Guide rates best and worst hospitals

The Daily Telegraph has today revealed the findings of the latest Dr Foster Hospital Guide. The guide, published annually, closely scrutinises a range of healthcare data to measure hospital performance and detect trends that could save lives.

As well as listing the hospital trusts in England that score above and below average on a range of different mortality measures, this year’s guide also found that:

The report also showed that the best-performing hospitals were generally in the south of England, while those found to be the poorest performers were mostly in the north. However, the report did not examine why this was the case.

Dr Foster Intelligence is a joint venture between the Department of Health and Dr Foster Holdings LLP and their research partners at Imperial College London. It aims to improve the quality and efficiency of health and social care through better use of information. It provides comparative information on health and social care services to health professionals and organisations to help improve the standard of healthcare.

The 2011 report is the tenth Dr Foster Hospital Guide to be published.

Despite overall improvements in mortality, some hospitals have consistently higher mortality rates than others.

For the first time, the Hospital Guide used four measures of mortality:

Four indicators give a more balanced view than a single indicator. Using these indicators, the report found that no trust is higher than expected on all four mortality measures, but two trusts – Hull and East Yorkshire Hospitals and University Hospital of North Staffordshire – are higher than expected on three out of four.

The following 19 hospital trusts have higher than expected mortality rates based on two measures – HSMR and SHMI:

Overall, 24% of trusts have higher than expected SHMI, 15% have higher than expected HSMR, 3% higher deaths in low-risk conditions, and 1% higher deaths after surgery.

One hospital, Chelsea and Westminster Hospital, achieved low mortality rates across all four mortality indicators.

The following hospitals were low (i.e. performed well) on three measures – HSMR, SHMI and deaths in low-risk conditions:

Royal Devon and Exeter NHS Foundation Trust was low for three different measures – HSMR, SHMI and deaths after surgery.

The following trusts were low (i.e. performed well) on two measures – the HSMR and SHMI:

Overall, 22% of trusts have lower than expected SHMI, 19% have lower than expected HSMR, 8% lower deaths in low-risk conditions, and 1% have lower rates of deaths after surgery.

Some trusts appeared to have both good and bad mortality results, which could be due to the way hospitals record deaths. For example, the Aintree University Hospitals NHS Foundation Trust has both a lower than expected HSMR and higher than expected SHMI.

Dr Foster’s guide draws attention to this inconsistency, explaining that this may be due to palliative care deaths being included within the HSMR. Different hospitals tend to code palliative care deaths in different ways, and higher rates of palliative care recording can lower a hospital’s mortality rate. If the relevant hospital has recorded their palliative care deaths, the HSMR adjusts for these deaths, which the report says makes it fairer on hospitals that care for terminally ill patients and who would otherwise be shown to have higher than normal in-hospital mortality rates.

Besides Aintree, nine other trusts coded a quarter of their HSMR as being palliative care cases. The SHMI measure, on the other hand, does not adjust for palliative care deaths. Dr Foster says it supports calls for palliative care coding guidelines to be made clearer.

In general, hospitals with the fewest senior doctors available at weekends have the highest mortality rates. A 2010 study by the Dr Foster Unit observed that people admitted to hospital over the weekend with common cardiovascular emergencies or with cancer were 7% more likely to die than those admitted from Monday to Friday.

Key factors that can contribute to higher mortality rates outside of normal working hours are:

The last point on staffing was considered to be a particular contributing factor that the report focused on. Dr Foster mapped senior staffing availability to the number of hospital beds against mortality rates for 130 trusts. They observed that more senior staff per bed at weekends is associated with a lower weekend mortality rate for emergency conditions, while more senior doctors (as a percentage of all doctors) is associated with lower rates.

The report noted nine trusts whose HSMR was within the expected range for people admitted Monday–Friday, but higher than expected for those admitted at the weekend:

The range of senior consultant availability at night was also wide. While almost a third of hospitals with an A&E unit had no consultants on site during the night, others had five or more consultants available in the hospital.

Of particular note is the risk associated with hip fracture at these times. Overall, patients who break their hip have a one in ten chance of dying, but the chance of surviving is much greater if they receive surgery within two days. For people admitted on Friday or Saturday, there is a lower chance of prompt treatment. In-hospital mortality in 2010/11 was observed to vary from 3.2% to 16.3% between providers. Numerous studies have shown that organisational factors in the patient’s treatment play a major part in determining patient survival.

2010 statistics on hip fracture showed:

In the following five trusts, 50% of all hip fracture patients waited more than two days for an operation:

Dr Foster says the answer may lie not necessarily in increasing the number of out-of-hours staff and services, but in reorganising the resources available to target where they are most needed. An example is networking with other hospitals in an area.

London has reorganised its stroke care in this way. Instead of all A&E departments treating strokes, a small number of hospitals now manage all stroke patients at a very high standard, seven days a week, 24 hours a day. Before the reorganisation (in 2009/10) 10% of stroke patients died within seven days of admission if they were admitted at the weekend compared to 8% who were admitted on weekdays. For weekend admissions in 2010/11, mortality has dropped to 7.3% compared to 6.4% for weekday admissions.

The report provides examples of other trusts that have reconfigured their services to provide more consistent out-of-hours care.

In general, patients treated in hospitals that perform operations rarely are more likely to die than in hospitals that perform a higher number of operations. This is particularly the case for major cardiovascular conditions such as an abdominal aortic aneurysm (a weakened section of the major artery that runs through the body, which has a very high mortality risk if it ruptures).

The risk of dying from major surgery used to treat this condition is 70% higher in hospitals that perform a lower number of these operations. Hospitals that perform fewer than 35 of these operations a year have a 13% patient mortality rate compared with 8% among hospitals that perform more than 35. (The report defines low-volume hospitals as those doing more than ten but 35 or fewer operations a year.)

Various factors may contribute to the difference in mortality across hospitals, including:

The report lists the large number of trusts which performed 35 or fewer procedures for abdominal aortic aneurysm in 2010/11.

Surgery for abdominal aortic aneurysm is the only area covered by this report. In other words, it cannot be inferred that you are at higher risk if you go into a hospital that performs fewer operations of any other type than another hospital.

The report also discusses factors that can help improve patient safety and outcomes.

It discusses the difference the rapidly increasing use of percutaneous coronary angiography (PCI, a technique to open up heart vessels blocked during a heart attack) has made to mortality from heart attacks: mortality has decreased by 2.5% since 2006. According to the report, it normally takes an estimated 15 years from the discovery of a new treatment to its widespread use by doctors, but the faster this happens, the greater the benefit seen.

Another factor in improving patient safety and mortality is following best practice and cost-effective patient care (care that is the safe and effective for patients and at the same time makes the best use of the NHS budget). The report discusses hip and knee replacements, which have increased over the past five years due to the increasing age of the population. Trusts that perform the best for these procedures had fewer patients with a long length of stay in hospital, fewer emergency readmissions within 28 hours, and lower rates of re-operation (a repeat operation done within one year of the initial procedure). Good care can also cost less in the long term.

The Hospital Guide Questionnaire looked at how certain trusts were improving patient recovery and reducing length of patient stay after these orthopaedic procedures. This is known as the Rapid Recovery Pathway. Factors that can improve patient recovery are:

In addition to hospital mortality rates, another important indicator of hospital performance is what patients say about their treatment. Online patient feedback can provide information that is not always clear from statistics, and websites such as NHS Choices and Patient Opinion now feature thousands of detailed comments on how patients view their treatment. The Dr Foster report says that comparing reports on these systems with national patient surveys has shown a reasonable degree of agreement:

The hospitals most often recommended percentage of people recommending the hospital) were:

Hospitals least often recommended were:

Where there was dissatisfaction, the five factors most likely to contribute to this were:

Private hospitals seemed to score well. It is difficult to tell the reasons for this. As these comments were registered on NHS Choices, they may reflect NHS patients being treated by private units. Also, the comparison between NHS and private may not be equal as private hospitals may be smaller and also manage less complex cases.

The report concludes with its Trusts of the Year, which had the best four mortality indicators and best scores in response to three questions on the national patient survey, which asked:

The four highest-performing hospitals according to these results were:

Only the Chelsea and Westminster Hospital NHS Foundation Trust scored low on all four mortality measures.

NHS Choices allows you to score the treatment you have received and leave specific explanations of what made your treatment good or bad. These opinions are publicly viewable, meaning you can read what other people have experienced before choosing where you want to be treated.

The service can be used to rate not only hospitals, but also a range of services, including GP surgeries and dentists. See our services finder to choose and rate your NHS services.