the greatest decrease by any broad cause group. This decrease meant that neoplasms (cancer and non-cancerous abnormal tissue growths), which fell by 17%, replaced cardiovascular disease in 2007 as the cause group responsible for the majority of avoidable deaths
  • Breaking the causes into more detail, Ischaemic (coronary) heart disease was the most common individual cause of avoidable death for all persons, accounting for 17% (19,951 out of 114,740) of these deaths in 2013
  • When the sexes were examined separately, Ischaemic heart disease was the most common cause of avoidable deaths for males, but for females it was lung cancer
  • In 2013, Ischaemic heart disease accounted for 22% (15,078 out of 69,245) of avoidable male deaths, while lung cancer accounted for 15% (6,823 out of 45,495) of avoidable female deaths
  • In both England and Wales, avoidable mortality rates were higher for males than females. While the gap between the sexes narrowed in England, it increased in Wales
  • Avoidable mortality rates fell significantly in all regions between 2001 and 2013. The greatest decrease was in London, where rates for males and females fell by 38% and 36% respectively. The smallest decrease was in the South West, where rates for males and females fell by 27% and 23% respectively
Back to table of contents

2. Definitions

We present mortality figures for causes of death that are considered avoidable in the presence of timely and effective healthcare or public health interventions (avoidable mortality). Figures are presented for England and Wales and the regions of England for the period 2001 to 2013. While a particular condition can be considered to be avoidable, this doesn’t mean that every death from that condition could be prevented. This is because factors, such as the lifestyle and age of the patient, the extent of disease progression at diagnosis or the existence of other medical conditions, are not taken into account when making a list of causes.

For most of the causes of death included there is an upper age limit of 74 years. This is because deaths at older ages are often difficult to attribute definitively to a single underlying cause, and the chances of death are more affected by coexisting medical conditions and other factors. A full list of the causes of death included in the definition, ICD-10 codes and age limits are included in the reference tables.

We also present trends in mortality by causes considered preventable (preventable mortality) or amenable to healthcare (amenable mortality), which are subsets of total avoidable mortality.

Amenable mortality:

  • a death is amenable (treatable) if, in the light of medical knowledge and technology at the time of death, all or most deaths from that cause (subject to age limits if appropriate) could be avoided through good quality healthcare

Preventable mortality:

  • a death is preventable if, in the light of understanding of the determinants of health at time of death, all or most deaths from that cause (subject to age limits if appropriate) could be avoided by public health interventions in the broadest sense

Avoidable mortality:

  • avoidable deaths are all those defined as preventable, amenable or both, where each death is counted only once; where a cause of death is both preventable and amenable, all deaths from that cause are counted in both categories when they are presented separately

We published a definition of avoidable mortality in 2011, following a period of consultation with statistics users, academics and experts. We received 20 responses to the consultation and a summary of these responses (174.8 Kb Pdf) was published on our website in August 2011. The final definition of avoidable mortality and the list of causes considered to be avoidable were published in the ‘Definition of Avoidable Mortality (306.9 Kb Pdf)’ document. This definition was used to produce an indicator of potentially avoidable deaths in England and Wales, and the first of a series of annual bulletins was published on 15 May 2012.

The list of causes considered avoidable, along with the associated age limits, will be reviewed every three years. This means that any cause of death which has been excluded from the current list due to concerns around the extent to which death can be avoided will be reassessed and may be included into future lists. We are now seeking your views to assist with this process. You can contribute to the consultation by answering the questions in the consultation document.

Back to table of contents

7. Background

It is widely accepted that the contribution of healthcare to improvements in population health ought to be quantified. Avoidable mortality, which is based on the concept that premature deaths from certain conditions should be rare, and ideally should not occur in the presence of timely and effective healthcare, is used as an indicator to measure this contribution.

Although avoidable mortality has been researched for the last 3 decades, there is little consensus among researchers about how to define it. According to Kossarova et al., (2009), the concept of “avoidability” dates back to the early twentieth century, where confidential enquiries were made into maternal deaths in an attempt to identify improvements.

The concept of avoidable mortality was first introduced by Rutstein et al. in the 1970s. Rutstein argued that in order to develop effective indicators of healthcare a number of disease lists should be drawn up, which should not, or should only infrequently, give rise to death or disability (Rutstein et al., 1976). Subsequently, several papers have reported on regional variation from conditions avoidable with medical intervention.

Rutstein also noted that the list of conditions considered to be avoidable would need to be updated in light of improvements in medical knowledge and practice, as well as social and environmental changes. As a result, their original lists were revised in 1980 to take these and the move to the Ninth Revision of the International Classification of Diseases (ICD-9) into account.

In England and Wales, Charlton et al., (1983) were the first to examine avoidable mortality. They investigated the geographic variations in mortality between 1974 and 1978 using 14 amenable causes of death selected from Rutstein’s list. As with several researchers after them, they excluded conditions such as lung cancer, whose avoidance was considered to be outside the scope of medical care. They found that even after adjusting for social factors, substantial variations in avoidable deaths remained and they urged future studies to examine this further in relation to health-service inputs.

Following Charlton’s work, an attempt was made to compile an atlas on avoidable mortality in the European Community (EC).The conditions included in this atlas were meant to provide warning signals of potential shortcomings in healthcare delivery. It also provided conditions for which a proportion of deaths can be prevented (Holland, 1997, cited by Kossarova et al., (2009).

The atlas also provided a basis for the more recent concept of avoidable mortality, which differentiates between conditions amenable to healthcare (treatable) and those preventable through wider public health policies. Some of the more recent lists of avoidable causes of death include those produced by Nolte and McKee (2004) and Page et al.,(2006). These lists were used by us as the basis for the definition of avoidable mortality. They have been amended and updated to make them more relevant to the UK and to reflect recent developments in health public policy.

Back to table of contents

8. Limitations

One of the main difficulties in producing an indicator of avoidable mortality is the selection of the causes of death for inclusion. While a particular condition can be considered to be avoidable, this doesn’t mean that every death from that condition could be prevented. This is because factors such as the age of the patient, the extent of disease progression at diagnosis or the existence of other medical conditions are not taken into account when compiling a list of causes.

Several studies have shown that deaths from causes amenable to healthcare are declining at a much faster rate than those from non-amenable causes, and that this decline has coincided with the introduction of specific improvements in healthcare. According to Nolte and McKee (2004) these studies are largely focused on quantitative measures, such as healthcare expenditure and the number of healthcare professionals, not necessarily measures of the quality of healthcare systems. They therefore argued that the lack of demonstrable association between avoidable mortality and healthcare resources is, in fact, not entirely surprising. Furthermore, there is likely to be a substantial time lag between change in resources, the introduction of a healthcare innovation or public health policy and a corresponding reduction in mortality. As a result, improvements in the healthcare system may not necessarily be evident from mortality figures in the short or medium term.

In a review of published work on amenable mortality, Mackenbach et al., (1990) noted that geographical variations were strongly linked to socioeconomic factors, which may reflect the differences in timely access to healthcare. Geographical variations may also simply be a result of random variations in disease incidence. For example, if there was a sudden increase in the incidence of a particular condition, and consequently an increase in the mortality rate for this condition, this might be mistakenly interpreted as a decrease in the quality of healthcare.

Note that avoidable mortality was not intended to serve as a definitive source of evidence of differences in effectiveness of healthcare systems. It was designed to highlight areas of potential weaknesses in healthcare that could benefit from further in-depth investigation (Nolte and McKee, 2004). Therefore, the statistics provided in this bulletin should help in assessing the quality and performance of healthcare, as well as wider public health policies. However, due to the limitations described, a degree of caution is required when interpreting the data.

Back to table of contents

9. ICD-10 coding changes implemented in 2011

In 2011 we updated the software used for cause of death coding from ICD-10 version 2001.2, to version 2010. Following this we carried out a bridge study to help users understand the likely impact of this change on mortality statistics for England and Wales. The main changes in the ICD-10 version 2010 are amendments to the rules used in selecting the underlying cause of death. Overall, the impact of these changes is small, although some cause groups are affected more than others.

Many of the conditions affected by the software version change are not included in the avoidable mortality definition. In the majority of cases where conditions are affected, deaths previously coded to one condition are now coded to another condition also listed in causes of avoidable deaths. Therefore, the coding change will have had little impact on the summary avoidable mortality figures published since 2011. For example, analysis presented in the 2011 drug-related deaths bulletin showed that the number of deaths coded as illicit drug use disorders (ICD-10 codes F11–F16 and F18–F19) decreased by 84% per cent in v2010, compared with v2001.2. However, these deaths were allocated to accidental poisonings by drugs (ICD-10 code X40–X44), which is also a cause of avoidable deaths. The impact of coding changes may be more pronounced for cause groups.

Back to table of contents

10. Methodological changes affecting age-standardised rates

The 2013 avoidable mortality indicators presented in this bulletin are based on the 2013 European Standard Population (ESP) introduced across our mortality outputs in 2014. Previously published indicators for 2001 to 2012, produced using the 1976 ESP, have been revised and published alongside data for 2013.

The difference between death rates based on the old and new ESP is purely methodological and doesn’t indicate an actual increase in previously published numbers of deaths or death rates. Further information on the change in ESP is available in our report examining the impact of the change in ESP on mortality data.

In future we intend to publish all age-standardised rates using the full 2013 ESP (with an upper age limit of 95+). However, until official population denominators are available for the oldest age group in the 2013 ESP, crude rates will continue to be standardised using the ‘abridged’ 2013 ESP, with an upper age limit of 90+. Our report found no significant difference between rates based on the upper age limits of 90+ and 95+.

Methods

Age-standardised rates were calculated using the number of potentially avoidable deaths registered in each year as the numerator and the mid-year population estimate for that year as the denominator. These rates were calculated for all avoidable, preventable and amenable deaths, as well as cause groups within these categories. Although age limits were set for some cause groups, rates were calculated using persons of all ages as the denominator. This method was adopted as the entire population is at risk of mortality due to causes considered to be avoidable at an aggregate level and also to allow data by cause group, amenable and preventable categories to be presented on a comparable basis.

Potential years of life lost (PYLL) is a measure of the potential number of years lost when a person dies prematurely from any cause. The basic concept underpinning PYLL is that deaths at younger ages are weighted more heavily than those at older ages. The advantage in doing this is that deaths at younger ages may be perceived to be of less importance if cause specific death rates alone were used to highlight the burden of disease and injury, as conditions such as cancer and heart disease often occur at older ages and have relatively high mortality rates.

In this bulletin, PYLLs are standardised using the 2013 ESP and referred to as standardised years of life lost (SYLL) for clarity. These rates represent the potential years of life lost if the population of England and Wales had the same population structure as the 2013 ESP. SYLL rates are presented as years of life lost per 100,000 population.

Formula for calculating SYLL

PYLL is calculated as the sum of the mortality rate in each age group, weighted by the potential number of years of life lost as indicated by the remaining period life expectancy for each age group. To calculate the SYLL, this is then standardised using the 2013 ESP.

Notes:

  1. Where:
    i is the age group (<1, 1-4, 5-9, 10-14….85-89, 90+)
    di is the number of deaths in age group i
    ai is the weight, or average age-specific period life expectancy in age group i for a given year
    ni is the population in age group i
    wi is the number of individuals in the standard population in age group i

Modifications to standard error and confidence interval calculations

The mortality data in this release are not subject to sampling variation, as they were not drawn from a sample. Nevertheless, they may be affected by random variation, particularly where the number of deaths or probability of dying is small. To help assess the variability in the rates, they have been presented alongside 95% confidence intervals (CI’s).

Traditionally, an approximation method is used to calculate the variance of age-standardised rates. The age-standardised is a weighted sum of the age-specific death rates where the age-specific weights represent the relative age distribution of the ESP. Therefore, its variance is now calculated as the weighted sum of those age-specific variances.

Traditionally, a normal approximation method is used to calculate confidence intervals on the assumption that the underlying deaths-data rates are based on are normally distributed. However, in some instances, for example in Wales, the annual number of avoidable deaths in certain cause groups may be relatively small (fewer than 100), and may be assumed to follow a Poisson probability distribution. In such cases, it is more appropriate to use the confidence limit factors from a Poisson distribution table to calculate the confidence intervals, instead of a normal approximation method.

The method used in calculating confidence intervals for age-standardised rates based on fewer than 100 deaths was proposed by Dobson et al., (1991), as described in APHO, (2008). A normal approximation method was used to calculate 95% CI’s, where there were 100 or more deaths in a year.

Full details of all the methodological changes are available in the avoidable mortality quality and methodology information notes (211.2 Kb Pdf).

Back to table of contents

11. Registration delays

The information used to produce mortality statistics is based on the details collected when deaths are certified and registered. In England and Wales deaths should be registered within 5 days of the death occurring, but there are some situations that result in the registration of the death being delayed. Deaths considered unexpected, accidental or suspicious will be referred to a coroner who may order a post mortem and/or carry out a full inquest to ascertain the reasons for the death.

Avoidable mortality statistics are presented based on the number of deaths registered in each calendar year, rather than deaths occurring in that year. This method is used due to the requirement for consistent and timely data, despite the potential limitation of data quality caused by registration delays.

In 2013 97% of avoidable deaths that were registered in 2013 also occurred in 2013.

In 2013, there were fewer avoidable, amenable and preventable deaths registered within 5 days than deaths from all causes. Approximately 59% of all avoidable deaths, 64% of amenable deaths and 58% of preventable deaths were registered within 5 days, whereas 73% of all causes were registered within 5 days. The average (median) time taken to register avoidable deaths in 2013 was 5 days, a day longer than all causes of death.

Back to table of contents

12. Uses and policy context

Statistics on avoidable mortality are used by central government, public health observatories, academia and charitable organisations working to reduce the prevalence of specific diseases and conditions deemed to be avoidable causes of death.

One of the main expected uses of these statistics is the monitoring of the quality performance of healthcare and public health policies. The Department of Health uses PYLL from causes considered amenable to healthcare for children and young population in its NHS Outcomes Framework 2015/16 (Department of Health, 2015) and mortality from preventable causes as an indicator in its Public Health Outcomes Framework to reduce preventable ill health, population dying prematurely and to reduce the gap between communities (Department of Health, 2014).

There has been considerable local and international interest in the development of statistics on avoidable mortality in the last 2 decades. In the UK, charitable organisations such as the Hepatitis C Trust, the British Lung Foundation and the British Association for the Study of Liver (BASL) are keen to see the conditions or diseases they campaign about included in the list of causes of death considered avoidable. It is thought that if these conditions were included on the cause list it would draw increased attention towards them and allow comparisons of trends to be made against other conditions.

At an international level, the European Union funded project ‘Avoidable mortality in the European Union: Towards better indicators for the effectiveness of health systems’ (AMIEHS, 2011) aims to develop a list of indicators (causes of death) for which mortality rates are likely to reflect variations in the effectiveness of healthcare, as defined by primary care, hospital care and personalised health services. To date, the project has developed an atlas containing trends in mortality for 45 possible amenable causes. Similarly, the Office for Economic Co-operation and Development (OECD) published a working paper in 2011, ‘Mortality amenable to healthcare in 31 OECD countries: estimates and methodological issues’. The study assessed the feasibility of using amenable mortality as an indicator of the performance of healthcare systems in OECD countries, concluding that there is lots of potential for cross-country comparisons of healthcare effectiveness.

Back to table of contents

13. Results on our website

Avoidable mortality figures for England and Wales combined, England, Wales and the regions of England are available on our website.

The workbooks contain results for England and Wales (combined and separately):

  • age-standardised rates per 100,000 population (with 95% confidence intervals), standardised years of life lost and the numbers of deaths for causes considered avoidable, amenable and preventable for the period 2001 to 2013 – data are available by cause group: infections; neoplasms; drug use disorders; cardiovascular diseases; respiratory diseases; injuries; and other conditions

  • the chances of eventually dying from avoidable causes based on the assumption that future mortality rates remain the same as those experienced in 2013

  • age-standardised rates per 100,000 population (with 95% confidence intervals) and the numbers of deaths for causes considered avoidable, amenable and preventable for the period 2001 to 2013 – data are for the regions of England, males, females and all persons

  • underlying dataset containing the number of deaths for each avoidable mortality cause group, broken down by sex and five-year age groups

Back to table of contents

14 .References

Association of Public Health Observatories (2008). Technical Briefing 3: Commonly Used Public Health Statistics and their Confidence Intervals. [accessed 07 April 2015]

AMIEHS (2011) Avoidable mortality in the European Union: Towards better indicators for the effectiveness of health systems, Final Report [accessed 07 April 2015]

Bajekal M, Scholes S, Love H, Hawkins N, O'Flaherty M, Raine R and Capewell S (2012) Analysing recent Socioeconomic Trends in Coronary Heart Disease Mortality in England, 2000–2007: A Population Modelling Study [accessed 07 April 2015]

Charlton JRH, Hartley RM, Silver R and Holland WW (1983)Geographical variation in mortality from conditions amenable to medical intervention in England and Wales. The Lancet 321(8326):691–696 [accessed 07 April 2015]

Department of Health (2015) The NHS Outcomes Framework 2015/16 [accessed 07 April 2015]

Department of Health (2014) Public Health Outcomes Framework [accessed 07 April 2015]

Department of Health (2013) Living well for longer (2014) [accessed 07 April 2015]

Dobson A, Kuulasmaa K, Eberle E and Scherer J (1991). Confidence intervals for weighted sums of Poisson parameters. Stat Med., 10:457-62. [accessed 07 April 2015]

Kossarova L, Holland W, Nolte E and McKee M (2009) Measuring ‘Avoidable’ Mortality, The London School of Economics and Political science [accessed 07 April 2015]

Mackenbach JP, Bouvier-Colle MH and Jougla E (1990) "Avoidable" mortality and health services: a review of aggregate data studies. Journal of Epidemiology and Community Health 44: 106-111 [accessed 07 April 2015]

Nolte E and McKee M (2004) Does health care save lives? Avoidable mortality revisited. The Nuffield Trust, London [accessed 07 April 2015]

Office for Economic Co-operation and Development (2014) Mortality amenable to healthcare in 31 OECD countries: estimates and methodological issues [accessed 07 April 2015]

Page A, Tobias M, Glover J, Wright C, Hetzel D and Fisher E (2006) ‘Australian and New Zealand Atlas of Avoidable Mortality’, University of Adelaide: Adelaide: PHIDU. [accessed 07 April 2015]

Rutstein DD, Berenberg W, Chalmers TC, Child CG, Fishman AP, Perrin EB, Feldman JJ, Leaverton PE, Lane MJ, Sencer DJ, and Evans CC (1976). Measuring the quality of medical care -A clinical method. New England Journal of Medicine 294:582-588 [accessed 07 April 2015]

Back to table of contents

15 .Background notes

  1. Statistics on mortality are derived from the information provided when deaths are certified and registered. Further information about the methods and quality of these statistics can be found in the mortality metadata (2.7 Mb Pdf) document on our website.

  2. In England and Wales deaths should be registered within 5 days of the death occurring. However, there are some situations which result in the registration of deaths being delayed. Deaths considered unexpected, accidental or suspicious will be referred to a coroner who may request a post mortem or carry out a full inquest to ascertain the reasons for the death. Further information on the impact of registration delays on the quality of mortality statistics can be found on our website.

  3. This bulletin presents age-standardised (also known as ‘directly-standardised’) rates, standardised to the 2013 European Standard Population (ESP). The ESP makes allowances for differences in the age structure of the population, over time and between sexes. The age-standardised rate for a particular cause of death is that which would have occurred if the observed age-specific rates for that cause had applied in the given standard population. Further information can be found on the revised European Standard Population 2013 page of our website. Previously published rates for 2001 to 2012, which were based on the 1976 ESP, have now been revised in light of the 2013 ESP.

  4. Figures are for deaths registered in each calendar year, while rates are based on mid-year population estimates as the denominator.

  5. Within this bulletin, a difference which is described as ‘statistically significant’ has been assessed using confidence intervals. Confidence intervals (CIs) are a measure of the statistical precision of an estimate and show the range of uncertainty around it. Calculations based on small numbers of events are often subject to random fluctuations. Significance is assigned on the basis of non-overlapping CIs. While more formalised and accurate methods of significance testing are available, the non-overlapping CI method is used because it is both simple to calculate and easily understood. As a general rule, if the confidence interval around an estimate overlaps with the interval around another, there is no significant difference between the 2 estimates.

  6. Special extracts and tabulations of mortality data for England and Wales are available to order for a charge (subject to legal frameworks, disclosure control, resources and agreement of costs, where appropriate). Such requests or enquiries should be made to

    Mortality Analysis Team, Life Events and Population Sources Division
    Office for National Statistics
    Government Buildings
    Cardiff Road
    Newport
    NP10 8XG

    Tel: 01633 456491
    E-mail: mortality@ons.gsi.gov.uk

    The ONS charging policy is available on our website.

  7. As a valued user of our statistics, we would welcome feedback on this release. Please send feedback to the postal or email address above.

  8. Details of the policy governing the release of new data are available by visiting www.statisticsauthority.gov.uk/assessment/code-of-practice/index.html or from the Media Relations Office email: media.relations@ons.gsi.gov.uk

  9. A list of the names of those given pre-publication access to the statistics and written commentary is available in ‘Pre-release access list to Avoidable Mortality in England and Wales, 2012 . The rules and principles which govern pre-release access are featured within the Pre-release Access to Official Statistics Order 2008

  10. Follow us on Twitter and Facebook

  11. National Statistics are produced to high professional standards and released according to the arrangements approved by the UK Statistics Authority.

    © Crown copyright 2015. You may re-use this document/publication (not including logos) free of charge in any format or medium, under the terms of the Open Government Licence v3.0; or write to the Information Policy Team, The National Archives, Kew, Richmond, Surrey, TW9 4DU; or email: psi@nationalarchives.gsi.gov.uk. Where we have identified any third party copyright information you will need to obtain permission from the copyright holders concerned. This document/publication is also available on the ONS website.

    Next publication:
    May 2016

    Issued by
    Office for National Statistics
    Government Buildings
    Cardiff Road
    Newport
    NP10 8XG

    Media contact:
    Tel: Media Relations Office 0845 6041858
    Emergency on-call 07867 906553
    Email: media.relations@ons.gsi.gov.uk

    Statistical contact:
    Tel: Andrew Tooley and Olugbenga Olatunde +44 (0) 1633 455397
    Email: mortality@ons.gsi.gov.uk

    Website: www.ons.gov.uk

  12. Details of the policy governing the release of new data are available by visiting www.statisticsauthority.gov.uk/assessment/code-of-practice/index.html or from the Media Relations Office email: media.relations@ons.gsi.gov.uk

Back to table of contents

Contact details for this Statistical bulletin

Andrew Tooley
mortality@ons.gsi.gov.uk
Telephone: +44 (0)1633 455397