Senior staff from 16 hospitals got together to brainstorm on reducing avoidable deaths. Hugh Rogers and Kate Jones report

Over recent years a number of hospitals have been making efforts to reduce avoidable deaths, some supported by the former Modernisation Agency, others by the Health Foundation. The service improvement team at the NHS Institute for Innovation and Improvement brought together senior managers and clinicians from 16 hospitals to share learning, and commissioned an objective evaluation, available on its website.

They have reduced overall mortality by approximately 10 per cent - over and above the underlying improvement of 2 per cent a year in the UK. A reduction of 10 per cent across England would save at least 10,000 lives per year. Some have cut mortality by as much as 20 per cent.

Tackling avoidable mortality means getting care right for every patient. Standardising care and adopting care bundles will help achieve this. More deaths can be avoided by ensuring the systems of care delivery are reliable than through reducing harmful errors, but both approaches are important.

'Chief executives need to know their hospital standardised mortality rates and support approaches to improving them. We should be deeply concerned about the avoidable deaths happening weekly in our own hospital,' says Luton and Dunstable Hospital foundation trust chief executive Stephen Ramsden.

An important first step is to review data on your trust's overall mortality rates, and break it down by specialty and by common clinical conditions. However, this will not be sufficient to make the issue real to clinicians and managers. Enlightened senior clinicians should carry out a case-note audit of 50 patients who died, using a 'trigger tool'.

This is a checklist of clues that suggest something has gone wrong, such as return to theatre or unplanned admission. The trigger tool makes the review quicker yet gives a high likelihood of picking up errors and adverse events. An adverse event rate of 10 per cent is to be expected, only a tiny proportion of which will be picked up by incident reporting systems or traditional audit meetings.

Admission analysis

Next, the deaths are arranged into a matrix showing whether the patient was admitted to an intensive treatment unit or general ward and whether their admission was for palliative or active care. This will highlight local issues and suggest relevant interventions.

This local information provides a very powerful argument for change. Reading about the incidence of harmful errors is quite different to seeing them occurring in your own service. Clinical leaders need to communicate the results widely and present the findings with stories from the case notes. Challenging questions should be asked.

Most clinicians are focused on care quality, so once they understand the problem they will be keen to get involved. They will need support setting up appropriate measurement systems and understanding improvement methods.

We found that trusts that work to increase clinical quality have increased their efficiency. This is because patients who receive the right care in a timely way recover faster (reducing length of stay) and because avoiding adverse events reduces the cost of putting things right. These costs can arise from medicines, further procedures, expensive admissions to a high-dependency unit or even litigation.

Supporting structures

All hospitals agree that the involvement of the chief executive and senior clinicians is important and that reliability and safety need to be key priorities for the executive team. Some trusts set up structures that align service improvement and clinical governance. It is important to build this relationship, as the staff will bring complementary skills.

Regular senior 'safety walk-rounds' to wards and departments encourage staff to highlight safety concerns.

The case-notes audit will identify issues that trusts should be working on. It is important to relate changes to your local findings, but a number of trusts have employed a common 'core' group of interventions. The first three all relate to patients whose deterioration on the ward goes unrecognised or, if recognised, is not met by an appropriate response to rescue them:

  • assessing clinical observations using early warning scoring systems;
  • communication tools such as 'SBAR' (situation, background, assessment, response);
  • rapid response teams to provide clinical support for sicker ward-based patients;
  • care bundles to standardise care;
  • daily ward rounds to speed decision-making;
  • safe administration of medicines;
  • reduction in healthcare-associated infections.

Always start small with tests of change. Encourage innovation by staff in their attempts to improve care systems. Monitor the reliability, effectiveness and impact of interventions at ward or department level - there is often plenty of room for improvement.

Measurement system Develop a regular measurement plan. You can only tell if interventions have made a positive impact if you measure things that are relevant to your changes.

At ward or department level, use process measures for assessing reliability such as percentage of patient observations completed, or percentage of prophylactic antibiotics given. Results can be posted on run charts in a corridor or ward office. This will remind staff that the issue is important to the organisation, as well as demonstrating progress.

An effective high-level outcome measure is the hospital's weekly deaths plotted on a statistical process control chart, to show significant improvements.

Hugh Rogers is associate and Kate Jones is head of the service transformation team at the NHS Institute for Innovation and Improvement.HSJ is working with the institute on the Good Management Live series of best practice events, including one on patient safety hosted by Luton and Dunstable Hospital foundation trust. Click here to book your place