Article Text

Download PDFPDF

Applying evidence: what’s the next action?
  1. Paul Glasziou
  1. Centre for Evidence-Based Medicine; University of Oxford; Oxford, UK

    Statistics from

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Every fortnight my practice group gathers for an early morning “journal club” where we discuss a paper that we think could change our practice. A large pot of tea, an informal atmosphere, and lively discussions have made the sessions popular. But our interest is not the entertainment of pulling apart papers or curiosity about pathophysiology: our central interest is in clinical practice improvement. As in other journal clubs, we critically appraise the paper, calculate the number needed to treat (NNT), and decide on the clinical bottom line. But most journal clubs end there, assuming that everyone will implement the change. But, as we’ve repeatedly experienced, that assumption is often false.1 So following the traditional EBM steps, we ask, “So what is the next action?”2

    The “next action” can be a variety of things. Sometimes the next action is garnering some additional information through another search, writing to the authors, or writing to a local expert. It doesn’t need to be a whole implementation plan, just the next physical action and who is responsible for it.2 For example, at a recent journal club, we discussed measuring blood pressure in both arms. A useful systematic review3 of blood pressure differences between arms made us aware of several important facts: (a) the prevalence is high, with around 6% of the adult population having at least a 6 mm Hg difference between arms, (b) the prevalence increases with cardiovascular risk factors such as age, diabetes, and smoking, and (c) the pressure in the higher arm is the “correct” one and more predictive of future risk. So we agreed we should do this but recognised the reality of our time pressure (or laziness): unless this was made easy, we wouldn't really do it. We agreed on 2 things: (i) we would only do this once on current patients with hypertension or those suspected of having high blood pressure, and (ii) we should ask the nurse who runs our blood pressure clinic to do this as part of the hypertension clinic.

    Clinical practice changes can be very simple (just do it!) or can involve many steps taking months or years. At the journal club session, we aim to identify only the next action to take and who will do it and by when. Next actions vary in type but might roughly be divided into “whether to” and “how to.” The “whether to” actions mean that either we are not convinced that the change is worth making or that it may not be practical. For “whether-to’s,” we may to need to gather some more external information by writing to the study author, doing further literature searches, or writing to a local expert. Or we may seek internal information about the number of patients in our practice that will be affected by the change. This information will be brought back to the next journal club meeting for discussion.

    Having decided the “whether to,” we need to consider the “how to,” which may include some training, purchasing equipment, altering a protocol, preparing a patient handout, etc. For example, when we decided it was a good idea to routinely check blood pressure in both arms of patients newly suspected of having raised blood pressure, it took several steps before making the change. First, we wrote to one of the guideline committees to ask how this should be done. When they were unable to provide the specifics, we next wrote to the systematic review authors for suggestions. We decided first that we would purchase 2 identical automated blood pressure machines to be kept in the nurse’s room. New patients would have a pair of measurements done, and if there was more than 10 mm Hg difference, this would be repeated several times to confirm the apparent difference. Patients with a difference would be told which arm should be measured, and a computer reminder would be set to alert clinicians at each visit.

    In this blood pressure example, there were several steps (a sequence of next actions) before we set up a satisfactory system to implement the change. There are no routine next actions: each clinical change involves its own problem-solving process. Some examples, given in the table, illustrate the diversity. It might be possible to ease this burden if research articles with implications for clinical practice suggested what steps needed to be taken to implement a change. While that may help, I suspect there will always be some tailoring needed for individual clinics and clinicians because of local resources and skills, and personal tastes. And even if what is needed is clear, someone in the practice must still decide the “what” and “when” of change and take responsibility to monitor that change has occurred. In quality improvement,4 this is known as “plan, do, check, and act”: plan the next action, do it, check that it works as planned, and finally, make it routine and monitor it.

    The journal club session and the follow-up actions constitute the plan and do of a plan-do-check-act cycle. The next step is to check on whether the changes have actually occurred. This follow-up check can be as simple as asking the clinicians a few months later, “Is everyone now doing the Epley manoeuvre?”1 to doing an audit of recent medication prescriptions. Some examples are given in the table.

    Examples of next actions arising from a general practitioner journal club

    Quality improvement4 in medicine has commonly been about more efficient appointments or discharge or other administrative processes and less about clinical improvement. However, this linking of evidence-based medicine and quality improvement is important if we are to reap the improvements in clinical outcomes that good research promises. By identifying the next action,2 we can break down complex changes into a series of easily do-able tasks. This is not an onerous “implementation plan,” but just the essential tasks for taking our evidence into practice. These extra 2–3 minutes at the end of the journal club can help overcome barriers to evidence-based practice and add to a sense of satisfaction that comes from continuous improvement of care.


    My thanks to the clinical team at 19 Beaumont St and to Amanda Burls and Brian Haynes who provided helpful comments.