OUP user menu

Experiences and views of specialist registrars in geriatric medicine on ‘do not attempt resuscitation’ decisions: a sea of uncertainty?

P.K. Myint, S. Miles, D.A. Halliday, L.K. Bowker
DOI: http://dx.doi.org/10.1093/qjmed/hcl096 691-700 First published online: 6 September 2006


Background: Recent cultural changes place doctors under increasing pressure to work with their patients to reach decisions about end-of-life care.

Aim: To survey the experience, practice and opinions of specialist registrars (SPRs) in geriatric medicine regarding ‘do not attempt resuscitation’ (DNAR) decisions.

Design: Postal questionnaire survey.

Methods: A questionnaire was mailed to all members registered as trainees (n = 408) with the British Geriatrics Society in November 2003; a reminder was sent nine weeks later. Responses were analysed using both quantitative and qualitative (thematic) approaches.

Results: Response rate was 62% (251/408), of whom 235 were still SpRs. Respondents played a major role in DNAR decision-making in their day-to-day clinical practice. Over a third of respondents did not feel that locally available guidelines were helpful. More than half sometimes disagreed with their consultants’ decision, and a fifth were concerned about the possibility of complaints regarding the decisions they made. The majority felt uncomfortable discussing the issue with the patient, and were more likely to discuss the issue with relatives than with patients.

Discussion: Further support and training may improve confidence and positive experiences in relation to DNAR decision-making among training-grade doctors in the UK.


Despite growing awareness of ‘do not attempt resuscitation’ (DNAR) orders and emergence of guidelines, around 20% of patients who die in UK hospitals will have one or more attempts at cardiopulmonary resuscitation (CPR) during their terminal admission.1 Recent cultural changes, including increasing patient and family involvement in health-care decisions, a political drive to offer choices to patients, and growing anxiety about litigation, mean that doctors are under increasing pressure to work together with their patients to reach decisions about end-of-life care.

Various professional bodies have produced CPR/DNAR guidelines since Doyal and Wilsher's publication in 1993,2 where the requirement of formal guidance on CPR was first proposed. In 2000, the Chief Medical Officer, who oversees the medical practitioners in England in Wales, stated that all NHS trusts should generate individual guidelines that ‘respect patients’ rights, are understood by all relevant staff, and accessible to those who need them, and that such policies are subject to appropriate audit and monitoring arrangements’.3 Earlier guidelines from the Royal College of Physicians,4 the British Geriatrics Society5 and the British Medical Association6 preceded the Human Rights Act, which was implemented in the UK in late 2000. Several of the articles of the Act aim, directly or indirectly, to make doctors more aware of their obligations to involve their patients in their DNAR/CPR decision-making.7 This has led to further guidelines being produced by the British Medical Association, in conjunction with the Resuscitation Council (UK) and the Royal College of Nursing (BMA/RCUK/RCN),8 and recent guidance from the General Medical Council (GMC).9 However, the legality of the GMC guidelines has been challenged recently.10 This case highlights the possibility of future challenges over DNAR decisions made by front-line clinicians, even when professional guidelines are followed.

O'Keefe described his own experience with guidelines in relation to the practicality of implementing them,11 and concluded that it is difficult to involve acutely ill patients in DNAR decision-making. Most studies to date, however, have mainly focussed on adherence to guidelines (e.g. following legible recording of orders or following steps) rather than the personal experience and preferred clinical practice of decision-makers.

In everyday clinical practice of DNAR decision-making, it is not uncommon to encounter unreasonable demands from patients themselves or their relatives, which can produce conflict, and take time and skill to resolve. Therefore, it is essential that doctors in higher specialist training receive appropriate training and support in this area. In the UK, geriatricians are the largest group of physicians who provide not only acute and community services to older people but also acute medical service to younger adults. Higher medical training at specialist registrar (SpR) level in geriatric medicine in the UK involves five years of clinical training after a post-graduate qualification, MRCP (UK) diploma or equivalent, and requires at least two years experience in acute hospital medicine at senior house officer level (medical officer equivalent). Following this, a certificate of completion of training in general internal medicine and geriatric medicine is awarded, enabling doctors to take up a consultant post in the NHS. Consultants in the UK are specialists who practice independently and also responsible for training of future consultants.

In this study, we examine current clinical experience and the routine approach to DNAR decision-making of specialist registrars (SpRs) in geriatric medicine. The rationale for focussing on higher trainees in geriatric medicine is mainly that in the UK, geriatricians are not only the largest group of physicians, but also most likely to be involved with the DNAR decision-making process. Therefore, it is important to identify the experiences and views of SpRs, as future leaders of geriatric medicine, to enable provision of adequate support and training.


In November 2003, a questionnaire designed to investigate DNAR decision-making was mailed to trainee members of the British Geriatrics Society (BGS), inviting their voluntary responses. The mailing list was obtained from the BGS registration office. The questionnaire incorporated both quantitative and qualitative items to explore respondents’ routine clinical practice and views on DNAR decision-making. Free entry into an author-funded prize draw was offered as an inducement. A reminder was sent to non-responders 9 weeks after the first mailing.

The questionnaire contained four domains (Appendix, questions 1–4) examining the respondents’ clinical experience, current decision-making practice, preferences concerning involvement with such decision-making processes, and personal experiences. These four domains were structured as follows: Question 1, clinical experience (years of practice in the UK since graduation, involvement in cardiac arrest situation, experience of cardiac arrest situation), self-reported frequency of DNAR decision-making outside team structure, awareness and experience of local guidelines; Question 2, current decision-making practice involving interaction with patients or relatives (responses to given scenarios); Question 3, preference concerning involvement with such decision-making processes; Question 4, personal opinion, memorable experiences and predictions as to the future of the DNAR decision-making process (free text).

Quantitative data were analysed using SPSS for Windows version 12.0.1 (SPSS Inc.) using a significance level of p < 0.05. Free text data were analysed qualitatively using a thematic coding schema. Themes were derived from the textual data. Frequency and percentages were given for each response. Two coders (SM and DAH) each analysed half of the reported data. Inter-coder agreement was checked by each coder analysing 10% of the other coders’ data. Disagreements were discussed until a consensus was reached. This process resulted in some codes being merged. Themes from the data were discussed with PKM and LKB to check for clinical and contextual accuracy.


After two mailings, the overall response rate was 62% (251/408). Of these, 16 were no longer SpR grade. Therefore, analyses were based on the remaining 235 responses from SpRs (94% of 251). Not every respondent answered all the questions.

Respondents had practiced medicine in the UK for a median 8 years (range 1–20) and attended cardiac arrest calls for a median 6 years (range 1–14).

The majority were involved with DNAR decisions ‘sometimes’ or ‘commonly’ in patients who were not under their care (195, 83%), and they did make DNAR decisions in younger adults (18–55 years) (186, 80%). Whilst almost all SpRs participating in this study (227, 97%) were aware of the presence or absence of local Trust guidelines in DNAR decision-making, only two-thirds (123, 63%) of the 195 respondents whose Trusts had such a policy thought it was helpful in their day-to-day practice.

In relation to this, when asked what document/guideline they would recommend as guidance for DNAR decisions for their colleagues, 28% (66) provided no response, indicating that over a quarter of the respondents had no document/guideline that they would recommend. Of those recommending a guideline, about a fifth recommended their own trust/local guidelines (48, 21%), or BMA guidelines (45, 19%). Joint BMA/RCUK/RCN guidelines were recommended by 29 (12%) respondents and GMC guidelines by 27 (11%). A few respondents stated that they would recommend discussion with the patient or senior clinician, or using common sense instead of recommending a particular guideline (Box 1, Quotes 4(b): 1102, 1244). There was a suggestion that there were various inadequacies with current guidelines (Box 1, Quotes 4(b), 1014, 1113). These inadequacies may have led to anxiety such as fear of complaint or litigation in relation to DNAR decisions (Box 1, Quotes 4(b), 1109).

Box 1. Quotations from the responses to questions 4 (a), (b), (d) and (e)

4 (a) What single factor would make DNAR decisions easier and why?

Greater patient/public awareness of outcomes of resuscitation and what DNAR means (i.e. does not mean withdrawal of active treatment)

1113 A public debate to explain the chance of success, what resuscitation is, and why it is often futile.

1200 If the issue were routinely raised with the patients outside the acute situation. If everyone understood that DNAR does not mean ‘not for active treatment’.

Advanced directives/Living wills

1119 Discussing this thorny issue of DNAR with a vulnerable, ill patient often compounds their existing worries that ‘the doctor believes I’m going to die because he is discussing resuscitation issues with me’. Directives made while ‘healthier’ may be more neutral and helpful to final decision-making.

Part of the standard procedure on admission

1097 Asking every patient who walks through the A/E door whether or not they would want resuscitation (not just the people who are frail), i.e. it should be part of the process and universal.

DNAR decisions based on clinical judgement/medical decision-making, individual clinician, teams, or ICU staff

1168 A wider appreciation of likely successful outcome stratified (for example) by disease, concomitant disease, etc. This would allow us to give patients and relatives a more concrete idea of the likelihood of long-term survival.

4 (b) Which document/guideline would you recommend as guidance for DNAR decision for your colleagues?


1102 The ‘Duties of a doctor’ booklet, but sensible discussion with experienced responsible clinician is usually better than general written guidelines.

1244 There are various things to take into account: (i) patient wishes; (ii) underlying medical problems; (iii) quality of life; (iv) chances of successful resuscitation; (v) decision made with colleagues and relatives; (vi) age.

Insufficiencies of guidelines

1014 I know the guidelines well, and I am not sure I would recommend them. The joint resuscitation/RCN guidelines are very contradictory, even within one sentence sometimes.

1113 None. There is no good guidance that covers the difficulties, some documents contradict themselves.

Concern over complaint/litigation

1109 Follow local policies: then the Trust has to support you if things go wrong.

4 (d) In your current clinical practice, what is the single most influencing factor in making DNAR decisions?

Likelihood of success

1002 The main factor is the likely outcome. I see it as a medical treatment. If it's not going to work, don't do it.

1103 Outcome/survival probability. I always ask myself the question ‘Should this patient be resuscitated?’ or ‘Will they survive an ITU stay?’

Medical condition

1141 Whether the underlying disease process is reversible and a resuscitation attempt is likely to have a positive outcome.

Quality of life

1126 Quality of life before and after.

Patient's wishes

1247 Patient's own wishes.

4 (e) Do you see any change in future regarding DNAR decision making? If so, what?

1224 Resuscitation forms completed on admission for all patients giving patient wishes/status.

1189 More open discussion about subject. Patient makes decision in advance.

1242 More use of advance directives, more open approach by clinicians and involving the patient earlier.

1253 Probably more patient involvement in decision-making, consequently with more CPR attempts that are almost certainly going to be unsuccessful.

Table 1 summarizes the responses from questions where respondents were asked about their current clinical experience with DNAR decisions. Relatively few of the respondents seek the patient's opinion on DNAR decisions. In contrast, the majority discuss the DNAR decision with a relative and seek relatives’ opinion as a ‘proxy’ to the patient's opinion in DNAR decision-making. Some 44% (n = 101) base their DNAR decisions primarily on expected quality of life after resuscitation over half the time, and 71% (n = 164) base decisions on the likelihood of successful resuscitation over half of the time. While 43% (n = 100) would never make a partial DNAR decision, an almost equal number of SpRs would make such a decision at times.

View this table:
Table 1

Extent to which DNAR decisions were based on various practices

PracticeRespondents⩽25% of the time26–50% of the time51–75% of the time⩾76% of the time
Seeking the patient's opinion on DNAR decisions233 (99%)119 (51%)67 (29%)24 (10%)23 (10%)
Seeking a relative's opinion as proxy for the patient230 (98%)60 (26%)64 (28%)31 (14%)75 (33%)
Discussing the decision with a relative232 (99%)9 (4%)39 (17%)38 (16%)146 (63%)
Basing the decision primarily on expected quality of life after resuscitation230 (98%)74 (32%)55 (24%)26 (11%)75 (33%)
Basing the decision primarily on likelihood of successful resuscitation231 (98%)26 (11%)41 (18%)30 (13%)134 (58%)
Making partial DNAR decisions232 (99%)214 (92%)14 (6%)4 (2%)0

When respondents were asked to detail the single most important influencing factor in making DNAR decisions, almost a third (72, 31%) said that likelihood of success was a key influencing factor (Box 1, Quotes 4(d): 1002, 1103), futility was mentioned as an influencing factor for 18 (8%) and the patient's prognosis for 17 (7%). For these respondents, the likelihood of a poor outcome, futility and poor prognosis would lead them to a DNAR decision. The patient's medical condition was a key influencing factor for 80 of the respondents (34%) (Box 1, Quotes 4(d), 1141). Of these, 14 stated that irreversibility was an appropriate reason for DNAR, 18 specifically mentioned that the patient's pre-morbid condition was an important consideration, and 27 considered co-morbidities to be important influences on their decision-making. Quality of life (QoL) was a key factor in making DNAR decisions for 42 (18%) respondents (Box 1, Quotes 4(d), 1126). About one third of these respondents did not provide further information, but another third referred to pre-morbid or pre-admission QoL, and for the remaining third, expected post-resuscitation QoL was the key factor. Forty-three respondents (18%) mentioned that the patient's wishes were important (Box 1, Quotes 4(d), 1247).

Table 2 shows agreement levels for questions regarding preferred clinical practice for DNAR decision-making. Around half of the SpRs who responded agreed that they found it uncomfortable to discuss DNAR with the patient. About half of the respondents agreed that they had had different opinions from their consultants regarding DNAR at some point, and that they were better skilled than their colleagues from other specialities in making such decisions. Just over half of the respondents agreed that their decision varied depending on whether or not the patient in question was under their care. The majority (84%) disagreed that decision-making with regard to DNAR decision should be left to consultants alone. These findings are particularly interesting because consultants in the NHS in the UK are not only independent practitioners who hold the responsibility for management decisions, but also educational supervisors of their juniors, including SpRs. A fifth of the respondents were concerned about complaints being made against them over DNAR decision-making. Over half disagreed that relatives should only be informed about decisions but not actively involved in the decision-making. Furthermore, half agreed that more patients should be designated DNAR, in comparison to only a quarter who disagreed. A more detailed analysis was conducted to see whether there was any association between having experience with arrest situations and having a strong view on this issue. Respondents who had <7 years of experience (UK experience or experience in holding arrest bleep) were more likely to give a neutral response to the statement that more patients should be designated DNAR, than those with ⩾7 years of experience, who were more likely to agree (χ2 = 6.006, p = 0.05).

View this table:
Table 2

Levels of agreement with various statements

StatementRespondentsStrongly agreeAgreeNeutralDisagreeStrongly disagree
I find it uncomfortable discussing DNAR with patients234 (99%)32 (14%)89 (38%)38 (16%)59 (25%)16 (7%)
I sometimes disagree with the decision made by my consultant235 (100%)5 (2%)121 (52%)44 (19%)50 (21%)15 (6%)
I think I am better at making DNAR decision than my colleagues from other medical specialties234 (99%)27 (12%)93 (40%)64 (27%)47 (20%)3 (1%)
My decision can vary depending on whether the patient is directly under my care234 (99%)27 (12%)103 (44%)38 (16%)52 (22%)14 (6%)
I worry about getting complaints about CPR235 (100%)9 (4%)39 (17%)66 (28%)98 (42%)23 (10%)
I believe relatives should only be informed about decisions but not involved in decision-making234 (99%)27 (12%)44 (19%)29 (12%)110 (47%)24 (10%)
I think more patients should be made not for resuscitation234 (99%)26 (11%)91 (39%)56 (24%)51 (22%)10 (4%)
DNAR decisions should be made by consultants only235 (100%)2 (1%)13 (6%)22 (9%)145 (62%)53 (23%)

When asked to detail the single factor that would make DNAR decisions easier and why, 54 (23%) of the respondents felt that greater patient/public awareness of the outcomes of resuscitation and what DNAR means (i.e. that it does not mean withdrawal of active treatment) would make decisions easier. Just under a fifth of the respondents felt that Advanced Directives or Living Wills (43, 18%) would make DNAR decision-making easier, as would a Trust policy including a standard procedure at admission to record patients’ wishes on DNAR (40, 17%). Some respondents felt that DNAR decision-making should be based on clinical judgement/medical decision-making either by individual clinician, teams or intensive care unit staff (25, 11%). Discussion of the issues around DNAR away from the clinical environment would improve DNAR decision-making for some respondents (13, 6%), e.g. consideration of DNAR issues before the patient is critically ill, prior to hospital admission, in a non-hospital environment such as with the GP. Other factors mentioned by fewer respondents included: having more time with patients, early discussion with patients and relatives, consensus between patient, relatives and staff, reduced power of next of kin, more realistic guidelines, adopting medical futility and less litigation (Box 1, Quotes 4(a)).

Interestingly, when asked what, if any, changes they saw in the future regarding DNAR decision-making, the SpRs’ responses indicated that these factors that would make decision-making easier were also the kind of changes that they could see coming. This said, 31 (13%) respondents did not respond and 20 (9%) said they did not see any changes. Among those who did see changes in the future (Box 1, Quotes 4(e)) greater patient involvement was the biggest single change predicted (65, 28%); this included asking all patients their wishes at admission or early during the hospital stay, and discussion of DNAR with patients. Also, 29 (12%) thought that there would be increased discussion and awareness about DNAR and CPR generally amongst the public, relatives, patients and health professionals. The involvement of advocates, in particular relatives, when patients are unable to give their wishes was seen as a possible change by 16 (7%) respondents. More use of Advance Directives or Living Wills was mentioned by 23 (10%) respondents. Twenty-one respondents (9%) mentioned guidelines, with some indicating that there were problems with current guidelines, others suggesting that new, better or more formal guidelines were required, and finally others simply mentioning that guidelines needed to be used, or that everyone needed to be more aware of the guidelines. Twenty (9%) thought that in the future there would be fewer DNAR decisions and, consequently, more CPR attempts. The consensus here appeared to be that this would lead to more inappropriate resuscitation or automatic resuscitation of everyone. Nineteen (8%) said that in the future DNAR decisions would be left to doctors, with 10 of these respondents specifically mentioning consultant-led decision.


This survey confirms that geriatric medicine trainees are extensively involved in both resuscitation efforts and decisions surrounding withholding such resuscitation. The response rate of 62% is not surprising in a population that is highly mobile, leading to inaccuracies in mailing lists. With a response rate of about two-thirds of higher specialist trainees in one medical subspecialty, this survey should not be viewed as a census of experience, clinical practice and attitude towards DNAR decision-making amongst trainees in higher medical education in the UK. However, the clinical experience of the respondents had a median of 8 years, reflecting the present day SpR grade experience across all medical specialities. Studying a selected group as we did in this survey is not necessarily a limitation. On the contrary, their specific patient population may give them insights less available to other specialties. Our survey also provide useful information as to how we might help promote experience and training of SpRs with regard to CPR and DNAR decisions.

In a study by Einav and colleagues of 79 practitioners on general medical wards in Jerusalem, Israel,12 survey participants rarely discussed DNAR orders with patients and their next of kin, tending instead to confer DNAR based on their personal value judgments rather than patient preferences. They also performed CPR when no pathophysiological benefit was expected, and performed limited resuscitation efforts frequently, which were hypothesized as efforts to circumvent the need for DNAR orders. The authors highlighted the important differences between the ideal guidelines such as joint UK guidance8 and existing practice, in end-of-life decision-making in their cultural setting.

Usefulness of national and local guidelines

A high percentage of respondents stated that the Trusts in which they are currently working had DNAR guidelines. However, one third of them did not find such policies helpful. There are some possible reasons why such guidelines are of limited value. Florin cautioned, for example, that a policy that makes patient consent a universal requirement of DNAR orders is likely to fail.13 Opinions regarding DNAR decisions often differ between decision-makers14 and this also makes rigid guidelines likely to fail.

Almost a third of respondents did not answer the question on which document/guideline they would recommend. Lack of response to this question could be interpreted in a number of ways: e.g. unfamiliarity with available guidelines, not valuing the guideline used in their Trusts, or feeling that guidelines alone did not provide all the advice and information required to make DNAR decisions in practice. It may be that despite there being a range of guidelines in use, no one guideline is considered to be the gold standard, and some respondents pointed out inadequacies. The lack of response may also suggest that these SpRs do not solely base DNAR decisions on a written document, but combine it with other decision-making techniques, such as their own clinical judgement, the advice of their senior clinicians and the local Trust position on DNAR.

Current experience and involvement

The fact that the experience of the respondents extended to making decisions about patients who are not under their care, including younger patients, may reflect the nature of the training involving general medicine and out-of-hours middle-grade cover for the whole hospital. It is possible that DNAR decisions are not being taken early enough by a patient's own team, leaving on-call doctors to make them out of hours. The need to make DNAR decisions on patients outside the usual team structure has implications for training, and may lead to trainees’ anxiety about increased chance of medico-legal challenge. This may have contributed to the concern about getting complaints, and also may have influenced the decision-making in over half of the respondents, who said that they would agree or strongly agree to the statement that their decision can vary depending on whether the patient is under their care or not.

Discussion with patients and relatives

We found a huge diversity of clinical practice among the respondents. There appeared to be tension between medical futility and patients’ or close relatives’ unrealistic expectations, causing problems in reaching a unanimous decision. While some ethicists hold the opinion that patients should always be offered CPR, a less extreme view is that it is within the legitimate professional role of doctors to define the absence of medical benefit, but not to make quality-of-life decisions without involving the patient.15,,16 Unfortunately, our study suggests that even at specialist registrar level in a speciality which routinely has to deal with DNAR issues, doctors are still uncomfortable discussing end-of-life care issues with their patients, and are more likely to discuss DNAR decisions with relatives than with patients. This highlights the need for support and further training in this area. Greater public awareness and involvement of patients in the decision-making process, including use of Advance Directives, was seen as one way forward. There was a suggestion amongst some respondents that early discussion with patients as a standard part of the admission process might make discussion of DNAR issues easier, as such discussion would be expected by all patients as soon as they arrive at hospital.

In critically ill patients, when the patient's opinion cannot be sought for various reasons, doctors either seek proxy decision from close relatives or make CPR decisions based on perceived quality of life of their patient.

Estimating quality of life and success of CPR

Although there has been an attempt to differentiate between unacceptable pre-arrest quality of life and where quality of life after CPR is unacceptable,17 there is little evidence to help physicians predict which patients will undergo a change in quality of life due to CPR. Physicians often rely on assumed quality of life of their patients in their DNAR decision, but unfortunately, tend to underestimate it.18 Furthermore, although patients predict that both their physicians and family members would accurately represent their wishes,18 doctors and relatives are very poor at estimating both their patients’ quality of life and their wishes.19–21 While some patients do not want to be involved in such decisions, most patients would like more communication with their doctors about resuscitation issues.22,,23 Approximately one third of the respondents would primarily base their decision on the perceived quality of life of the patient after resuscitation at least 50% of the time. This practice, therefore, could potentially be improved with increasing awareness of the literature among the trainees.

Varying opinions among doctors in making DNAR decisions

Wagg and colleagues reported that doctors who had attended more than 25 CPR attempts were better at making realistic DNAR decisions, and that there was an inverse relationship between expectation of survival and number of arrests attended.24 DNAR recommendations also vary with medical speciality, and with years of training and experience.14 In this study, partial DNAR decisions were made by some of our respondents, but not by others. Over half had dissenting views on decisions made by their consultants at some point. Considering that the SpRs are fairly senior team members, there is an intriguing possibility that the trainers (consultants) are making inappropriate CPR orders or not making appropriate DNAR orders for various reasons including the fear of litigation, which should be examined in future studies. Worryingly, more than half of the respondents stated that their practice could vary depending whether the patient is under their care or not, which gives some insight into the difficulties that a decision-maker encounters outside the structure of the clinical team. Appropriate training and support may be the way forward to improve the clinical practice and experience of trainees in dealing with such delicate issues.

It could be argued that the large diversity in their opinions was related to the sociodemographic characteristics such as age, sex and religious beliefs of the SpRs. We deliberately did not collect any identifiable data from the participants, to assure complete confidentiality. The major drawback of this approach is that we are not able to provide such data about the respondents, or correlate these factors with their responses. Nevertheless, the primary aim of our study was to examine the experience, practice and views of higher specialist trainees in the UK on DNAR decision-making, regardless of their gender, racial or religious background.

One limitation of the survey is that the questions used were not validated. This made us unable to report one of the questions in this report, due to the ambiguity of the question (question 4c). However, the respondents did not appear to encounter any major issue in answering the remaining part of the questionnaire. This is probably because we have used standard validated methods, e.g. level of agreement, with which all higher trainees are familiar.

A sea of uncertainty?

Our study provides a snapshot view of current experience and involvement in DNAR decision-making of higher specialist trainees in a major medical subspecialty. Although our study cannot be regarded as representative of the situation among all higher specialist trainees, respondents in our survey are fairly representative of this grade in terms of clinical experience, and may be more familiar with this issue compared to SpRs in other subspecialties.

Our results suggest that even in a speciality which commonly encounters such ethical issues, SpRs are being tossed around in a sea of uncertainty. There is considerable inconsistency in opinion and practice in the DNAR decision-making process among these SpRs, and many do not find the available guidelines to be particularly helpful. Furthermore, these SpRs find it stressful to speak to patients about DNAR decisions, and they sometimes disagree with their consultants.

Various issues highlighted in this report, such as inadequacies of guidelines, stressful nature of discussion the DNAR issue with patients and relatives, and wide variation in the clinical practice among trainees, indicate that there is potential for improvement in further training and support from senior colleagues. Training and education in medico-legal aspects and increased awareness of the literature surrounding DNAR issues may also improve SpRs’ confidence in their decision-making process. It may be possible to adopt well-established methods used in some medical specialities (e.g. debriefing), particularly when SpRs are faced with traumatizing personal experience which could potentially influence their clinical practice in the future.

On the other hand, professional guidelines could also be clearer and more user-friendly. Some respondents pointed out inadequacies in some of the currently available guidelines. Continual review and user involvement may improve the standard and applicability of future CPR/DNAR guidelines. However, it would be impossible to provide clear guidance on every possible clinical scenario, and this kind of detailed approach has to be balanced with the generalizability of guidance and some clinical freedom for the practitioner. Further research is required to help formulate training initiatives and develop effective training methods for SpRs, and to find the most effective way to educate the public regarding DNAR decision-making. Ideally this should involve a partnership between the public, ethical experts, legal experts and the front-line clinicians, including trainees themselves.


We would like to thank the former Honorary Secretary of the British Geriatrics Society (BGS), Dr Kevin Kelleher, and Ms Recia Atkins from the BGS office for giving permission to conduct this national survey on trainee members of the BGS, and Mrs Marlene Gabriel and Mr Tony Winhall from Norfolk and Norwich University Hospital for their assistance with the project. We gratefully acknowledge the participants in this questionnaire. We also would like to acknowledge the contribution of one of the co-authors of this paper (DAH) who contributed on behalf of the East of England Research Development & Support Unit. Some of the data were presented in British Geriatrics Society Autumn Meeting, 6–8 October 2004 as a poster entitled ‘Do not attempt resuscitation (DNAR) decision making and trainees in Geriatric Medicine’, and published as an abstract in Age and Ageing, July 2005.

Appendix: Questionnaire survey on ‘Do Not Attempt Resuscitation’ decisions

1. About your experience with cardiac arrests and DNAR decisions

  1. How many years have you been practising in the UK? (Please include PRHO* year if it is applicable).

  2. During this practice, how many years have you held a crash bleep?

  3. Are you asked to make DNAR decisions for patients who are not under your team's care? (commonly/sometimes/rarely/never).

  4. Do you ever make DNAR decisions on younger adults (18 to 55 years of age)?

  5. Do you have a trust policy for DNAR in your hospital? If yes, do you find it helpful?

2. About your current practice with DNAR decisions (Please answer by approximate percentages, e.g. 20% of decisions)

  1. How often do you seek the patient's opinion on DNAR decisions?

  2. How often do you seek the relative's opinion as proxy for the patient?

  3. How often do you discuss your decision with a relative?

  4. How often do you base your decision primarily on expected quality of life after resuscitation (determined by you)?

  5. How often do you base your decision primarily on likelihood of successful resuscitation?

  6. How often do you make partial DNAR decision? [e.g. for DC shock if shockable rhythm but not for CPR if PEA(EMD)]

3. About your feelings regarding DNAR decisions (SA = strongly agree, A = agree, N = Neutral, D = disagree, SD = strongly disagree)

  1. I find it uncomfortable discussing DNAR with patients.

  2. I sometimes disagree with the decision made by my consultant.

  3. I think I am better at making DNAR decision than my colleagues from other medical specialities.

  4. My decision can vary depending on whether the patient is directly under my care.

  5. I worry about getting complaints about CPR decisions.

  6. I believe relatives should only be informed about decisions but not involved in decision-making.

  7. I think more patients should be made not for resuscitation.

  8. DNAR decisions should be made by consultants only.

4. About your opinion/practice in future (Please describe in 50 words or less. Continue overleaf if necessary)

  1. What single factor would make DNAR decisions easier and why?

  2. Which document/guideline would you recommend as guidance for DNAR decision for your colleagues?

  3. Briefly describe your worst or most memorable experience of DNAR (data not reported here).

  4. In your current clinical practice, what is the single most influencing factor in making DNAR decisions?

  5. Do you see any change in future regarding DNAR decision making? If so, what?

*PRHO, Pre-registration House Officer (equivalent to house officer/internship). The ‘crash bleep’ is a bleep held by several members of the medical on-call team during their duty. It is activated by a cardiac arrest, and the persons holding the bleep need to attend the cardiac arrest call immediately. A typical cardiac arrest team consists of a medical registrar (SpR) as the team leader, a medical senior house officer (SHO), a medical junior house officer, a resuscitation officer, an anaesthetist (either SHO or SpR) who can perform intubation, a nursing sister, a senior nurse and two porters.


View Abstract