Intended for healthcare professionals

Editor's Choice Editor's choice

Evidence to inform

BMJ 2007; 334 doi: (Published 21 June 2007) Cite this as: BMJ 2007;334:0
  1. Fiona Godlee, editor
  1. fgodlee{at}

    The BMJ aims to help doctors make better decisions. Often this may best be achieved by helping doctors to help patients make better decisions. Yet recent research from the Picker Institute has found that doctors in the UK are worse at engaging patients in healthcare decisions than doctors in comparable countries. Use of decision aids is only one aspect of engaging patients but it's a potentially important one. A Cochrane review has found that decision aids improve people's knowledge of the options, create realistic expectations of benefits and harms, reduce difficulty with decision making, and increase participation.

    The trial on decision aids for mode of delivery by Alan Montgomery and colleagues (doi: 10.1136/bmj.39217.671019.55) adds to this list of successes and, as Jeremy Lauer and Ana Betran point out in their editorial (doi: 10.1136/bmj.39247.535532.80), it raises intriguing possibilities about how decision aids could be made even more effective. Focussing on the woman's preferences rather than explicitly presenting probabilities for different options was associated with higher rates of vaginal delivery—a highly desirable outcome when up to a quarter of births in developed countries, and even larger proportions in transitional countries, are by caesarean section. Crucially, this outcome required not only an informed patient but one who had the power to influence her mode of delivery.

    How long before it will be unethical to ask a patient to consent to treatment without using a decision aid? I predict it will not be long—despite the time needed to use aids and the limited availability of good ones. The Ottawa Health Decision Centre is working on this. Run by Annette O'Connor, it provides a list of decision aids, each scored for quality against an international standard (

    Also this week, two bits of advice from specialists to generalists in areas where good evidence is hard to come by. There is no strong evidence that perioperative use of β blockers or statins is beneficial, say Stephan Bolsin and colleagues (doi: 10.1136/bmj.39217.382836.BE). Two large trials are planned or underway, but recruiting enough participants will be challenging, they say. It will be sobering to unpick why perioperative β blockade was so widely adopted with so little evidence to support it.

    Meanwhile, Ketan Dhatariya highlights an evidence-free zone in which specialists and generalists differ on a key aspect of care (doi: 10.1136/bmj.39237.661111.80). Diabetologists have long believed that saline is the intravenous fluid of choice for patients with diabetic ketoacidosis, a view enshrined in national guidelines. But most patients are first seen by emergency and intensive care physicians, who tend to prefer Hartmann's solution. There are shades here of the albumin debate, which was resolved by a systematic review of RCTs. This concluded controversially that albumin, which had been used for over 50 years in critically ill patients, increased rather than decreased mortality (BMJ 1998;317:235-40). Ironically, it is the potential dangers of Hartmann's solution—still so widely used by non-specialists—that may prevent a randomised trial, says Dhatariya. But experience suggests that only hard evidence will change such long held beliefs and practices. And there's always a chance the specialists would be proved wrong.