Intended for healthcare professionals

Observations Ethics Man

The death of DNR

BMJ 2009; 338 doi: (Published 30 April 2009) Cite this as: BMJ 2009;338:b1723
  1. Daniel K Sokol, lecturer in medical ethics and law, St George’s, University of London
  1. daniel.sokol{at}

    Can a change of terminology improve end of life care?

    In a scene in the film Dumb and Dumber, Lloyd Christmas, played by Jim Carrey, sees a fellow diner collapse in a restaurant. The man clutches his abdomen and complains of an ulcer. “It’s OK,” Christmas reassures the victim, “I know CPR.” The man resists mouth to mouth resuscitation. “It’s a lot easier if you just lay back,” Christmas notes.

    While perhaps not quite as ignorant as the well meaning Christmas, many non-clinicians hold rosy views about the nature and effectiveness of cardiopulmonary resuscitation (CPR). Several studies have underlined their misplaced optimism: in one, the 269 respondents reported a mean expected survival rate for CPR of 65%1; in another, 81% of respondents over 70 years old believed the likelihood of leaving the hospital after a cardiac arrest to be at least 50%.2 The real figure, for all in-hospital cardiac arrests, is roughly 14%, and many survivors will have new functional or neurological impairments.3

    The illusion of CPR’s effectiveness can lead patients and relatives to make ill informed choices about care at the end of life. To emphasise the fallibility of the exercise, many institutions have abandoned the term “do not resuscitate” (DNR) in favour of “do not attempt resuscitation” (DNAR). Still, the discussion about the suitability of a DNAR order can be difficult for patients, relatives, and clinicians alike. So awkward can it be that many such discussions, which should form an important component of the future care plan, are avoided entirely.4

    Although raising the issue of death is seldom easy, part of the struggle is to dispel misunderstandings about DNR orders. DNR does not mean “do not treat,” much less “do not bother.” With the exception of those in intensive care, many patients with DNR orders survive to discharge. DNR means “if the patient has a cardiac arrest, do not attempt cardiopulmonary resuscitation.” Indeed, some trusts now use the acronym DNACPR. The manner in which the situation is described is arguably more important in resuscitation decisions—when tension, fear, and guilt may be palpable—than in any other area of medicine. To help dispel myths and improve understanding, a further change of terminology has been suggested: “allow natural death” (AND).5

    A study published earlier this year on the views of nurses, nursing students, and laypeople in south Texas showed that changing the title from DNR to AND increased endorsement of the order in all three groups, reaching statistical significance in the second two groups.6 It is not a surprising result, given the gentler, more benevolent tone of “AND.” It is devoid of the cold negativity of “do not resuscitate,” with its connotations of abandonment and a death sentence. AND better reflects what so many of us believe should happen when the bell tolls: the peaceful, unobstructed flow from life to death.

    Adopting the change should help reduce stress and feelings of guilt among all parties and may encourage clinicians to initiate the discussion with suitable patients or relatives more often than they currently do. The situation where patients who should have had a DNR order are resuscitated and are left on the ward, in a hopeless condition, to die a second time should become less frequent. Not only will the indignity of CPR on the inexorably dying occur less often, but the finance managers, recognising the potential savings of fewer days on the ward, should rejoice at the likely cost implications of the change.

    There are problems with AND, not least the potential for mistaking AND with the conjunction “and.” In the early days of implementation we should follow the example of one US institution that used AND/DNR to accustom staff to the new terminology.5 Once the new acronym is widely known and the DNR part can be dropped, we should find a way to distinguish AND from its more pedestrian homograph, perhaps by circling the term or some other simple method.

    AND lacks the specificity of DNR.7 Allowing natural death, understood literally, may require withholding or withdrawing all sorts of treatment from the patient: no ventilation, no antibiotics, no dialysis, no palliation. Yet, often it may be appropriate to treat a DNR patient therapeutically.8 If the term is introduced, we must determine exactly what we mean by it to avoid misinterpretation. AND, like DNR, does not necessarily entail forgoing aggressive treatment; and admittedly this fact does not sit comfortably with the literal interpretation of “allow natural death.” As with DNR, any discussion of AND would be accompanied by a discussion of what care should and should not be offered. The vagueness of AND, rather than being a disadvantage, could encourage clinicians to have that discussion with patients and relatives.

    As Hippocrates noted many years ago, clinicians should try to benefit patients with minimum harm. If a change of terminology can improve end of life care by reducing anxiety and costs, then surely such change is morally desirable, if not perhaps morally obligatory. The immediate priority is to identify, as exhaustively as possible, the logistical and practical challenges of making the transition from DNAR to AND—and to address them. We have much to learn from those pioneering institutions in the United States and the United Kingdom that have already effected the change. While changing the language alone will not overcome all the problems with resuscitation decisions, it is a step in the right direction, towards a healthier relationship between patients, relatives, and clinicians and a more peaceful end for many.


    Cite this as: BMJ 2009;338:b1723


    • Thanks to Robert Elias, Raanan Gillon, Tom Palser, Dee Traue, Rob George and Ronald P Sokol for their comments on earlier drafts.


    View Abstract