Keogh review on hospital deaths published

The findings of a review into the quality of care and treatment provided by 14 hospital trusts in England has prompted widespread coverage in the press.

The review, started in February 2013, was led by Professor Sir Bruce Keogh, the National Medical Director for the NHS in England. It looked at the quality of the care and treatment provided by 14 trusts identified as having higher than average death rates in the two years before the start of the review.

Eleven of these trusts are to be put under ‘special measures’ in order to improve governance.

The review has revealed problems in care that had not been exposed before. While the report says immediate safety issues found were dealt with straight away, it also calls for co-ordinated efforts to improve care and accountability in the longer term.

Why was the Keogh review commissioned?

The review was commissioned by the prime minister, David Cameron, and the secretary of state for health, Jeremy Hunt, in response to the findings of the Mid Staffordshire Public Inquiry.

It aimed to look into the quality of care and treatment being provided by English hospital trusts with higher than average death rates in the previous two years.

While above average death rates can often be accounted for by other factors (such as the hospital serving an area with an older population), previous health scandals have shown that particularly unusual results in data ("outliers") should never be ignored.

The 14 trusts were selected on the basis of having higher than average scores on one of two well-established measures of death rates. These are:

• the hospital standardised mortality ratio (HSMR), which compares the expected rate of death in a hospital with the actual rate of death
• the summary hospital level mortality indicator (SHMI), which compares death rates between individual hospitals

The report set out to:

• determine whether there are any ongoing failings in the quality of care provided to patients at these 14 hospital trusts
• identify whether the trusts' actions to improve quality is adequate and whether additional steps are needed
• identify if any additional support should be made available to the trusts
• identify any areas that may require legal (regulatory) action to protect patients

What data did the Keogh review look at?

The review was carried out in three stages and considered the performance of the hospitals across six main areas:

• deaths
• patient experience
• workforce
• clinical and operational effectiveness
• leadership
• governance

Stage 1 – information gathering and analysis

All information covering the six key areas was gathered for each trust and analysed. Findings were then compared with national average standards. Areas of concern were followed up in a visit to the hospital involved.

Stage 2 – rapid responsive review

Review teams were trained to carry out planned and unannounced site visits at each of the 14 trusts for two or three days at a time. These teams were made up of 15-20 people and included patients, doctors, nurses, managers and regulators. The visits involved walking the wards and talking to patients, trainees, staff and senior executives. Findings were recorded in a rapid responsive review report. Individual interviews and approximately 70 staff focus groups were carried out as part of a cultural assessment.

Stage 3 – risk summit and action plan

Once the reviews were completed, a meeting ("risk summit") was held to agree a co-ordinated plan of action with each trust, including support to speed up improvements and identify who was accountable.

What were the key findings of the Keogh report?

The report found examples of good care as well as areas where improvement is urgently required. In the report, Professor Sir Bruce Keogh says: "We found pockets of excellent practice in all 14 of the trusts reviewed. However, we also found significant scope for improvement, with each needing to address an urgent set of actions in order to raise standards of care."

Key findings from the review include:

• Understanding that concepts such as excess deaths and avoidable deaths are more complex than analysing a single-level summary death rate indicator (two widely used death rate indicators were the basis of results of this review).
• There are many different causes of high death rates and there is no "magic" solution.
• Death rates in NHS hospitals have been falling over the past 10 years and the rate of improvement in the 14 hospitals under review has been similar to other NHS hospitals.
• Factors often claimed to be associated with higher death rates (such as access to funding and poor health of the local population) were not found to be statistically associated with the results of these hospitals.
• Accuracy of clinical coding (the way hospitals make a computerised record of diseases, operations and other "healthcare episodes") can impact on death indicator numbers. For example, the review says that coding patients to make them appear sicker or identifying a higher amount of multiple conditions can improve death rates, but arguably represents an attempt to "fix the figures". Some hospitals were said to not be responding to the signals the figures were identifying as they felt they were incorrect, which is potentially a matter of concern.
• More than 90% of deaths in hospital happen when patients are admitted in an emergency rather than for a planned procedure. The review says it is therefore not surprising that all of the 14 hospital trusts had higher deaths in urgent and emergency care, and only one trust  (Tameside General Hospital) had high death rates for elective procedures.
• Understanding the causes of higher death rates is said not to be about finding a "rogue surgeon" or problems occurring in a single specialty area. The review says it is more likely to be a combination of problems that all hospitals in the NHS experience, such as busy A&E departments and wards, treatment of the elderly, and the need to recruit and keep excellent staff. 

Where areas of concern were found in any of the trusts, immediate action was taken, including:

• immediate closure of operating theatres
• suspension of out-of-hours stroke services
• instigating changes to staffing levels
• dealing with backlogs of complaints from patients

The review identified areas of action in the next two years as well as some common themes and barriers to delivering high quality care. These themes are:

• A limited understanding of how important and simple it can be to genuinely listen to the views of patients and staff, and engage them in how to improve services.
• The ability of hospital boards and leaders to use data to drive quality improvement. This theme is made more difficult by how hard it is to access data held in different places and different ways across hospitals systems.
• The complexity of using and interpreting summary measures of death (HSMR and SHMI).

How accurate was the media's reporting of the Keogh review?

The review was widely covered in the media with a variety of headlines, and some inaccurate reporting. The Guardian reports that secretary of state for health Jeremy Hunt is sending "hit squads" into failing NHS trusts, while the Daily Mail reports that Hunt "vows to sack hospital bosses if they reject sweeping changes to improve care".

There is also a widely quoted figure that NHS failings have led to 13,000 unavoidable deaths. This figure was given by Professor Sir Brian Jarman, a member of the review's national advisory group, in a BBC radio interview. It is currently unclear from the media coverage what evidence Professor Jarman  used to make these claims, but the media has reported this figure as a fact that emerged from the main review itself, when actually the report gives no such figure.

In another example, Mail Online says: "NHS medical director, Professor Sir Bruce Keogh, says there were thousands of needless deaths". In fact, Keogh said: "However tempting it may be, it is clinically meaningless and academically reckless to use such statistical measures to quantify actual numbers of avoidable deaths."

The review reports that the staff of the 14 trusts involved embraced the review, were open and honest, and showed commitment towards improving quality of care for patients.

As would be expected, many of the headlines had a political angle, with the Daily Mail reporting, "20,000 extra NHS deaths on Labour's watch amid calls for on-site inspectors", and The Telegraph saying, "Thousands may have died because of Labour NHS failings".

BBC News had the most accurate reporting of the review's findings.

Conclusion

In a letter to the secretary of state, Professor Keogh reports that assessments of the 14 hospital trusts have been highly rigorous and uncovered problems in care that had not been exposed before. He warns against hasty reactions and pointing the finger of blame.

Any immediate safety issues discovered are said to have been dealt with. Professor Keogh states that considered debate is needed, as well as co-ordinated efforts to improve care with a future focus on accountability.