Comparison of medical and nursing attitudes to resuscitation and patient autonomy between a British and an American teaching hospital

Pergamon

Plh S0277-9536(97)00187-1

Soc. Sci. Med. Vol. 46, No. 3, pp. 415-424, 1998 © 1997 Elsevier Science Ltd. All rights reserved

Printed in Great Britain 0277-9536/97 $19.00 + 0.00

COMPARISON OF MEDICAL AND NURSING ATTITUDES TO RESUSCITATION AND PATIENT AUTONOMY

BETWEEN A BRITISH AND AN AMERICAN TEACHING HOSPITAL

M I C H E L L E MELLO'* and CRISPIN J E N K I N S O N 2

'Department of Health Policy and Administration, School of Public Health, University of North Carolina at Chapel Hill, Chapel Hill, NC, U.S.A. and 2Health Services Research Unit, Department of

Public Health and Primary Care, University of Oxford, Oxford, U.K.

Abstract--In the last 30 years, cardiopulmonary resuscitation (CPR) has evolved from an intervention indicated only in cases of acute insult to an otherwise healthy body to a default measure employed in virtually all cases of cardiac failure. The high cost and low efficacy rate of CPR has provoked questions about the moral and economic wisdom of its routine use, particularly for elderly patients with serious comorbidity. This paper presents the results of a comparative study of decision making practices concerning "Do-Not-Resuscitate" (DNR) orders in British and American hospitals. Thirty-four physicians and nurses in one American and one British hospital were interviewed about their decision making practices. Qualitative methods of data analysis were employed. The study revealed that while the Amer- ican and British hospitals had adopted similar formal protocols for DNR decision making, in practice the British physicians often made DNR decisions unilaterally, whereas the American physicians sought the patient's or surrogate's consent in every instance, even where it was not legally required. The British decision making model enables physicians to reduce the inappropriate use of resuscitation, but at the expense of patient autonomy. In contrast, the American approach fully respects patient autonomy, but except in cases of medical futility grants physicians no authority to refuse to render treatments that are in their judgment contraindicated. © 1998 Elsevier Science Ltd. All rights reserved

Key words--CPR, resuscitation, clinical ethics, autonomy, decision making

INTRODUCTION

In the last 30 years, cardiopulmonary resuscitation (CPR) has gradually evolved from an intervention

indicated only in cases of acute insult to an otherwise strong, healthy body to a default measure

employed in virtually every in-hospital arrest. In the

United States today, 55% of all patients who receive CPR are elderly (United States Congress, 1987), many of them with serious underlying

comorbidity. For a large percentage of these patients, resuscitation has lost its restorative func-

tion and is being used merely to postpone death.

This trend is of great concern to health care providers and health care payors alike. Physicians are

increasingly uneasy with a legal environment which seems to require them to obtain the patient's or surrogate's consent for a "Do-Not-Resusci ta te"

*Author for correspondence: 10 Livingston St, B32, New Haven, CT, 06511, U.S.A.

tWhile the term "DNR" was used in both the American and the British hospitals, the phrase is ambiguous since it refers to both successful resuscitation and the resuscitation attempt itself. Some hospitals prefer the term "Do Not Attempt Resuscitation (DNAR)", and it is the attempt which is meant by the term "DNR" in this paper.

( D N R ) t order in every instance, even in cases where the physician believes that attempting to resuscitate the patient would be profoundly wrong. There is a growing feeling that physicians have lost control over a decision which is essentially medical, though it is also imbued with moral elements. The use of life-prolonging technologies by the dying elderly has also become a concern to those seeking to control health care costs. The 12% of the popu- lation that is age 65 and over accounts for more than one-third of America's total health care expen- ditures (United States Senate, 1989), and 27-30% of Medicare payments each year are for the 5 -6% of beneficiaries who die in that year (Emanuel and Emanuel, 1994). Given these concerns, it is perhaps time to reevaluate the routine use of CPR in the hospital.

This paper discusses the findings of a comparative study of how D N R decisions are made in American and British hospitals. British decision making practices are of interest because CPR is not used as frequently as in the United States and because the balance of power in the physician- patient relationship is different. While the United States has seen a fundamental reorientation of the medical relationship from paternalism to a relationship centered on the patient's autonomy, in Britain

415

416 Michelle Mello and Crispin Jenkinson

this transition has not yet occurred in full. Patient autonomy and consent are not held as sacred and physicians retain a larger degree of decision making authority and clinical freedom. It is interesting to ask how this dynamic has structured DNR decision making and kept the use of CPR from becoming as routine as in the U.S.

BACKGROUND

Studies consistently demonstrate that CPR has a very low efficacy rate. A recent multicenter survey and two meta-analyses of CPR survival studies which synthesized a large proportion of the over one hundred survival studies published in the last 30 years suggest that in-hospital CPR has an aver- age initial success rate (i.e. the patient survives the arrest incident) of around 38% (Tunstall-Pedoe et al., 1992; Von Gunten, 1991; Schneider et al., 1993). About 25% of these initial survivors die before being discharged; thus, only about 15% of all patients who receive in-hospital CPR (or 75% of the initial survivors) ever leave the hospital. Von Gunten points out that even this low figure may give an exaggerated impression of the efficacy of CPR because many patients who were considered poor candidates for CPR were given DNR orders and were therefore not included in the sample groups.

In addition to having an 80 to 90% failure rate (measured by survival to discharge), CPR is expens- ive. Vrtis' analysis of the cost of a CPR program in a community hospital, which calculated both direct costs (the costs of maintaining a CPR program and actually performing CPR) and indirect costs (the costs of ongoing care for patients who survive CPR), estimated the cost of CPR to be $7002 per patient who received a resuscitative attempt and $60,327 per patient who survived the incident (Vrtis, 1992). Isolating the indirect costs, Vrtis found that the cost of care per patient who survived initially but died before discharge averaged $9476 with a mean post-resuscitation length of stay of 9.9 days, and the cost of care for those who survived to discharge averaged $19,430 per person with a mean post-resuscitation length of stay of 20.3 days. If hospitals were to decrease the use of CPR in cases where its appropriateness is questionable, the at- tendant decrease in costs would derive primarily from lower indirect costs of CPR (since the direct costs of maintaining a CPR program would still be incurred). Vrtis' study indicates that these indirect costs constitute the bulk of the per-patient cost of CPR.

Given its relatively low efficacy and high cost, why is CPR used so routinely in the United States? Perhaps the answer lies in the new emphasis on patient autonomy in the physician-patient relationship. The reorientation of this relationship to trans- fer the locus of control to the patient has meant

that patients or their surrogates must now be consulted and given full information about their medical options and permitted to decide for themselves what is "medically appropriate" and in their "best interest" (Crimmins, 1993). The implication for CPR has been the introduction of the DNR order, which conveys the patient's consent for non-intervention. Legislation and hospital policies recogniz- ing the legal status of DNR orders have effectively codified the use of CPR as a default measure to be used in the absence of a properly executed order by the patient to the contrary. Few patients execute DNR orders, so resuscitation is attempted in nearly all cases.

Patients in the United States and their families have been aggressive in exercising their right to be consulted about DNR orders. A study of American patients found that 60% wished decisions about resuscitation to be shared between medical staff and family and only 40% felt that the physician alone should make the decision (Stolman et al., 1990). Interestingly, studies show that British patients are less interested in exercising their autonomy in this area. In one study of elderly British patients, over half claimed that the leading role in DNR decision making should be taken by medical and nursing staff, one-third claimed that the patient should take the leading role, and the remainder thought the decision should be made jointly by the patient and the medical staff (Gulati et al., 1983). Overall, American patients and their relatives are more likely than their British counterparts to want to participate in medical decisions, to scrutinize the actions of their physicians, to complain and to liti- gate when they believe their rights or interests have been abridged.

Although autonomy is the centerpiece of the modern medical relationship in the United States, in practice it is impinged upon by the ability of physicians to override a patient's demands for treatment on the ground that the treatment is medically futile. No ethical principle or law has ever required physicians to assent to treatment that is of no potential benefit to the patient. The concept of medical futility has two components, one quantitative and one qualitative. The quantitative element refers to the statistical likelihood that a treatment will produce its intended medical effect. Thus, CPR could be considered futile in the quantitative sense if there was a very low probability that it would result in the restoration of normal cardiac rhythm and breathing. The qualitative component of futility is a measure of whether or not CPR has any medical benefits to the patient, which is a more complex question than whether it has any medical effects. Benefits have to do with the impact of a treatment on the patient's quality of life, and consequently they are difficult to quantify. While the assessment of effects is essentially a question of medical science and is therefore best made by doctors, the assess-

Comparison of attitudes toward CPR and patient autonomy 417

ment of benefits involves value judgments which should be made by the patient.

The principle of patient autonomy requires that when a physician believes that CPR might be successful but should be foregone because it would result in an unacceptable quality of life for the patient, an obligation exists to obtain the patient's or family's permission for a DNR order. In contrast, when a physician has decided that CPR is futile in the quantitative sense, his/her only obligation is to talk with the patient or family with the aim of securing an understanding of the decision that has already been made. Obtaining the patient's or surrogate's consent is not required because the decision is a technical one based on medical exper- tise (Tomlinson and Brody, 1988).

In the United States, despite the injunction that all competent patients must be consulted about DNR orders unless CPR is judged to be futile in the quantitative sense, discussions to obtain informed consent do not always take place (Vance et al. , 1993; Bedell and Delbanco, 1984). In one hospital studied, only 19% of patients with DNR orders had been consulted about the order, though 86% of the patient group was competent to participate in the decision (Bedell and Delbanco, 1984). Furthermore, only 33% of the families had been consulted, and only fifteen of the 157 physicians surveyed said they discussed resuscitation with their patients (Bedell and Delbanco, 1984). Many physicians are dissatisfied with the requirement that patients or relatives be consulted in every instance, because it makes DNR orders "hostage to patient and family ambivalence" (Baker, 1993) and sometimes forces physicians to deliver care that goes against what their conscience tells them is right.

It is perhaps less surprising that patient consultation about DNR orders does not occur on a widespread basis in Britain, either. In one study of a teaching hospital, only 6% of competent patients with DNR orders had taken part in the decision (Dickinson and Sabo, 1991a), and in another study only three of 627 patients who died without a resuscitation attempt had documented evidence that this had been discussed with the patient or family (Keating, 1989). At least two studies have found evidence that junior physicians operate informal decision making systems in which some patients are designated "not-for-resuscitation" in the absence of a DNR order (Dickinson and Sabo, 1991b: Aarons and Beeching, 1991).

Health care providers have been vocal about their discomfort with the current state of CPR util- ization. In one survey of American physicians, 58% disagreed with the practice of providing CPR routinely in in-hospital arrest situations and 95% believed that there exist some medical conditions for which in-hospital CPR is futile (Miller et al., 1993). A survey by Baker, Dersch and Fein found that physicians object to being required to adminis-

ter futile CPR by default in cases where neither the patient nor the surrogate requested it, because it is not only medically useless but, since it was not requested, it also lacks symbolic meaning for the patient or family (Baker e t al. , 1995).

METHODS

A qualitative study of DNR decision making was conducted from May to December 1994 involving in-depth interviews with 17 physicians and 17 regis- tered nurses from two hospitals, one in the United States and one in England. Interview participants were recruited with the aid of the division chief and a nurse manager in each hospital's geriatric care division. Each participant was informed that responses would be kept confidential and was interviewed for 30-50 minutes using a semistructured format. Interviews were conducted in a private room in the hospital by the first author, who received prior training in interviewing technique. Topics covered in the interviews included respondents' knowledge of the clinical outcomes of CPR for elderly patients, procedures for making decisions about resuscitation for their patients, and beliefs about the appropriate use of CPR, their obligations to consult patients about DNR orders, and the uti- lity of formal DNR policies. All conversations were tape recorded and fully transcribed.

Transcripts were analyzed using a thematic content analysis method adapted from Glaser and Strauss' "grounded theory" approach (Glaser and Strauss, 1967) and used in health care research by Burnard (Burnard, 1991). This method involves making a detailed and systematic inventory of the issues addressed in interview responses and generat- ing theory by noting common themes. Transcripts are read through closely and notes are made on general themes within each transcript (for example, "hospital DNR policy is helpful" or "CPR is futile in some cases"). Transcripts are then read again and all aspects of the content are categorized into an exhaustive system of headings. The list of headings is then organized into hierarchical groupings and refined to eliminate near-duplications. In order to check the validity of the category system and guard against researcher bias in the selection of categories, two colleagues are then asked to read a randomly chosen sample of the transcripts and independently generate their own category systems. The three lists of categories are then discussed and adjustments are made as necessary. Transcripts are then read a third time alongside the final category list and fully coded into the agreed-upon headings. Responses from different interviews which fall under the same heading are then juxtaposed. The comparison of responses under each heading enables the researcher to generate theoretical prop- erties of the category. For example, for the category "hospital DNR policy is helpful", the researcher


can identify the range of opinions on the issue, the degree of consensus, the reasons underlying responses, and the implications for clinical practice. A theory concerning physicians' and nurses' views of hospital DNR policies can be developed based on the responses. The generalizability of the theory can then be considered.

In an effort to maximize the degree of compar- ability of the American hospital and the British hospital studied, hospitals with similar modes of financing and administration were selected. The British hospital is a National Health Service (NHS) Trust hospital in southeastern England and the American hospital is a Veterans Administration (VA) Medical Center in the southeastern United States. Both are teaching hospitals affiliated with major universities. As an NHS hospital, the British hospital is publicly owned and funded. However, since 1993 the hospital has been an NHS Trust, a self-governing public corporation within the NHS (Laing, 1994). Nearly all of the patient care ren- dered in the hospital is free at the point of delivery. For the American case study, a VA hospital was selected because the VA system is the closest thing the United States has to a national health service.

In selecting physicians and nurses to recruit for the interviews, identifying those likely to be most knowledgeable about resuscitation practices at their hospital took priority over selecting sample groups in each country that would have the same composition in terms of medical specialty and level of experience. The geriatrics division chief and a nurse manager at each hospital were consulted to deter-

mine which physicians and nurses would be most knowledgeable.

RESULTS

A total of 34 informants were interviewed. At the British hospital, nine physicians and eight nurses were interviewed. At the American hospital, the sample consisted of 10 physicians and seven nurses. None of the physicians contacted for interviews refused to participate, but four nurses declined to participate because they were too busy. Both the British and the American sample groups of nurses were the most senior nurses on their wards and the two groups had roughly the same qualifications (RN) and levels of experience (the median number of years of practice was 10 years for the British group and 8 years for the American group). The British nurses all worked on the geriatrics ward of the hospital; the American nurses all worked on the VA Hospital's Extended Care and Rehabilitation Center, a long-term care facility located within the hospital.

The two samples of physicians differed somewhat in composition. The American group consisted entirely of residents and fellows (the median number of years of practice was 3 years), while the British group consisted of 4 senior house officers, 3 senior registrars, and 2 consultants (the median number of years of practice was 8 years).

Informants were queried about a variety of issues related to their practices and beliefs concerning CPR. Responses are summarized in Table 1.

Table 1. Summary of responses a

(% giving a "yes" response) American British

Physicians Nurses Physicians Nurses

Would never override a DNR order Would never override patient's wish to receive CPR Felt obligated to provide futile CPR if patient requests it

lmportance ofhonoringpat~nts' w~hes 40 100

40 100 89

40 b 56

Consulting patients about theft resuscitatwn status Felt "great responsibility" 80 to consult patients Felt they need only consult 90 very sick patients Satisfied with current 90 amount of patient consultation about CPR in their hospital

86

56

78

67

Hospital DNR poliey Level of knowledge of Low Medium Medium hospital DNR policy Level of support for Low High High hospital policy Perception of level of High High Low compliance with policy by physicians in their hospital

100

I00

38

75

83

High

High

Low

aSmall sample size precluded tests of statistical significance. blndicates question not asked to this group.


Importance of honoring patients' wishes

Respondents ' beliefs about the importance of honoring patients' wishes about resuscitation were explored by asking whether they felt it was ever appropriate to override a patient 's wish to receive C P R or to be "not-for-resuscitation". Only one of the American physicians stated that she did not always honor D N R orders. However, six of the other nine physicians stated that they could envision a circumstance in which it would be appropriate to ignore one. The most commonly mentioned circumstances were if the patient 's situation was very different from the scenario he/she envisioned when he/she executed the D N R order, and if the patient was in an acute episode that the physician believed could easily be resolved, as in the following anecdote:

This one patient was a DNR and never wanted to be on a mechanical ventilator, and we put him on it because we thought that we could get him over his acute illness. We wanted to give him 48 hours and see what happened. In the end we ended up extubating him so that when we extubated him he just died. That was horrible for the wife; she was just so upset. She felt guilty because she had put him through something he didn't want. But on the other hand, medically, we felt that we could get him through it, and they didn't understand [American junior physician number 3 (AJP3), 11. 326-343].

Five of the American nurses stated that they could not think of any situation in which it would be justifiable to override a D N R order.

All of the British physicians and nurses stated that they always followed D N R orders, and all of the physicians and five of the nurses said they could not envision a circumstance in which a competent, non-depressed patient's wish to be D N R should not be respected. The consensus in favor of honoring D N R requests was quite different from the consensus of the American physicians that D N R requests could be overlooked in some situations.

Respondents were also asked their opinion about honoring a patient 's request to have full resuscitative efforts made. All of the American nurses said that such requests should always be honored. Physicians were more divided on the matter. Four felt that a patient 's request to receive CPR could legitimately be denied if CPR was deemed to be medically futile, four felt that it could not be denied, and two were uncertain.

Only one British physician believed it could ever be justifiable to override a competent patient's wish to receive CPR. Some British physicians felt very strongly indeed about respecting patients' wishes, as is revealed in this comment by a junior doctor:

Oh, if they're of sound mind, they might have every can- cer under the sun and they might be about to drop dead next week. If they said to me, "I want you to resuscitate me", then I'm sorry, but that's what I'm going to do, really [British junior physician number 1 (BJPI), 11. 219- 221].

Perceived obligations to provide futile CPR

The majority of the American and British physicians interviewed had strong views about whether or not they were obligated to provide futile CPR. Four of the American physicians felt obliged to provide CPR for any patient who wanted it, regard- less of its futility. One American physician explained, " I f I think it's futile, no, I don' t think it should be done. But do I do it? Sure. I think the patient has the right to make that determination at this point in our system" (AJP4, 11. 185-87). Three of the four American physicians expressed a wish that physicians be given the authority to declare a treatment futile and withhold it. Apparently none of these four physicians knew that they already pos- sessed this authority. In addition to the wealth of medical literature supporting the idea that physicians are not obligated, legally or morally, to render futile treatment, there is a hospital policy at the VA Medical Center which explicitly states that physicians are not required to provide or even offer futile treatments.

The four American physicians who were aware that they were not obligated to render futile care appeared to feel very secure in their position. One said,

I feel very comfortable not coding someone if 1 feel like it's totally futile. I feel totally comfortable. 1 know most physicians don't, but in the [Intensive Care] Unit, many times we just didn't "'code" them [provide CPR]. Because it was really futile, and the family can't understand (AJP3, 11. 381-384).

Another physician related an anecdote of having refused to continue with futile resuscitation attempts and appeared to be confident that the right decision was made:

I remember as a medical student we had this lady who was in florid kidney failure. The resident was trying to get [the family] to sign a DNR, but they refused. Then she coded three or four times that one night. And finally the resident picked up the phone again and said, "Look, I'm not doing this anymore." And that was that. And I agreed with it completely, because, you know, that's ridiculous (AJP2, 11. 243 249).

Of the British physicians, all of the junior physicians said they would feel obligated to give futile CPR, but four senior doctors stated that they would not. The contrast between senior and junior doctors on the question of futility is interesting. Those physicians who supported a physician's right to withhold futile treatments emphasized two factors: that the treatment was completely useless and a waste of energy and resources, and that it was unfair to patients to offer them futile interventions and raise their hopes unduly.

Consulting patients about their resuscitation status. British and American physicians differed in the degree to which they involved patients in decisions to withhold CPR. One senior British physician who had worked in both the United States and England


was able to contrast the tendency of American physicians to offer CPR to all patients to the practice of British physicians to offer it only to those for whom it might be beneficial. He stated that not offering it to every patient is

very much different from what we used to do when I was in North America. But I think that there are some people who one is under no obligation to offer what is in essence a useless treatment. When you ask people, and if they say "no" they feel that they have deprived their morn of something that could have been useful, then 1 think you put them under an unfair burden [British senior physician number 1 (BSP1), 11. 127-132).

Thus, by eliminating CPR as one of the treatment options patients and families had to choose among, this physician felt he was easing the burden of decision making for them. While some American physicians commented on this advantage of not offering CPR to every patient, they did not make it their practice to offer CPR selectively. Although both the British and the American respondents exhibited a great deal of respect for patients' right to make autonomous choices about their course of treatment, the British respondents believed that autonomy did not extend into the realm of futile interventions.

Both the American and the British respondents reported that most physicians in their hospital did not make it a practice to ask patients about their preferences for resuscitation. One American physician said,

Most patients, it's not discussed with them. I just did the [Intensive Care] Unit last month, and it was amazing how many times the general medicine resident hadn't talked to the patient about whether they wanted to be resuscitated or not (AJP3, 11. 205 207, 211 214).

Given that it is not discussed with many patients, and that the default option is for patients to receive CPR, it is logical to assume that some patients in the American hospital who did not have a D N R order on the chart were designated "for resuscita- t ion" by default rather than because they had actually expressed a preference to be so. Discussions with patients about their wishes were relatively infrequent in the British hospital as well. A British doctor remarked, "I think people might pay lipser- vice to having D N R orders discussed with patients, but in reality that happens very rarely" (BSP2, 11. 192-193).

The British doctors ' sense of their responsibility to talk to patients about their CPR status was not as strong as the American doctors'. Eight of the 10 American physicians said that they believed doctors had a great responsibility to consult patients, as compared to only five out of nine British physicians. A British physician attributed this lower sense of responsibility to a perception that British patients do not often assert their right to make their own treatment choices. He observed that "the British, compared to other nationalities, are much

more accepting of what doctors say and a lot of British people will say, 'Oh, whatever you say, Doc tor ' " (BJP2, 11. 168-170).

Both the British and the American physicians believed that their responsibility to consult patients about CPR extended only to very sick patients and those for whom cardiac arrest is likely to become an issue. Nine out of 10 American physicians and seven of nine British physicians stated that it was not their practice to discuss CPR status with all of their patients, but only with select patients. The primary reason the American doctors gave for not discussing CPR with all patients was that many of their patients were relatively young and healthy, and it was difficult for these patients to imagine themselves in a cardiac arrest situation. In contrast, over half of the British respondents were full-time geriatricians, and their primary reason for not discussing it with everyone was that it was upsetting to patients. One British nurse commented that patients tend to think, " M y God, where have I come? What are they planning to do to me?" [British nurse number 3 (BN3), 11. 233].

While nearly all of the American respondents were satisfied with the current degree of patient consultation about CPR status in their hospital, three British physicians and two British nurses felt that CPR was not being discussed enough. One American nurse objected to the way the subject of D N R orders was broached by doctors sometimes:

It was an admission that came in [on] a change of shift and he said, "'If your heart should stop beating, would you want us to get it started again by using machines?" And it was kind of an abrupt question and the guy had just gotten settled in the room and it wasn't the appropriate time at all to be talking about this stuff. The guy goes, "Well, yeah, I suppose". Bingo! You know, he had a code order. I think a patient needs to be settled in a little bit, kind of get [over] the stress of the change of environment and kind of get used to the place [American nurse number 2 (AN2), 11. 233 240).

Several British respondents expressed the view that although discussions with patients about CPR were still fairly rare, British physicians were moving towards discussing it more frequently and openly. However, one British physician who had worked in the U.S. explained that British doctors would never go to as great lengths as their American counterparts to involve patients in CPR decisions:

It's not like in the States where it's almost a reflex. Everybody is asked about CPR, always asked. I never saw physicians in the States make the decision on their own bet, no matter how terrible the situation was. I think that in the style of practice in this part of the world, there will always be some room left for clinical judgment. Rather than having to do things because it's in the algorithms or there's a law book saying you should do it. I don't think we'll ever go that far. And I think that's probably just as well (BSP1, 11. 383-390).

It is important to note that the American respondents all stated that no D N R orders were entered in their hospital without discussions having taken


place with the patient or the patient 's surrogate, unless CPR had been deemed medically futile. Thus, those patients with who D N R orders were not discussed remained "for-resuscitation" (unless a D N R order was entered on the grounds of futility). The interviews yielded no support for the finding in the medical literature (Vance et al., 1993; Bedell and Delbanco, 1984) that a significant percentage of D N R orders in American hospitals are entered without the patient 's consent.* In contrast, the British respondents indicated that in their hospital, a large percentage of D N R orders were written by doctors without consulting the patient or the patient 's family. A junior physician described this problem:

It sounds really awful to say, but I've been in hospitals where the nursing staff will come up to one and say--and this has happened to me even as a house officer months, just merely months after qualification--the nurse will come up to me and say, "Do you think this patient should be resuscitated?" My immediate response to that has always been, "I'm sorry, but it's not for me to make that sort of decision." Now that's just speaking for myself. Now if these nurses ask me questions like that, I'm sure they ask others of my rank questions like that, and who's to say that some of them don't give them answers? (BJP1, 11. 274-284).

Hospital DNR policies

Respondents were queried about their awareness of and attitude towards their hospital's written policy on DNR orders. In the United States, as of 1988 all hospitals have been required by the Joint Commission on Accreditation of Health Care Organizations to have formal, written D N R policies (Schecter, 1994). The VA Medical Center's policy on life-sustaining treatmentst states that such treatments may be withheld or withdrawn under the following circumstances: (1) at the oral or written request of a competent patient; (2) as specified by a valid advance directive when the patient lacks decision making capacity; (3) at the request of the surrogate decision maker on behalf of an incompetent patient; or (4) when the treatment is considered medically futile. When a treatment is futile, it need not be offered to the patient. Pursuant to the man- dates of the Patient Self-Determination Act of 1990, the policy states that patients will be encouraged to

*It should be noted that the Bedell and Delbanco study was conducted in 1984, before the passage of the Patient Self-Determination Act and several state DNR laws and at a time when fewer American hospitals had formal DNR policies. It is possible that the authors would have found different results in today's policy environment. However, the 1993 Vance, Ciancio and West study indicates that these policy changes have not eliminated the practice of writing DNR orders without the patient's expressed consent.

tThis policy was an internal memorandum and cannot be referenced for reasons of confidentiality.

:~This policy was also an internal memorandum and cannot be referenced.

complete advance directives and will be given information about their right to refuse treatment upon admission to the hospital.

British law does not require hospitals to have formal D N R policies, but in December 1991 the Chief Medical Officer issued a directive to all Consultants in England and Wales stating that clear policies should be formulated on the making of D N R decisions and that Consultants have a responsibility to ensure that their DNR policy is understood by all members of the medical staff, particularly junior doctors (Saunders, 1992). Pursuant to this directive, a number of hospitals have drafted their own policies on DNR decisions (regrettably, there are no published studies indicating the percentage of hospitals that have done so). Guidelines produced by the Royal College of Physicians (Williams, 1993), the British Medical Association/Royal College of Nursing/U.K. Resuscitation Council (British Medical Association, 1993), and British Medical Journal (Doyal and Wilsher, 1993) have been influ- ential in determining the content of hospital policies. The salient features of these guidelines are the stipulation that patient consent is not required for D N R orders in all cases and the idea that the views of the family should be considered but have no real legal or moral force. Ultimately the decision is based on the physician's medical judgment of the risks and benefits of CPR.

The most important provisions of this British hospital's policy on DNR orders$ are (1) the overall responsibility for a D N R decision rests with the consultant (the senior, specialist physician) in charge of the patient's care; (2) a consultant may enter a DNR order without consulting the patient only if the patient is irreversibly close to death in the short term, resuscitation presents an unacceptably high probability of death or severe brain damage, or the patient already has irreversible and severe brain damage; (3) when a senior clinician thinks that a D N R order is warranted for reasons other than the three listed above, the patient's informed consent must be obtained; and (4) consultants may choose to discuss DNR orders with patients' relatives, but relatives have no legal rights in the treatment of adult patients.

Overall, the American nurses were far more fam- iliar with their hospital's policy than the American physicians. Three of the 10 physicians did not know that the hospital policy existed. All seven nurses knew that patients were asked about advance directives when they were admitted, but five of the 10 physicians were quite surprised to learn this. When respondents were asked to summarize the main provisions of the hospital policy, both nurses and physicians gave incomplete descriptions, mentioning only those provisions which were relevant to their own job responsibilities.

Although none of the American physicians or nurses objected to the principle of having a hospital


policy on DNR orders, opinions of the content and implementation of this particular policy were mixed. Nurses were in general very supportive of the policy, noting that it gave the patient and family a chance to express their views. Most physicians found fault with the policy. Their most common complaint was that the admissions clerk did not do a thorough job of discussing advance directives with patients upon admission.

The British respondents were more knowledgeable about and supportive of their hospital's DNR policy than were the American respondents. Most respondents knew not only that the policy existed, but also when it was adopted, where the guidelines came from, and what the content of the policy was. However, some gaps in British respondents' knowledge of the policy were apparent regarding the circumstances in which patients had to be consulted about DNR orders. Several were not aware that the policy provided any specific guidelines as to when the patient should be consulted, and believed it was up to them to decide which patients it would be appropriate to discuss it with. Most of the British nurses were quite accurate in their description of the policy. Both physicians and nurses in the British hospital approved of the hospital's DNR policy. The nurses commented that the policy was helpful and had just the right amount of flexibility, while a physician stated that the guidelines "have been a great help in ensuring that decisions get made by appropriately senior people at an appropriate stage" (BSP4, 11. 247-249).

There appeared to be a high degree of compliance with the hospital policy in the American hospital. While the discussions with the admissions clerk were by most accounts inadequate, the consensus of the respondents was that discussions did occur in some form with all admissions and were followed up by the nurses. Moreover, DNR orders were not entered without having been discussed with the patient or surrogate. Compliance appeared to be lower in the British hospital. A British physician estimated that there was less than 50% compliance with the requirement that all patients who might arrest should have a resuscitation status decided beforehand. Several of the respondents conveyed the impression that DNR orders were often entered without the expressed consent of the patient in cases that did not fit into the policy's conditions of severe brain damage, being irreversibly close to death, or CPR presenting an unacceptably high risk of death or brain damage.

Another compliance problem highlighted by two of the British nurses was that senior physicians sometimes opted not to resuscitate patients when no DNR order had been formally entered. A nurse mentioned a competent elderly patient for whom CPR had not been discussed, but whom she believed the doctors would not resuscitate because her current quality of life was thought to be extre-

mely poor. The nurse stated that "they've just made the decision she's not fit to be 222" (The term "222" is British medical slang for a full resuscitation attempt) (BN4, 1. 217). This determination is clearly not in compliance with the policy's mandate that all DNR orders based on quality of life con- siderations must be consented to by the patient. Such comments indicate that the extent of compliance with this hospital policy was not optimal.

DISCUSSION

Obligations to render futile care

One of the most important findings of the study was that the physicians interviewed were not well informed about their legal obligations to render care that is medically futile. Many American doctors believed that the were legally obligated to provide futile CPR if the patient or the patient's family requested it, when in fact both the law and their hospital's policy clearly state that they are not. The American physicians offered CPR to every patient out of a sense of duty, despite the fact that they were not actually required to do so and despite their claims that they believed CPR, like all other treatments, should only be offered to those patients who might possibly benefit from it. The British doctors did not seem to share the American physicians' concern about legal obligations.

In a related vein, the attitudes of British junior physicians concerning their moral duty to perform futile CPR differed markedly from those of the British senior physicians and the American physicians. Although they were not concerned about legal obligations, the British junior doctors believed that they had a positive moral obligation to give futile CPR if the patient wanted it. They did not believe they had the moral authority to override such a request, even though all but one of them claimed that they would override patients' requests for other medical treatments that could not be of any benefit to them. This suggests that they did not consider CPR to be like other medical treatments, in the sense of having indications and contraindica- tions. They viewed the decision to be DNR as an ethical decision rather than a technical one. In contrast, the British senior physicians felt quite comfortable refusing to render futile CPR. The American physicians felt even more strongly that they should not have to provide futile CPR, perhaps because some of them had had negative experiences in which they were legally bound to provide inappropriate CPR because a patient's family would not consent to a DNR order. They believed that the DNR decision was primarily a medical, not a moral, decision.


Policy versus practice

In both the British and the American hospital, there appeared to be considerable divergence between the official DNR policy of the hospital and the actual decision making practices of the physicians. Compliance with the hospital policy in the British hospital was low. Discussions with patients about CPR were infrequent and orders were sometimes entered for competent patients without any consultation with the patient. The British physicians did not know the CPR preferences of most of their patients. The mandate of the hospital policy that physicians should seek the views of the patient wherever possible was not being fulfilled to the maximum possible extent.

The American physicians were in compliance with their hospital policy's requirement that patient consent for DNR orders be obtained where CPR would not be medically futile. The study found no evidence that DNR orders were being made without the patient's consent at the American hospital, despite reports in the medical literature that as many as half of all DNR orders in the U.S. are made without consulting the patient or surrogate. The American physicians departed from the man- dates of their hospital's DNR policy, however, by seeking the patient's consent for DNR orders even where such consent was not required, such as in cases in which CPR would be futile. This practice resulted in physicians having to provide CPR in futile cases where consent for a DNR order could not be obtained.

The official DNR policies of the British and American hospitals were quite similar in their content, suggesting that DNR decisions would be made in much the same way in the both hospitals. However, in actuality decisions were made quite dif- ferently due to physicians' differing beliefs about their legal and moral duties in the provision of futile care and about the proper roles of physicians and patients in medical decision making. The British approach, in practice, relied heavily on the physician's clinical judgment, and the guiding principle was whether or not CPR was medically appropriate for a particular patient. The American approach focused on the autonomy of the patient, and the guiding principle was the patient's or surrogate's wishes. The British approach appeared to result in far fewer cases of CPR being provided when physicians felt it be a disservice to the patient, but in many of these cases patients were not given the opportunity to give their consent to the withholding of care. The American approach showed full respect for the patient's right to make autonomous choices, but at the price of physicians' ability to withhold CPR because in their clinical judgment, attempting to resuscitate this patient would be wrong.

LIMITATIONS OF THE STUDY

This study revealed some interesting differences in the way British and American physicians and nurses dealt with the decision not to resuscitate. However, the small sample precluded tests of statistical significance. Additionally, the higher proportion of older physicians and physicians who specialized in geriatrics in the British sample com- plicates the issue of whether the observed differences in DNR practices between British and American physicians can be attributed to cultural differences or other factors. It may be that younger physicians in both countries are more or less likely than older physicians to seek patient consent for procedures, value patient opinions, comply with formal guidelines, exercise their authority to withhold futile treatments, and so on. Similarly, one could theorize that geriatricians might be more likely than non-geriatricians to elicit and respect the treatment preferences of elderly patients. Study findings concerning national differences should be interpreted cautiously in light of these differences in sample composition. Future research could be di- rected at the substantiation of these results with a larger, more homogeneous sample.

The generalizability of the study findings is lim- ited by the study's focus on only two hospitals. There may be unknown organizational or other characteristics which make these hospitals unrepre- sentative of other NHS and VA hospitals. In particular, the VA hospital's academic affiliation may have some impact on physician practices in the hospital which was unmeasurable in the present study. This study is an exploratory study of how DNR decisions are made in an American and a British hospital and its conclusions are intended to inspire further qualitative and quantitative research into this important issue.

The present study addresses the prevalence of DNR discussions among physicians, patients, and families, but provides little information about the quality of these discussions. It was not possible to fairly judge the quality of these communications based on data from physician and nurse interviews; observational studies of physician patient-family interactions are needed. One such study by Tulsky et al. (1995) found that DNR discussions between medical residents and patients are frequently domi- nated by the physician in terms of speaking time, usually do not provide the patient with quantitative information about the risks and survival probabil- ities associated with CPR, and tend to focus on the patient's treatment preferences under various medical scenarios rather than on the patient's general concerns and values. The authors concluded that due to these shortcomings, most DNR discussions were of poor quality and did not provide the patient with sufficient information to make an informed decision about resuscitation. Some phys-


icians, including many of the physicians interviewed in the present study, might dispute this conclusion and argue that quantitative data are not useful in D N R discussions because survival rates and rates of adverse events from large groups of patients do not necessarily apply to any particular patient. Moreover, it is reasonable that the physician would consume the majority of the speaking time in a D N R discussion since one of the purposes of these discussions is to educate the patient about his options, their risks and their potential benefits. It is also understandable that physicians would orient discussions away from the general concerns of patients and toward specific medical scenarios because this results in clearer, more specific directives. However, if one draws the conclusion, based on the Tulsky study or other evidence, that the quality of D N R discussions is generally poor, then the finding of the present study that the American physicians showed great respect for patient autonomy would be somewhat mitigated. An essential element of respect for a patient's autonomous choices is ensuring that the patient has enough information to make an informed decision and that the patient makes decisions under appropriate conditions (e.g. without undue pressure from others, including the physician). Physicians must safeguard these conditions through their communications with patients if they are to fully respect their patients' autonomy.

Acknowledgements--The assistance of Sally C. Stearns, Ph.D. and Elizabeth Frazer, D. Phil. in the preparation of this manuscript is gratefully acknowledged.

REFERENCES

Aarons, E. J. and Beeching, N. J. (1991) Survey of "Do- Not-Resuscitate" orders in a district general hospital. British Medical Journal 303, 1504-1506.

Baker, R. (1993) The ethics of medical futility. Critical Care Clinics 9, 575-584.

Baker, R., Dersch V. and Fein, I. A. (1995) A comparative study of the effects of the New York DNR law and the JCAHO DNR policies. In Legislating Medical Ethics. A Study of the New York DNR Law, eds R. Baker and M. A. Strosberg. Kluwer Academic Press, Dordrecht.

Bedell, S. E. and Delbanco, T. L. (1984) Choices about cardiopulmonary resuscitation in the hospital: when do physicians talk with patients? New England Journal of Medicine 310, 1089-1093.

British Medical Association (1993) Decisions relating to cardiopulmonary resuscitation: a statement from the BMA and RCN in association with the Resuscitation Council (U.K.). Unpublished BMA memorandum.

Burnard, P. (1991) A method of analysing interview transcripts in qualitative research. Nurse Education Today 11, 461-466.

Crimmins, T. (1993) Ethical issues in adult resuscitation. Annals of Emergency Medicine 22, 495-501.

Dickinson, E. J. and Sabo, R. (1991a) Cardiopulmonary status decisions for elderly patients in a British teaching hospital. Journal of the American Geriatrics Society 39, A35.

Dickinson, E. J. and Sabo, R. (1991b) Operation of an unwritten cardiopulmonary resuscitation policy. Clinical Science 81, 25.

Doyal, L. and Wilsher, D. (1993) Withholding cardiopulmonary resuscitation: proposals for formal guidelines. British Medical Journal 306, 1593-1596.

Emanuel, E. J. and Emanuel, L. L. (1994) The economics of dying: the illusion of cost savings at the end of life. New England Journal of Medicine 330, 540-544.

Glaser, B. and Strauss, A. L. (1967) The Discovery of Grounded Theory. Aldine, Chicago.

Gulati, R. S., Bhan, G. L. and Horan, M. A. (1983) Cardiopulmonary resuscitation of old people. Lancet 2, 267-269.

Keating, R. M. (1989) Exclusion from resuscitation. Journal of the Royal Society of Medicine 82, 402~,05.

Laing, W. (1994) Managing the NHS: Past, Present and Agenda .for the Future. Office of Health Economics, London.

Miller, D. L. et al. (1993) Factors influencing physicians in recommending in-hospital cardiopulmonary resuscitation. Archives oflnternal Medicine 153, 1999 2003.

Saunders, J. (1992) Who's for CPR? Journal of the Royal College of Physicians London 26, 254-257.

Schecter, W. P. (1994) Withdrawing and withholding life support in geriatric surgical patients: ethical consider- ations. Surgical Clinics of North America 74, 245-259.

Schneider, P. A., Nelson, D. J. and Brown, D. J. (1993) In-hospital cardiopulmonary resuscitation: a 30-year review. Journal of the American Board of Family Practice 6, 91-101.

Stolman, C. J. et al. (1990) Evaluation of patient, physician, nurse and family attitudes towards do-not-resuscitate orders. Archives of Internal Medicine 150, 653-658.

Tomlinson, T. and Brody, H. (1988) Ethics and communi- cation in Do-Not-Resuscitate orders. New England Journal of Medicine 318, 43-46.

Tulsky, J. A., Chesney, M. A. and Lo, B. (1995) How do medical residents discuss resuscitation with patients? Journal of General Internal Medicine 10, 436-442.

Tunstall-Pedoe, H. et al. (1992) Survey of 3765 Cardiopulmonary Resuscitations in British Hospitals (the BRESUS Study): methods and overall results. British Medical Journal 304, 1347-1351.

United States Congress, Office of Technology Assessment (1987) Life-Sustaining Technologies and the Elderly. Government Printing Office, Washington, DC.

United States Senate, Special Committee on Aging (1989) Aging America." Trends and Projections. Government Printing Office, Washington, DC.

Vance, R. P., Ciancio, J. and West, E. (1993) Advance directives and do not resuscitate orders at Bowman Gray-Baptist Hospital Medical Center. North Carolina Medical Journal 54, 373-377.

Von Gunten, C. F. (1991) CPR in hospitalized patients: when is it futile? American Farnily Physician 44, 2130- 2134.

Vrtis, M. C. (1992) Cost/benefit analysis of cardiopulmonary resuscitation: a comprehensive study, Part II. Nursing Management 23, 44-51.

Williams, R. (1993) The 'Do Not Resuscitate' decision: guidelines for policy in the adult. Journal of the Royal College of Physicians of London 2, 139-140.

Documents

Comparison of medical and nursing attitudes to resuscitation and patient autonomy between a British and an American teaching hospital