Intended for healthcare professionals


Hospital tables “should prompt authorities to investigate”

BMJ 2001; 322 doi: (Published 20 January 2001) Cite this as: BMJ 2001;322:127
  1. Zosia Kmietowicz
  1. London

    The publication last weekend of death rates in hospital trusts throughout Britain has been hailed as a “courageous attempt” to compare the performance of hospitals, but one that offers little help to patients.

    The research shows that for every 10 people who die in the best hospitals, 17 die in the worst, even after adjusting the figures for patients' ages, sex, diagnoses, whether the admission was an emergency, and length of stay.

    The biggest predictor of death rates was the number of doctors in the hospital and the area. The more doctors per bed and the more GPs per head of population, the lower the death rate. The analysis was carried out by Sir Brian Jarman, emeritus professor of primary health care at Imperial College of Science, Technology and Medicine in London.

    Heading the league table for fewest deaths was University College London Hospitals, which boasted 63 doctors for every 100 hospital beds, whereas Walsall Hospitals performed the worst, with just 23 doctors per 100 beds.

    Sir Brian and his team used published death rates for NHS hospital trusts, adjusting them as far as possible to compare like with like. Social background was examined but found to be less significant than other factors included in the analysis.

    A guide to hospitals in Britain and Ireland was published with the research in the Sunday Times which, together with publisher Dr Foster Ltd, commissioned the work.

    Martin McKee, professor of European public health at the London School of Hygiene and Tropical Medicine, praised the research, calling it “a courageous attempt to try to compare the quality of hospitals.”

    But he expressed concern about the methods used to compile the statistics, which took account only of deaths in hospitals and number of admissions while omitting 30 day mortality, multiple admissions, and whether there was a hospice close to the hospital.

    Professor McKee added that, although patients may not find the figures very helpful—“a hospital that is performing poorly is generally known to the local population”—the results should prompt health authorities and primary care trusts to question why one hospital is performing less well than another.

    Dr Azeem Majeed, senior lecturer in general practice at University College London and an epidemiologist with the Office for National Statistics, welcomed the increased availability of information on hospital standards but was concerned that the public would not be aware of its pitfalls.

    “There are a number of problems if you are going to use this type of information as a guide to hospitals. The most important is that there is huge variation within a hospital between the performance of different departments and the individual doctors within those departments, and this variation is wider than the differences between hospitals,” said Dr Majeed.

    “These statistics tell us something, but not much. For example, if you had bowel cancer it won't tell you whether your doctor is any good.”

    Embedded Image

    The best predictor of death rates is the number of doctors

    (Credit: SUNDAY TIMES)

    View Abstract