You are here

Article Text

Article menu

Download PDFPDF

Clinical ethics
Paper
Including patients in resuscitation decisions in Switzerland: from doing more to doing better
  1. Samia A Hurst1,
  2. Maria Becerra2,
  3. Arnaud Perrier2,
  4. Noelle Junod Perron3,
  5. Stéphane Cochet4,
  6. Bernice Elger5
  1. 1Institute for Biomedical Ethics, Geneva University Medical School, Geneva, Switzerland
  2. 2General Internal Medicine Service, Geneva University Hospitals, Geneva, Switzerland
  3. 3Department of Primary Care and Community Medicine, Geneva University Hospitals, Geneva, Switzerland
  4. 4Oncology Service, Geneva University Hospitals, Geneva, Switzerland
  5. 5Center for Legal Medicine, Geneva University Medical School, Geneva, Switzerland
  1. Correspondence to Dr Samia A Hurst, Institute for Biomedical Ethics, Geneva University Medical School, CMU/1 rue Michel Servet, CH-1211 Genève 4, Geneva, Switzerland; samia.hurst{at}unige.ch

Abstract

Background Decisions regarding Cardio-Pulmonary Resuscitation (CPR) and Do Not Attempt Resuscitation (DNAR) orders remain demanding, as does including patients in the process.

Objectives To explore physicians’ justification for CPR/DNAR orders and decisions regarding patient inclusion, as well as their reports of how they initiated discussions with patients.

Methods We administered a face-to-face survey to residents in charge of 206 patients including DNAR and CPR orders, with or without patient inclusion.

Results Justifications were provided for 59% of DNAR orders and included severe comorbidity, patients and families’ resuscitation preferences, patients’ age, or poor prognosis or quality of life. Reasons to include patients in CPR/DNAR decisions were provided in 96% and 84% of cases, and were based on respect for autonomy, clinical assessment of the situation as not too severe, and the view that such inclusion was required. Reasons for not including patients were offered in 84% of cases for CPR and in 70% for DNAR. They included absent decision-making capacity, a clinical situation viewed as good (CPR) or offering little hope of recovery (DNAR), barriers to communication, or concern that discussions could be emotionally difficult or superfluous. Decisions made earlier in the patient's management were infrequently viewed as requiring revision. Residents reported a variety of introductions to discussions with patients.

Conclusions These results provide better understanding of reasons for CPR/DNAR decisions, reasons for patient inclusion or lack thereof, and ways in which such inclusion is initiated. They also point to potential side-effects of implementing CPR/DNAR recommendations without in-depth and practical training. This should be part of a regular audit and follow-up process for such recommendations.

  • Clinical Ethics
  • Right to Refuse Treatment
  • Autonomy
  • Quality of Health Care
  • Living Wills/Advance Directives

https://doi.org/10.1136/medethics-2012-100699

Statistics from Altmetric.com

    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.

    Decisions regarding Cardio-Pulmonary Resuscitation (CPR) and Do Not Attempt Resuscitation (DNAR) orders are increasingly considered a necessary part of hospital practice in several countries, including Switzerland. However, such decisions remain demanding, and CPR/DNAR orders frequently misunderstood.1 ,2 Moreover, although most patients would like to be consulted on these decisions,3 and despite consensus that such inclusion is ethically required in most cases,4–6 including patients in these decisions is difficult.7 Such conversations can be emotionally laden and lead to conflict,8 and come up against public misconceptions on the efficacy of resuscitation.9 As a result, hospital doctors are often uncomfortable regarding whether, why, and how to include patients in CPR/DNAR decisions. Education on such issues is often lacking in hospital training.10

    To date, almost all studies designed to understand the practice of decisions regarding CPR/DNAR orders have been conducted in the United States of America. Little information is available in Europe,11 and only three studies have been conducted in Switzerland.1 ,11–13 Moreover, most previous studies of CPR/DNAR orders focused on their association with quantitative factors such as patients’ age14–21 or comorbid conditions,14–20 22–27 quality of life,14 or life expectancy.14 ,20 ,21 ,24 ,28 Arguments on which real life CPR/DNAR decisions are based in hospital practice have not been explored.

    Decisions regarding CPR/DNAR orders are, however, recognised as a frequent cause of ethical difficulty.29 In the General Internal Medicine Service of the Geneva University Hospital, a tertiary university teaching hospital in Switzerland, a CPR/DNAR order using a specific form is required for each patient. This was based on a position issued by the Clinical Ethics Committee (CEC) in 2001.30 This topic was addressed because according to the experience of committee members discussion with patients on their preferences regarding CPR remained exceptional, many patients were nevertheless the object of DNAR decisions, patients in palliative management often did not have a DNAR order, orders were often decided by a single person without discussion between doctors or the health care team, CPR/DNAR orders were seldom revised in light of changing circumstances, rarely documented and even more rarely explicitly justified in the patient's chart. The CEC recommended that the default position should be to resuscitate patients, except in cases where a competent and informed patient refused CPR, or when the patient was in end-of-life care, or where CPR would be considered futile. The definition provided for ‘end-of-life’ is ‘death is expected in the following days’. Futile was defined as: ‘any treatment which does not enable the restoration of a patient's vital functions, or enable him to recover an acceptable quality of life, as defined based on the patient's own appreciation criteria’. The CEC further recommended systematic discussion of CPR/DNAR orders for all patients between at least two doctors and information of the health care team, as well as systematic documentation and regular review of CPR/DNAR orders.30 Finally, the CEC recommended the inclusion of patients in such decisions unless they were incapable of decision-making, or in end-of-life care, or if CPR would be futile in their case. A further exception to patient inclusion is that a CPR order can be given without patient inclusion in cases where the disease is not serious and cardiac arrest is highly unlikely.

    In this setting, we conducted a study to assess the prevalence and factors associated with CPR/DNAR orders and with patient inclusion in these decisions,31 physicians’ justification for CPR/DNAR orders and their decisions regarding patient inclusion, and how discussions with patients had been initiated. In this paper, we report results from the analysis of open-ended questions on participants’ reported reasons for CPR/DNAR orders, their reasons for including or not including patients in these decisions, and the way in which they reported initiating these discussions with patients.

    Participants and methods

    Participants

    We identified residents based on a weekly screening of CPR/DNAR orders forms in six wards (approximately 100 patient beds) of the General Internal Medicine Department of the Geneva University Hospitals in Switzerland. We identified all available CPR/DNAR forms in the context of our quantitative study.31 We contacted residents for a short face-to-face questionnaire in a sample of cases containing approximately equal numbers (about 50 cases) of patients from four categories defined on the basis of the resuscitation order, and of whether or not the decision had been discussed with the patient. These groups were: discussed DNAR, undiscussed DNAR, discussed CPR, and undiscussed CPR. Cases were included consecutively into each category, as identified through the development of our quantitative database.31 Patient cases were included if the resident in charge of the patient was available in his/her office and consented to filling in a short face-to-face questionnaire.

    Data collection

    Questionnaire items were based on a literature review of factors associated with CPR or DNAR orders, and with patient participation in end of life decisions. Two versions of the questionnaire were used, taking into account whether the patient had been included in the CPD/DNAR decision or not. Survey items are further described elsewhere.31

    In all questionnaires, open-ended questions addressed justification provided for the CPR/DNAR order by the resident. In cases where a discussion with the patient took place, residents were asked the justifications for discussing the decision with the patient as follows: ‘What were your reasons to discuss [this] with the patient?’. They were then asked how that discussion had been initiated: ‘If [you initiated the discussion]; how did you start?’. In cases where such a discussion did not take place, residents were asked the justification for not discussing the decision with the patient as follows: ‘What were your reasons for not discussing [this] with the patient’?

    Closed-ended questions addressed whether justifications for orders had been discussed within the medical team, who had initiated the discussion when patients were included, and resident demographics. Questionnaires were administered face-to-face, and responses noted by the research assistant. The full questionnaire is available on request.

    To minimise recall problems, residents who consented to participation were asked to fill in the questionnaire 1–6 days after the patients’ admission to the ward. Data collection was open during 12 months, from April 2004 to May 2005.

    Protection of human participants

    Participation was voluntary and responses were made anonymous before analysis. Questionnaires and answers were kept strictly confidential, the research assistant was an advanced medical student with no power over residents, identities were kept from senior members of the research team who worked in the surveyed department, and this was made clear to respondents. This study was submitted to the chair of the hospital research ethics committee who designated it as quality control and exempted it from full ethics committee review.

    Data analysis

    Descriptive statistics were used for closed-ended responses. Responses to open-ended questions were transcribed and coded. As the aim of our study was to describe a phenomenon—residents’ justifications and approaches to CPR/DNAR decisions—we used a conventional content analysis approach in our qualitative data analysis, and approach considered appropriate ‘when existing theory or research literature on a phenomenon is limited’ as was the case here.32 The first step involved examining the data, breaking them down, and making comparisons and conceptualisations, which were then labelled with a set of codes. Codes for the participants’ justification for CPR/DNAR orders, their decisions regarding patient inclusion, and how discussions with patients had been initiated, were thus developed and refined, and grouped into categories. As answers to our open-ended items were relatively short, we reached theoretical saturation for this first level but were not able to pursue theory development beyond the content analysis stage. Earlier data were recoded using the main categories that emerged during later stages of the analysis. Quotations presented in this article are translated from the original French.

    Results

    Respondents

    Weekly screenings resulted in the inclusion of 1446 records (47% of the 2911 admissions during the study period). Of these, 21.2% contained a DNAR order, 61.7% a CPR order, and 17.1% gave no indication. From this sample, we selected 100 CPR orders, of which 51 had been discussed with the patient and 49 had not, and 106 DNAR orders, of which 56 had been discussed with the patient, and 50 had not, for the resident questionnaire. Almost all approached residents consented to participation (98%). Refusal was due to time constraints in three cases, and one resident answered an insufficient number of questions to be included. The 206 patients were treated by 61 different residents with postgraduate clinical experience ranging from 1 year to over 5 years. 89% of residents reported general training in ethics during medical school, and 31% specific training on CPR/DNAR orders. The mean number of cases per resident was 3.3.

    Reasons given for CPR/DNAR orders

    Justifications were given for 38% and 59% of CPR/DNAR respectively, and are outlined in table 1. Residents gave four major justifications for DNAR orders: important comorbid conditions, patient or family resuscitation preferences, patient age, poor prognosis or quality of life.31 Diseases most frequently reported as determining a DNAR order were cancer, mostly described as untreatable, advanced, or terminal, heart disease, or the presence of multiple disorders. For the CPR patients the residents indicated the same factors in reverse as justifications: patient younger age, patients’ resuscitation preferences, or causes for optimism such as an ongoing therapeutic plan. A good quality of life as assessed by residents was also a salient justification. Cancer and heart disease were also the most frequent diseases reported as determining a CPR order. In such case, however, cancer was rarely described as terminal, advanced, or untreatable. Reasons for CPR/DNAR orders were discussed within the health care team in 55% of DNAR and 32% of CPR orders.

    View this table:
    Table 1

    Residents’ justifications for CPR/DNAR orders

    Reasons given to include or not include the patient

    Justifications given for including—or not including- patients in decisions regarding CPR/DNAR orders are outlined in table 2. Reasons to include patients were based on routine (topic is always discussed), diagnoses other than cancer or central nervous system disease, changes or prospective changes in the patient's health status, previous occurrence of ‘close calls’, as well as ethical considerations (respect for patient autonomy).

    View this table:
    Table 2

    Residents’ justification for including, or not including, the patient

    Reasons given for not including patients included previous discussions with family members or in other healthcare teams, communication barriers or emotional difficulties, judgment that inclusion was superfluous, and central nervous system disease.

    A good quality of life or poor prognosis were included as justifications both to include and not to include patients in decisions regarding CPR/DNAR status.

    Introducing the discussion

    Descriptions of how the discussions to include the patient in CPR/DNAR decisions were initiated are outlined in table 3. Respondents reported starting by telling the patient that this discussion was always included in admission work-ups, by asking what the patient did or did not want from a stated list, by focusing on the patient's present state of health, or on the risk of it worsening. Other reported starting with a discussion of a recent health event, or previous advance care planning, and making the initial question about CPR/DNAR a follow-up. Participants also reported announcing that this would be a difficult discussion or asking patients if they had already deliberated, either alone or with their primary care doctor, on what they did or did not want. Discussions regarding DNAR orders were initiated by residents, first residents, and patients in 68%, 2%, and 16% of cases, respectively. Discussions regarding CPR orders were initiated by residents, first residents, and patients in 86%, 2%, and 8% of cases, respectively.

    View this table:
    Table 3

    Initiating the discussion

    Discussion

    Our study provides insight into residents’ reasoning regarding CPR/DNAR decisions and into why, and how, they include patients in these decisions. Such findings are important both to tailor post-graduate and continuing education to doctors’ perceptions and needs, and sometimes to revise or adapt ethical recommendations to fit with pitfalls of clinical practice.

    Residents used arguments based on respect for patient autonomy and decision-making capacity, as well as arguments based on their own clinical assessment of the situation, to justify including or not including patients in CPR/DNAR decisions. In view of efforts made to improve a situation where patient involvement was perceived as exceptional, this may represent progress. Other studies where recommendations for end-of-life care were implemented showed mixed results, with an increase in advance care planning33 and end-of-life management34 reported by some, but little effect on patient involvement reported by others.35 Some justifications, such as reports that previous refusals or an unstable clinical situation are a reason to discuss treatment preferences, or that these discussions happen because they always should, also suggest a heightened awareness of the importance of patient involvement.

    Residents, however, also reported several justifications for not including patients which may give cause for concern. Communication barriers were invoked, even when they were based on language or on an absent dental prosthesis. Decisions made at earlier stages in the patient's management tended not to be viewed as requiring revision. Finally, residents seem to operate with a much broader definition of a patient who is ‘too well’ or ‘too sick’ for DNAR to be discussed than the Ethics Council recommendation. They reported patients with a diagnosis of advanced cancer, but whose death was not expected in the next days, as sick enough for a DNAR order to be considered, and for it to be made without patient consultation. In reverse, their reports suggest that patients with many different kinds of clinical situations are deemed too well for their treatment preferences to be discussed.

    This discrepancy between residents’ thresholds and the Ethics Council's threshold for allowing a decision, either CPR or DNAR, without patient involvement, was also found in previous studies,36 ,37 including in Switzerland,13 and can be interpreted in several ways. First, residents may be rationalising situations when their true reason for not including patients is the anticipation of an emotionally difficult discussion. They may underestimate the number of patients who want to discuss DNAR status with their doctor, even when they state that they would find such a discussion distressing.3 Second, residents may underestimate the differences that can exist between their own assessment of the patient's interest, and this patient's wishes and priorities.31 Third, this discrepancy may be associated with the thresholds of medical indication and futility, with their attendant difficulties,38 and misuses.39

    Among the different thresholds proposed for futility, the Ethics Council recommendation took a position which required the intervention itself to lack the prospect of restoring vital functions, and any quality of life considerations to be based on the patient's own appreciation. Residents, in contrast, considered factors such as patients’ age, or likelihood of recovery could be included in the assessment of futility as regards CPR. This would be consistent with the extension of the ‘end of life’ label to patients with an incurable disease threatening their lives in the foreseeable future, rather than in the next few days. It would also be consistent with the fact that residents tend to evaluate patients’ quality of life themselves and include this consideration into CPR/DNAR decisions.31

    Residents reported using a variety of introductions to discussions to include patients in CPR/DNAR decisions. These included stating that the discussion was systematic, explaining the patient's present state of health, outlining the risk of getting worse, and asking what the patient wanted or presenting a default. Some discussions were initiated by following up on a complication or on a previous discussion, or asking if patients had previously thought about their treatment preferences. These reports illustrate a great diversity of strategies and experience. While some examples show sensitivity to context and to the difficulties associated with such discussions for both doctors and patients, some examples are concerning or might be misunderstood. Given public misconceptions of CPR, asking if a patient wants doctors to ‘start it up again’ following cardiac arrest may play into excessive expectations. Some examples of how discussions are started are technical, and cast the patient in an inappropriate expert role regarding CPR techniques. Other examples show asymmetry between CPR and DNAR orders in the manner in which discussions are initiated. While this study's sample and recruitment does not allow us to search for associations between discussion styles and CPR/DNAR orders, it is intriguing that patients told not to worry all ultimately had CPR orders, as did most of those told that this discussion is routine or that the discussion could be difficult. These exploratory findings raise questions as to whether such introductions cue the patient's answer. In a recent Canadian study, patients with CPR and DNAR orders to which they had personally agreed understood these orders very differently. Patients with DNAR orders ‘described resuscitation in graphic concrete terms that emphasised suffering and futility, and DN(A)R orders in terms of comfort or natural processes.’ Patients with CPR orders, in contrast, ‘understood resuscitation in an abstract sense as something that restores life, while DN(A)R orders were associated with substandard care or even euthanasia.’40 It is credible that doctors’ choice of words could influence such perceptions. An alternative interpretation could be that residents may adapt their wording to their sense of the patient's state of health, and write CPR/DNAR orders based on similar considerations. In either case, greater awareness would contribute to better conversations.

    Several of these aspects suggest that more in-depth as well as practical training of residents regarding how to reach CPR/DNAR decisions, as well as when and how to include patients, would be useful. Calls for such training have already been made elsewhere.6 ,10 Although training on CPR/DNAR could take place in many settings, it is a missed opportunity that the reasons for CPR/DNAR decisions were rarely discussed within the medical team. There are potential side-effects of implementing CPR/DNAR recommendations without simultaneous practical training. Although each of these points would require further study, patient reaction to discussions regarding their treatment preferences might be influenced by the manner in which they take place, varying from surprise, shock, or confusion, to information, respect and support. This is further reason why regular audit and follow-up of such policies should be welcome.41

    Our study has several limitations. Our questionnaire sample included only 206 cases, which were chosen with the purpose of recruiting as many different residents as possible. This allowed us to explore different views on CPR/DNAR orders and patient inclusion, as well as experiences with such discussions, but as a result our findings reflect diversity and salience within our sample rather than frequency in clinical practice. These cases were further sampled from a broader database of 1446 patient files. In collecting this larger database, we included only 47% of admissions to the study wards despite very frequent collections. The most likely explanations are that we undersampled patients whose stay was very short, either because they left the hospital, were transferred early to a different service, or died within a few days of admission. These patients are as likely to have been more severely ill, or less severely ill, leading the opposing sampling biases to cancel each other at least in part. A low percentage of patients (58 patients, 2% of the admissions) died during our study period in the six medicals wards of the study, but we could have missed patients who were transferred quickly to the intensive care unit or to rehabilitation hospitals. As previously outlined,31 we were dependent on the availability of residents and despite our efforts we may have selected residents who were more interested in ethical questions than those who might have tried to ‘escape’ the attention of the research assistant. Our sample included a number of temporary residents who were substituting and who are less experienced than their more stable colleagues employed for a period of several years during residency. However, this is likely to have been attenuated by our broad sampling strategy. As in other questionnaire studies, a bias could exist towards obtaining socially accepted answers. We tried to reduce this bias by guaranteeing complete confidentiality regarding respondents’ identity and their answers, especially towards the clinical hierarchy. That residents did report socially/professionally undesirable attitudes, such as failing to include patients because they did not dare to do so and admitted ignorance as to why orders and resuscitation preferences were not discussed with patients, indicates that this bias was low. Recall bias is also possible due to the delay between CPR/DNAR decisions and our interviews. Keeping the interval to less than a week, however, would tend to minimise this. An exception could be recall of the first sentence introducing discussion with patients. It is likely, however, that residents unable to recall their exact words would give what they deem to be a representative example of how they initiate such discussions. Finally, as with any exploratory single centre study, any generalisation to other contexts should be cautious.

    Conclusion

    These results provide better understanding of reasons for CPR/DNAR decisions, reasons for patient inclusion or lack thereof, and ways in which such inclusion is initiated. Some of our findings also provide cause for concern. Reasons to leave patients out of CPR/DNAR decisions included barriers to communication, previously made decisions, or a sense that the discussion was emotionally difficult or might be superfluous. Although the Ethics Council took a restrictive view of the cases in which futility would warrant a DNAR order without patient inclusion, residents seem to operate with a much broader definition of these cases, leading to more restrictive patient inclusion in DNAR decisions. Such results point to potential side-effects of implementing CPR/DNAR recommendations without simultaneous in-depth and practical training. This should be part of a regular audit and follow-up process for such recommendations.

    Acknowledgments

    The authors wish to thank Malik Babaker for his assistance in data collection, as well as all the busy residents who took time to answer our questions.

    References

    1. ↵
      1. Imhof L,
      2. Mahrer-Imhof R,
      3. Janisch C,
      4. et al
      . Do not attempt resuscitation: the importance of consensual decisions. Swiss Med Wkly 2011;141:w13157.
      OpenUrlPubMed
    2. ↵
      1. Fritz Z,
      2. Fuld J,
      3. Haydock S,
      4. et al
      . Interpretation and intent: a study of the (mis)understanding of DNAR orders in a teaching hospital. Resuscitation 2010;81:1138–41.
      OpenUrlCrossRefPubMedWeb of Science
    3. ↵
      1. Gorton AJ,
      2. Jayanthi NV,
      3. Lepping P,
      4. et al
      . Patients’ attitudes towards ‘do not attempt resuscitation’ status. J Med Ethics 2008;34:624–6.
      OpenUrlAbstract/FREE Full Text
    4. ↵
      1. Solbakk JH,
      2. Zoloth L
      . The tragedy of translation: the case of ‘first use’ in human embryonic stem cell research. Cell Stem Cell 2011;8:479–81.
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Weijer C
      . I need a placebo like I need a hole in the head. J Law Med Ethics 2002;30:69–72.
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      Swiss Academy of Medical Sciences. Resuscitation decisions 2008 March 11th 2011. http://www.samw.ch/dms/fr/Ethique/Directives/actuel/Reanimationsentscheidung_F_09/Reanimation_F_09.pdf (accessed 1 November 2012).
    7. ↵
      1. Albin RL
      . Sham surgery controls are mitigated trolleys. J Med Ethics 2005;31:149–52.
      OpenUrlAbstract/FREE Full Text
    8. ↵
      1. Myint PK,
      2. Rivas CA,
      3. Bowker LK
      . In-hospital cardiopulmonary resuscitation: Trainees’ worst and most memorable experiences. QJM 2010;103:865–73.
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. Groarke J,
      2. Gallagher J,
      3. McGovern R
      . Conflicting perspectives compromising discussions on cardiopulmonary resuscitation. Ir Med J 2010;103:233–5.
      OpenUrlPubMed
    10. ↵
      1. Pitcher D,
      2. Smith G,
      3. Nolan J,
      4. et al
      . The death of DNR. Training is needed to dispel confusion around DNAR. BMJ 2009;338:b2021.
      OpenUrlFREE Full Text
    11. ↵
      1. van Delden JJ,
      2. Lofmark R,
      3. Deliens L,
      4. et al
      . Do-not-resuscitate decisions in six European countries. Crit Care Med 2006;34:1686–90.
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      1. Perron N Junod,
      2. Morabia A,
      3. de Torrente A
      . Quality of life of do-not-resuscitate (DNR) patients: how good are physicians in assessing DNR patients’ quality of life? Swiss Med Wkly 2002;132:562–5.
      OpenUrlPubMed
    13. ↵
      1. Perron N Junod,
      2. Morabia A,
      3. De Torrente A
      . Evaluation of do not resuscitate orders (DNR) in a Swiss community hospital. J Med Ethics 2002;28:364–7.
      OpenUrlAbstract/FREE Full Text
    14. ↵
      1. Hakim RB,
      2. Teno JM,
      3. Harrell FE Jr.,
      4. et al
      . Factors associated with do-not-resuscitate orders: patients’ preferences, prognoses, and physicians’ judgments. SUPPORT Investigators. Study to understand prognoses and preferences for outcomes and risks of treatment. Ann Intern Med 1996;125:284–93.
      OpenUrlCrossRefPubMedWeb of Science
    15. ↵
      1. Stolman CJ,
      2. Gregory JJ,
      3. Dunn D,
      4. et al
      . Evaluation of the do not resuscitate orders at a community hospital. Arch Intern Med 1989;149:1851–6.
      OpenUrlCrossRefPubMedWeb of Science
    16. ↵
      1. Wenger NS,
      2. Pearson ML,
      3. Desmond KA,
      4. et al
      . Epidemiology of do-not-resuscitate orders. Disparity by age, diagnosis, gender, race, and functional impairment. Arch Intern Med 1995;155:2056–62.
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      1. Bedell SE,
      2. Pelle D,
      3. Maher PL,
      4. et al
      . Do-not-resuscitate orders for critically ill patients in the hospital. How are they used and what is their impact? JAMA 1986;256:233–7.
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      1. Vetsch G,
      2. Uehlinger DE,
      3. Zuercher-Zenklusen RM
      . DNR orders at a tertiary care hospital—are they appropriate? Swiss Med Wkly 2002;132:190–6.
      OpenUrlPubMed
    19. ↵
      1. Messinger-Rapport BJ,
      2. Kamel HK
      . Predictors of do not resuscitate orders in the nursing home. J Am Med Dir Assoc 2005;6:18–21.
      OpenUrlCrossRefPubMed
    20. ↵
      1. Goodlin SJ,
      2. Zhong Z,
      3. Lynn J,
      4. et al
      . Factors associated with use of cardiopulmonary resuscitation in seriously ill hospitalized adults. JAMA 1999;282:2333–9.
      OpenUrlCrossRefPubMedWeb of Science
    21. ↵
      1. Charlson ME,
      2. Sax FL,
      3. MacKenzie CR,
      4. et al
      . Resuscitation: how do we decide? A prospective study of physicians’ preferences and the clinical course of hospitalised patients. JAMA 1986;255:1316–22.
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      1. Ghusn HF,
      2. Teasdale TA,
      3. Boyer K
      . Characteristics of patients receiving or foregoing resuscitation at the time of cardiopulmonary arrest. J Am Geriatr Soc 1997;45:1118–22.
      OpenUrlPubMedWeb of Science
    23. ↵
      1. Wachter RM,
      2. Luce JM,
      3. Hearst N,
      4. et al
      . Decisions about resuscitation: inequities among patients with different diseases but similar prognoses. Ann Intern Med 1989;111:525–32.
      OpenUrlPubMedWeb of Science
    24. ↵
      1. Phillips RS,
      2. Wenger NS
      , SUPPORT investigators. Choices of seriously Ill patients about cardiopulmonary resuscitation: correlates and outcomes. Am J Med 1996;100:128–37.
      OpenUrlCrossRefPubMedWeb of Science
    25. ↵
      1. Lawrence VA,
      2. Clark GM
      . Cancer and resuscitation: does the diagnosis affect the decision? Arch Intern Med 1987;147:1637–40.
      OpenUrlCrossRefPubMedWeb of Science
    26. ↵
      1. Bacchetta MD,
      2. Eachempati SR,
      3. Fins JJ,
      4. et al
      . Factors influencing DNR decision-making in a surgical ICU. J Am Coll Surg 2006;202:995–1000.
      OpenUrlCrossRefPubMedWeb of Science
    27. ↵
      1. Miller DL,
      2. Gorbien J,
      3. Simbartl LA,
      4. et al
      . Factors influencing physicians in recommending in hospital cardiopulmonary resuscitation. Arch Intern Med 1993;153:1999–2003.
      OpenUrlCrossRefPubMedWeb of Science
    28. ↵
      1. Wenger NS,
      2. Pearson ML,
      3. Desmond KA,
      4. et al
      . Outcomes of patients with do-not-resuscitate orders. Toward an understanding of what do-not-resuscitate orders mean and how they affect patients. Arch Intern Med 1995;155:2063–8.
      OpenUrlCrossRefPubMedWeb of Science
    29. ↵
      1. Hurst SA,
      2. Forde R,
      3. Reiter-Theil S,
      4. et al
      . Ethical difficulties in clinical practice: experiences of European doctors. J Med Ethics 2007;33:51–7.
      OpenUrlAbstract/FREE Full Text
    30. ↵
      Conseil d’éthique clinique des HUG. L'ordre ‘Not To Be Resuscitated’ (NTBR) chez l'adulte et l'enfant. Geneva: Geneva University Hospitals Clinical Ethics Council, 2001.
    31. ↵
      1. Becerra M,
      2. Hurst SA,
      3. Junod Perron N,
      4. et al
      . ‘Do not attempt resuscitation’ and ‘cardiopulmonary resuscitation’ in an inpatient setting: factors influencing physicians’ decisions in Switzerland. Gerontology 2011;57:414–21.
      OpenUrlPubMedWeb of Science
    32. ↵
      1. Hsieh HF,
      2. Shannon SE
      . Three approaches to qualitative content analysis. Qual Health Res 2005;15:1277–88.
      OpenUrlAbstract/FREE Full Text
    33. ↵
      1. Hockley J,
      2. Watson J,
      3. Oxenham D,
      4. et al
      . The integrated implementation of two end-of-life care tools in nursing care homes in the UK: an in-depth evaluation. Palliat Med 2010;24:828–38.
      OpenUrlAbstract/FREE Full Text
    34. ↵
      1. Esteve A,
      2. Jimenez C,
      3. Perez R,
      4. et al
      . Factors related to withholding life-sustaining treatment in hospitalized elders. J Nutr Health Aging 2009;13:644–50.
      OpenUrlCrossRefPubMedWeb of Science
    35. ↵
      1. Castle N,
      2. Owen R,
      3. Kenward G,
      4. et al
      . Pre-printed ‘do not attempt resuscitation’ forms improve documentation? Resuscitation 2003;59:89–95.
      OpenUrlCrossRefPubMedWeb of Science
    36. ↵
      1. Harris D,
      2. Davies R
      . An audit of ‘do not attempt resuscitation’ decisions in two district general hospitals: do current guidelines need changing? Postgrad Med J 2007;83:137–40.
      OpenUrlAbstract/FREE Full Text
    37. ↵
      1. de Vos R,
      2. Koster RW,
      3. de Haan RJ
      . Impact of survival probability, life expectancy, quality of life and patient preferences on do-not-attempt-resuscitation orders in a hospital. Resuscitation committee. Resuscitation 1998;39:15–21.
      OpenUrlCrossRefPubMedWeb of Science
    38. ↵
      1. Kite S,
      2. Wilkinson S
      . Beyond futility: to what extent is the concept of futility useful in clinical decision-making about CPR? Lancet Oncol 2002;3:638–42.
      OpenUrlCrossRefPubMedWeb of Science
    39. ↵
      1. Curtis JR,
      2. Park DR,
      3. Krone MR,
      4. et al
      . Use of the medical futility rationale in do-not-attempt-resuscitation orders. JAMA 1995;273:124–8.
      OpenUrlCrossRefPubMedWeb of Science
    40. ↵
      1. Downar J,
      2. Luk T,
      3. Sibbald RW,
      4. et al
      . Why do patients agree to a ‘do not resuscitate’ or ‘full code’ order? Perspectives of medical inpatients. J Gen Intern Med 2011;26:582–7.
      OpenUrlCrossRefPubMed
    41. ↵
      1. Gabbott D,
      2. Smith G,
      3. Mitchell S,
      4. et al
      . Cardiopulmonary resuscitation standards for clinical practice and training in the UK. Resuscitation 2005;64:13–9.
      OpenUrlCrossRefPubMedWeb of Science

    Footnotes

    • Contributors SH wrote the first draft. All authors had full access to all the sources in the study. SH, MB, NJP, SC and BE participated in the conception and design of the study. SH, MB, NJP, SC, AP and BE participated in the  analysis and interpretation of data, and revision of the manuscript for important intellectual content. All authors approved the final version. SH is the guarantor.

    • Funding This work was funded by the Geneva University Hospitals Quality Improvement Projects. The views expressed here are the authors’ own and do not reflect the position of the Geneva University Hospitals. SAH was funded by grants from the Swiss National Science Foundation (3233B0-107266/1-2; PP00P3_123340). The funding sources were not involved in the study design, the collection, analysis, or interpretation of data, in the writing of the report or in the decision to submit the paper for publication.

    • Competing interests None.

    • Ethics approval Central Ethics Committee of the Geneva University Hospitals.

    • Provenance and peer review Not commissioned; externally peer reviewed.