Intended for healthcare professionals


What has evidence based medicine done for us?

BMJ 2004; 329 doi: (Published 28 October 2004) Cite this as: BMJ 2004;329:987
  1. Sharon E Straus, assistant professor,
  2. Giselle Jones, assistant editor BMJ (gjones{at}
  1. Toronto General Hospital, 200 Elizabeth Street, Toronto, Canada M5G 2C4

    It has given us a good start, but much remains to be done

    In this theme issue we ask how the evolution of evidence based medicine (EBM) has made a difference to the way we practice medicine and whether it has improved care for patients. So what is the evidence so far? The material we have collected shows that the answer is not straightforward, and it is still early days to be definitive about the success of the EBM movement. However, we hope this issue will fuel the debate by reflecting on the progress that has been made in practising and teaching EBM over the last 10 years and by drawing attention to those areas with which we continue to struggle.

    EBM is now an integral part of many undergraduate, postgraduate, and continuing education activities.14 Coomarasamy and Khan identified 23 studies of educational interventions involving EBM in the postgraduate environment and found that clinically integrated teaching improved knowledge, skills, attitudes, and self reported behaviours (p 1017).2 Teaching EBM is also being incorporated successfully into journal club environments, with positive effects on changing practice and improving the care of patients.4

    Accurate, accessible, and continually updated sources of evidence such as the Cochrane Library and Clinical Evidence are now available widely, and freely to some middle and low income countries. Garner et al outline how sources such as the Cochrane Library are being used in a collaborative effort to translate evidence into practice in middle and low income countries.5 Perhaps as important as identifying where evidence exists, sources of evidence also highlight gaps in the evidence that inform the focus of future research.

    Sceptics would argue that producing research is one thing but showing that it changes practice is another. However, a cluster randomised trial published in this issue shows that a collaborative quality improvement intervention to promote surfactant therapy in neonates lead to improved outcomes (p 999).6 Significantly more infants in the intervention group received surfactant in the delivery room, and fewer received the dose more than two hours after birth. Results from trials such as this one provide encouraging evidence that efforts to change behaviour, though difficult, are possible even when applied across multiple health settings.

    Despite these advances several challenges remain. Few articles address the impact of teaching EBM on clinical outcomes, and in particular those that matter to patients as well as clinicians. Coomarasamy and Khan did not identify any studies in their systematic review that evaluated the impact of postgraduate teaching of EBM on clinical outcomes,2 and scant data exist on changing behaviours other than from self reports. Although this has not been a requirement for other modes of medical practice or education, its importance should not be neglected. We suggest that to facilitate these efforts, those who teach EBM around the world establish a collaborative network not only to share educational materials but also to evaluate educational interventions and determine their impact on important clinical outcomes (p 1029).7

    Another barrier is that providing evidence from clinical research is a necessary but not sufficient prerequisite to change behaviour and improve patient care. A study by Sheldon et al on the impact of guidance from the National Institute for Clinical Excellence on prescribing practice and use of evidence based interventions in the United Kingdom found that implementation was variable.8 Proponents of knowledge translation would argue that changing behaviour is not a simple task and involves a complex process requiring comprehensive approaches directed towards all relevant stakeholders including patients, healthcare professionals, managers, and policy makers.9 In particular more emphasis needs to be placed on understanding and incorporating patients' values (which often differ widely from those of their doctors) into the process and working together towards a mutual evidence based decision.10

    We now have too many sources of evidence, compiled with a variable mix of scientific rigour and opinion, resulting in confusing messages. Sometimes evidence may favour an intervention, but health policy may prevent clinicians from providing it.11 Little wonder that the uptake of evidence is so piecemeal. The challenge will be to provide immediate access to high grade evidence in user friendly formats that are tailored to meet the needs of stakeholders.

    Finally, some practitioners are concerned about the legal implications that EBM poses. Could they be considered negligent by the courts for not applying evidence based guidance in decision making? This fear could result in the inappropriate, broad brush application of guidelines to every patient whereby the art of practising medicine is replaced purely with science—a soul destroying prospect for any clinician. These and other legal challenges that EBM practitioners have faced (and might face in the future) are reviewed by Hurwitz.12

    But we should not let these challenges deter us; this issue shows that EBM has achieved milestones since its evolution a decade ago. We already have enough evidence to answer most of the common clinical questions practitioners face, so the focus of the next 10 years should be on how to use it and how best to measure how we are doing.


    • Competing interests None declared.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.
    View Abstract