Intended for healthcare professionals


Getting evidence into practice

BMJ 1998; 317 doi: (Published 04 July 1998) Cite this as: BMJ 1998;317:6

This article has a correction. Please see:

Needs the right resources and the right organisation

  1. Fiona Godlee, Editor, Clinical Evidence.
  1. BMA House, London WC1H 9JR

    NHS's 50th anniversary pp 61, 72

    The person who did perhaps more than anyone to bring the rigours of systematic review into clinical research, Tom Chalmers, once asked: Why do doctors kill more people than airline pilots do? He suggested 10 reasons. These included the fact that pilots are required to have time off to sleep, that they do everything in duplicate, and that they follow protocols. But his final reason was that if doctors died with their patients they would take a great deal more care.

    Taking more care means, among other things, practising evidence based health care, and, even to enthusiasts, death for failing to do so seems harsh. After all, as Gina Radford, director of Britain's new National Institute for Clinical Excellence, said at a recent meeting on evidence based medicine in York, no one goes to work to do a bad job. If they are to improve how they care for patients, clinicians need to know what they are doing wrong, or badly, and how to put it right. At the moment this is difficult.

    Firstly, the medical literature is unwieldy, disorganised, and biased. Most research published in medical journals is too poorly done or insufficiently relevant to be clinically useful. In a recent survey, over 95% of articles in medical journals failed to reach minimum standards of quality and clinical relevance.1 Good research on important questions is often analysed and presented in ways that make it hard to apply in clinical practice. In answer to a question about the risks associated with the oral contraceptive pill, only five of 74 articles identified by a systematic review contained information in a useful form.2

    Secondly, many of the questions that arise daily in clinical practice remain unaddressed by well designed research. Studies have suggested that up to 80% of clinical decisions are based on good evidence, 3 4 but these studies looked mainly at prescribing decisions. Evidence on many other types of decisions—such as when to investigate, which test to use, and when to refer, not to mention the complex mix of sociology, mythology, and pastoral care that make up general practice—is sparse and its quality poorly defined.

    Nor are clinical practice guidelines the long term solution they once appeared to be. They are slow and expensive to produce, mostly of poor quality, and hard to update. Although they can change practice in some circumstances—when they are locally developed, involve a specific education strategy, and have patient specific reminders at the time of consultation5—anecdotal evidence suggests that they are not widely used.

    Finally, there is the problem that medicine is traditionally a solitary profession—one clinician dealing with one patient. Finding out how well you are doing and how you could do better can be difficult without the help of well designed and administered systems for audit and feedback.

    But help is at hand, as described in a series starting this week on getting research evidence into practice (p 72).6 Thanks to the Cochrane Collaboration and others, good systematic reviews are now available in many areas of health care, overcoming the biases inherent in the biomedical literature and providing a firmer base for clinical decisions. These are available on the Cochrane Library CD Rom. Abstracting journals such as Evidence Based Medicine, Evidence Based Mental Health, and Evidence Based Nursing identify the best and most relevant clinical research in their areas; the Best Evidence CD Rom presents a cummulative record. The major electronic databases are making searching easier by incorporating quality filters for different types of search question. Training courses and books on critical appraisal are helping clinicians to become educated consumers of these new resources. And because the information may still seem hard to access and understand, a new tool for clinicians, Clinical Evidence, will soon bring this concentrated wisdom a few steps closer to patient care (see box).

    “Taking more care” involves more, of course, than getting research evidence into practice. At the York meeting, Liam Donaldson, director of the NHS Executive's Northern and Yorkshire region, listed his three ingredients for success in health care organisations—culture, culture, and culture. He warned that the design of the organisation (in Britain's case, the NHS) must be right for evidence based medicine to flourish. The new framework for organisational change in England is clinical governance, and on p 61 Scally and Donaldson explain what this means and what we can expect if it succeeds.7


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.


    Clinical Evidence

    Clinical Evidence is a compendium of summaries of the best available evidence on a range of important clinical questions. Produced jointly by the BMJ Publishing Group and the American College of Physicians, it will be updated and expanded twice a year, both as a book and on the web. It does not make recommendations, and where there is no good evidence it says so. Contributions are written by practising clinicians with expertise in evidence based medicine. The first issue will be available in January 1999. For more information, contact mnasser{at}

    Embedded Image

    See also advertisement for clinical editors in Classified Supplement (special appointments).

    View Abstract