Press release

New hospital mortality indicator to improve measurement of patient safety

New information on mortality rates in NHS hospitals has been published today

This was published under the 2010 to 2015 Conservative and Liberal Democrat coalition government


New information on mortality rates in NHS hospitals has been published today as part of plans to give patients and the public more transparent and robust information about their local NHS.

The new Summary Hospital-level Mortality Indicator (SHMI) compares the actual number of patients who die following treatment at a trust with the number who would be expected to die, given the characteristics of the patients treated there.

For the first time, it considers all deaths that take place both in hospital and within 30 days of discharge, offering a more comprehensive picture of deaths following hospital care.

The SHMI shows mortality rates for every acute non-specialist trust in England - providing a single comprehensive indicator that will be used consistently across the NHS.  It will also highlight trusts with the lowest mortality which can provide valuable learning on how quality of care can be improved.

Each trust has a single SHMI value but the data has been published with two different methods of categorising trusts as having ‘as expected’, ‘higher than expected’ and ‘lower than expected’ mortality rates. One method reduces the potential for falsely identifying borderline trusts as ‘higher than expected’, and therefore identifies fewer trusts as higher or lower than expected. The other method is more sensitive, identifying more trusts as higher or lower than expected.

The data shows:
• the vast majority of trusts have a mortality rate that falls within an expected range - 119 using the less sensitive control limits and 79 using the more sensitive control limits
• for Trusts with higher than expected mortality, 14 are identified using the less sensitive control limits and 36 using the more sensitive control limits
• 14 trusts have lower than expected mortality using the less sensitive control limits and 32 Trusts for the more sensitive control limits.

Health Secretary Andrew Lansley said:

“We are determined to improve patient safety and shine a light on poor performance by giving patients, public and the NHS more robust information about their hospital trust.

“As I have highlighted this week, we are doing all we can to improve care for patients and help turn around struggling hospitals.

“This new measure will help ensure patient safety by acting like a smoke alarm to prompt further investigation. Alongside other data, this will help the NHS in future to spot and act on poor care as soon as possible. We are determined to learn the lessons of the appalling events at Mid Staffordshire - this data will help us avoid a repeat of that tragedy.

“A more transparent NHS is a safer NHS where patients can be confident of receiving high quality care.”

NHS Medical Director Professor Sir Bruce Keogh, who commissioned the review of mortality indicators on behalf of the National Quality Board, said:

“The SHMI adds to our understanding of hospital mortality, but no one indicator alone can give us a complete picture of a hospital’s performance.

“For example, no-one would buy a car based only on the mileage or how many miles you get to the gallon - you would look at lots of information before making a decision.  In the same way, to truly understand the quality of care at a hospital, you must look at this alongside other information.

“All hospital trusts, regardless of whether they are ‘outliers’, need to examine, understand and explain their SHMI - and identify where performance may be falling short. Trusts with a low mortality rate could also provide valuable learning about how quality of care can be improved.”

Hospital mortality ratios are complex indicators which have prompted international debate about their definition and interpretation. The Francis Review into Mid Staffordshire NHS Foundation Trust recognised the uncertainty this debate was causing and recommended a national review into hospital mortality ratios so that variations and trends associated with hospital deaths could be better understood. That review was carried out by a wide range of the experts and reached a consensus on SHMI as a new indicator.

The indicator is still new and ‘experimental’ - refinements will have to be made in the future. No statistical model can ever perfectly estimate the risk of mortality.

Notes to editors

1. The Summary Hospital-level Mortality Indicator (SHMI) shows mortality rates for every acute non-specialist trust in England for the period from 1 April 2010 to 31 March 2011.

2. Summary Hospital-level Mortality Indicators are intended to compare the observed number of deaths that actually occurred at a hospital with a statistical estimate of the number of deaths that might have been expected, based upon national average death rates and the particular characteristics of the patients treated in each hospital.

3. It is statistically invalid to rank hospitals’ quality of care in a league table based on their SHMI value because the SHMI on its own does not measure the quality of care. But it is valid to identify those hospitals with a higher or lower than expected mortality ratio using SHMI because it is designed to assess whether the mortality rate at an individual hospital is within the expected range or not after taking into account the risk profile of patients served by that hospital.

4. Following the Review of the Hospital Standardised Mortality Ratio (HSMR) in 2010, the Department of Health committed to adopting the SHMI as the single summary-level indicator for hospital mortality. It commissioned the NHS Information Centre to develop, deliver and publish the SHMI as it does all national indicators.

5. The SHMI gives a broader picture of hospital mortality than the Hospital Standardised Mortality Ratio (HSMR). It includes all deaths in all settings - not just in hospital - up to 30 days after discharge and covers all clinical codes rather than just those relating to 80 per cent of the most common causes of death in hospital. Therefore it is not possible to compare SHMIs with HSMRs.

6. The SHMI values for each non-specialist acute trusts are being published by the NHS Information Centre.

7. Information on the SHMI for non-specialist hospital trusts will also be available to patients on the NHS Choices website.

8. A consensus statement on the SHMI, supported by a range of stakeholders, has been published on the Department of Health website.

9. The Francis Review recommended that:
 ”In view of the uncertainties surrounding the use of comparative mortality statistics in assessing hospital performance and the   understanding of the term ‘excess’ deaths, an independent working group should be set up by the Department of Health to examine and report on the methodologies in use. It should make recommendations as to how such mortality statistics should be collected, analysed and published, both to promote public confidence and understanding of the process, and to assist hospitals to use such statistics as a prompt to examine particular areas of patient care.”

10. This indicator has been developed in collaboration with a range of national stakeholders following a review commissioned in 2010 by medical director for the NHS in England Sir Bruce Keogh and chaired by Ian Dalton, then chief executive of the North East Strategic Health Authority. It involved a wide range of stakeholders, including the Department of Health, representatives from strategic health authorities and trusts, the NHS IC, the Care Quality Commission, Monitor, the Kings Fund, the Academy of Royal Colleges, the NHS Confederation, Dr Foster Intelligence, CHKS, University Hospitals Birmingham, the National Patient Safety Agency and Professor Sir Brian Jarman and colleagues from Imperial College.  During 2011, several members have continued to support and contribute to the technical work associated with the development and construction of the SHMI (from the Department of Health, the Care Quality Commission, Dr Foster Intelligence, Dr Foster Imperial , CHKS and University Hospitals Birmingham). 

11. For more information, please contact the Department of Health press office on 020 7210 5221.

Published 27 October 2011