login | login to stanford medicine web | stanford medicine

7
Roadblocks to Do-Not-Resuscitate Orders A Study in Policy Implementation Barbara E. Cammer Paris, MD; Victor G. Carrion, MD; James S. Meditch, Jr, MD; Carol F. Capello, MEd; Michael N. Mulvihill, DrPH Background: Cardiopulmonary resuscitation, a poten- tially lifesaving procedure, is initiated on hospitalized pa- tients who have an arrest in the absence of a written do\x=req-\ not-resuscitate (DNR) order. New York State Law specifies that attending physicians may write a DNR order on an adult patient either with his/her consent or that of a sur- rogate. Under specified circumstances, concurring physi- cian and witness signatures are also required. This study examines potential obstacles physicians may encounter when implementing a DNR order for a hospitalized patient. Methods: Sixty house staff officers and 45 attending phy- sicians at two New York City medical centers responded to a questionnaire listing 18 potential problems in ob- taining a DNR order. Using a Likert scale, respondents rated the prevalence of each problem. Results: Analysis of the data indicates that attending physician's failure to discuss DNR issues with patients and situations involving surrogate decision making are considered major obstacles to obtaining a DNR order. Procedural regulations, including abundant paperwork and witnessed signatures, are not identified as major obstacles. Conclusions: This study suggests a need for improved communication among physicians, patients, and surro- gates about advance directives, when feasible, either prior to hospitalization or early in its course, in an effort to comply with DNR legislation in a manner that reflects the patient's wishes and best interests. (Arch Intern Med. 1993;153:1689-1695) CARDIOPULMONARY resus¬ citation (CPR) is a spe¬ cific lifesaving medical intervention to restore heartbeat and support respiration in the event of cardiac or respiratory arrest. Although New York State's Do Not Resuscitate (DNR) Law, issued in 1988, intended to clarify the rights and responsibilities of patients, surrogates, and health professionals, it also resulted in presumed consent to CPR for patients whose medical records had no written DNR order.1 Such non- selective CPR may result in prolonging existence when there is no hope for re¬ covery, a situation that may be burden¬ some to the patient, the family, the health care providers, and the economic system. Hospitals have attempted to mitigate such problems by implementing formal DNR policies that state that a patient with DNR status be given all appropriate medical care except for CPR. As early as 1974, the Amer¬ ican Medical Association suggested that de¬ cisions not to resuscitate be written on the patient's chart and shared with all who care for the patient.2 However, it was not until 1988 that New York state legislation ef¬ fected regulation of DNR orders.3 The DNR order may be established either with the consent of a patient who has decision-making capacity or with a sur¬ rogate when, in the opinion of two phy¬ sicians, the patient lacks such capacity. Con¬ sent should be sought if the patient has a terminal condition, is permanently uncon¬ scious, or if resuscitation would be med¬ ically futile and would impose an extraor¬ dinary burden on the patient in light of See Materials and Methods on next page From the Department of Geriatrics and Adult Development, The Mount Sinai Medical Center, New York, NY. DownloadedFrom:http://archinte.jamanetwork.com/byaSTANFORDUnivMedCenterUseron03/12/2014

Upload: doantram

Post on 28-Jan-2017

229 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Login | Login to Stanford Medicine web | Stanford Medicine

Roadblocks to Do-Not-Resuscitate OrdersA Study in Policy ImplementationBarbara E. Cammer Paris, MD; Victor G. Carrion, MD; James S. Meditch, Jr, MD;Carol F. Capello, MEd; Michael N. Mulvihill, DrPH

Background: Cardiopulmonary resuscitation, a poten-tially lifesaving procedure, is initiated on hospitalized pa-tients who have an arrest in the absence of a written do\x=req-\not-resuscitate (DNR) order. New York State Law specifiesthat attending physicians may write a DNR order on anadult patient either with his/her consent or that of a sur-rogate. Under specified circumstances, concurring physi-cian and witness signatures are also required. This studyexamines potential obstacles physiciansmayencounterwhenimplementing a DNR order for a hospitalized patient.Methods: Sixty house staff officers and 45 attending phy-sicians at two New York City medical centers respondedto a questionnaire listing 18 potential problems in ob-taining a DNR order. Using a Likert scale, respondentsrated the prevalence of each problem.

Results: Analysis of the data indicates that attendingphysician's failure to discuss DNR issues with patientsand situations involving surrogate decision making areconsidered major obstacles to obtaining a DNR order.Procedural regulations, including abundant paperworkand witnessed signatures, are not identified as majorobstacles.

Conclusions: This study suggests a need for improvedcommunication among physicians, patients, and surro-gates about advance directives, when feasible, either priorto hospitalization or early in its course, in an effort tocomply with DNR legislation in a manner that reflects thepatient's wishes and best interests.

(Arch Intern Med. 1993;153:1689-1695)

CARDIOPULMONARY resus¬

citation (CPR) is a spe¬cific lifesaving medicalintervention to restoreheartbeat and support

respiration in the event of cardiac orrespiratory arrest. Although New YorkState's Do Not Resuscitate (DNR) Law,issued in 1988, intended to clarify therights and responsibilities of patients,surrogates, and health professionals, italso resulted in presumed consent toCPR for patients whose medical recordshad no written DNR order.1 Such non-selective CPR may result in prolongingexistence when there is no hope for re¬covery, a situation that may be burden¬some to the patient, the family, thehealth care providers, and the economicsystem.

Hospitals have attempted to mitigatesuch problems by implementing formal DNRpolicies that state that a patient with DNRstatus be given all appropriate medical care

except for CPR. As early as 1974, the Amer¬ican Medical Association suggested that de¬cisions not to resuscitate be written on thepatient's chart and shared with all who carefor the patient.2 However, it was not until1988 that New York state legislation ef¬fected regulation of DNR orders.3

The DNR order may be establishedeither with the consent of a patient whohas decision-making capacity or with a sur¬rogate when, in the opinion of two phy¬sicians, the patient lacks such capacity. Con¬sent should be sought if the patient has aterminal condition, is permanently uncon¬scious, or if resuscitation would be med¬ically futile and would impose an extraor¬dinary burden on the patient in light of

See Materials and Methodson next page

From the Department ofGeriatrics and AdultDevelopment, The Mount SinaiMedical Center, New York, NY.

Downloaded From: http://archinte.jamanetwork.com/ by a STANFORD Univ Med Center User on 03/12/2014

Page 2: Login | Login to Stanford Medicine web | Stanford Medicine

MATERIALS AND METHODS

Two sites were chosen for the study: The Mount Sinai Med¬ical Center and Beth Israel Medical Center, both large Man¬hattan-based tertiary referral centers with diverse socioèco-nomic and multiethnic patient populations. The forms forhandling DNR orders are similar at each site and complywith New York State Law.

A questionnaire (Table I ) was distributed at both hos¬pitals during medical conferences and put in the mailboxesof the Department of Medicine house staff and attendingphysicians. It lists 18 potential problems encountered in im¬plementing a DNR order for patients with and without de-cisional capacity. Physicians were asked to estimate the prev¬alence of each problem, using a Likert scale of 1 to 5 (1,rare; 5, common); a "not applicable" category was also pro¬vided. Since "competency" and "incompetency" are deter¬mined by the court, these terms, as used in the question¬naire, refer to patients with and without decisional capacity.

The first five questions in the questionnaire relate toissues for patients with decisional capacity: the patient re¬fuses DNR status; the patient verbally agrees but refuses tosign the DNR form; the attending physician does not dis¬cuss DNR status with the patient; and the house staff is toobusy or too uncomfortable to discuss DNR with the patient.

Questions 6 through 9 address problems implement¬ing a DNR order for all patients: disagreement or no com-

munication between attending physician and house staff re¬garding DNR status; attending physician fails to write theorder; and the house staff is too busy to complete the DNRorder.

Questions 10 through 14 address issues for patientswithout decisional capacity and who have no previous DNRstatus: surrogate fails to agree to DNR or verbally agrees butrefuses to sign the forms; surrogate is not contacted beforethe patient arrests; uncertainty as to who is the surrogate;and potential surrogates disagree about the patient's DNRstatus.

Questions 15 and 16 address issues regarding patientswho, during a previous hospital admission, consented to aDNR order but who have since become decisionally inca¬pable: the surrogate objects to the DNR order; and the at¬tending physician and house staff disagree regarding the pa¬tient's capacity to make decisions. Questions 17 and 18 addressproblems relating to availability of a witness and agreementto sign a DNR form for a patient who lacks decisional ca¬pacity.

For each question, the mean score of the Likert-scaledresponses was calculated with a 95% confidence interval.The "not applicable" category was treated as missing data.The questions were then ranked according to the most com¬monly to least commonly occurring barriers.

Additional demographic data were collected, includ¬ing participant age, gender, extent of postgraduate training,and prior instruction in DNR procedures.

I_

the medical condition and the expected outcome of re¬suscitation. By law, before obtaining consent the attend¬ing physician must provide the person giving consent withthe patient's diagnosis and prognosis, the reasonably fore¬seeable risks and benefits of CPR for the patient, and theconsequences of a DNR order. The attending physicianmust complete the necessary hospital forms in a timelymanner, including obtaining the patient's or surrogate'ssignature and that of a witness and concurring physician,and the patient's DNR status must be documented in themedical record.

See also page 1641

Deficiencies in implementation of a DNR order havebeen attributed to many factors, including delay in discus¬sions until the patient becomes too ill to participate4 and/orthe surrogate's lack of knowledge of the patient's wishes.5One study found that physicians who theoretically feel pa¬tients should partake in decisions about resuscitation ac¬tually rarely discuss this issue with their patients, due todiscomfort in approaching patients and/or familymembers.6This study attempts to identify the relevance of these andother issues as potential obstacles to the implementation

of a DNR order for a group of attending and house staffphysicians at two New York City hospitals.

RESULTS

Of the 229 questionnaires distributed, 114 were re¬

turned; nine completed by physicians outside the De¬partment of Medicine were excluded from analysis. In to¬tal, 105 questionnaires were evaluated: 45 by attendingphysicians (25 from Beth Israel, 20 from Mount Sinai)and 60 by house staff (30 from each site).

Table 2 provides information on physician age, gen¬der, level of training, and previous instruction in DNRimplementation. The median age of attending physicianswas 45 years, 89% were men, and they had practiced anaverage of 14 years. The median age of the house staffwas 28 years, and 67% were men. Less than one third ofthe attending physicians but almost half of the house staffhad prior instruction in DNR issues.

Table 3 summarizes the rank order results for allquestions. Table 4 subdivides responses by training leveland Table 5 by survey site. Questions focusing on com¬munication between attending physicians, patients, andsurrogates were identified as the most common obstaclesto DNR implementation by attending physicians and housestaff at both sites. Specifically, the highest-ranked prob-

Downloaded From: http://archinte.jamanetwork.com/ by a STANFORD Univ Med Center User on 03/12/2014

Page 3: Login | Login to Stanford Medicine web | Stanford Medicine

Table 1. Questionnaire*

For questions 1 through 5 , assume the patient is competent (¡e,has capacity to decide for himself/herself)1. Patient does not want DNR status

Common Rare N/A5 4 3 2 10

2. Patient gives verbal consent but does not want to sign theDNR formCommon Rare N/A5 4 3 2 10

3. Private attending physician does not discuss DNR status withpatientCommon Rare N/A5 4 3 2 10

4. House staff is too busy to discuss DNR status with patientbefore patient arrestsCommon Rare N/A5 4 3 2 10

5. House staff does not discuss DNR with the patient becausehouse staff does not feel comfortable with the issueCommon Rare N/A5 4 3 2 10

For questions 6 through 9, assume the patient may be eithercompetent or incompetent6. Difference of opinion between attending physician and housestaff regarding DNR statusCommon Rare N/A5 4 3 2 10

7. Failure of communication between attending physician andhouse staff regarding DNR statusCommon Rare N/A5 4 3 2 10

8. Attending physician fails to write order, even though thepatient has given verbal consentCommon Rare N/A5 4 3 2 10

9. House staff is too busy to complete DNR form before patientarrestsCommon Rare N/A5 4 3 2 10

For questions 10 through 14, assume the patient is incompetent,has no previous DNR status, and his/her wishes are unknown.The "surrogate" is the person selected to make a decisionregarding resuscitation on behalf of the patient10. Failure of surrogate to agree to DNR status

Common Rare N/A5 4 3 2 10

11. Surrogate agrees verbally but refuses to sign DNR formsCommon Rare N/A5 4 3 2 10

12. Failure to contact surrogate before patient arrestsCommon Rare N/A5 4 3 2 10

13. Uncertainty about who is the proper surrogateCommon Rare N/A5 4 3 2 10

14. Difference in opinion between potential surrogatesconcerning the patient's DNR statusCommon Rare N/A5 4 3 2 10

Table 1. Questionnaire (cont)D. For questions 15 and 16, please consider the patient whopreviously consented to a DNR order but who returns to thehospital with a new problem and is incapable of making adecision15. The surrogate objects to the DNR order

Common Rare N/A5 4 3 2 10

16. Difference in opinion between attending physician and housestaff regarding patient's capacityCommon Rare N/A5 4 3 2 10

E. For questions 17 and 18, consider the situation of an incompetentpatient whose surrogate agrees to DNR status and a witness isneeded17. A witness is unavailable

Common Rare N/A5 4 3 2 10

18. Witness does not agree to sign the DNR formCommon Rare N/A5 4 3 2 10

Demographic Data1. Were you ever instructed as to how to discuss DNR status with apatient or surrogate? Yes_No_

2. If yes, at what level of training were you instructed aboutdiscussing DNR status with patients or surrogates?A. Medical schoolB. House staffC. Attending physicianD.Other_

3. Please complete the following:Age;-Gender: male femaleMarital status: unmarried married divorced widowedPosition: PGY-1 PGY-2 PGY-3 PGY-4Fellow AttendingIf attending, years in practice_

*When obtaining do-not-resusitate (DNR) status on patients for whom itis medically advisable, a physician may encounter different problems;these may not be resolved before the patient undergoes cardiopulmonaryarrest. To help us determine which problems are most common inestablishing a DNR order, please evaluate the following potential problems,using a scale of 5 (Common) to 1 (Rare). A "not applicable" choice (N/A)is also included.

lems were (1) attending physician does not discuss DNRwith the patient (highest-ranked problem by house staff);(2) surrogate does not agree to DNR; (3) surrogate is notcontacted before the patient arrests (highest-ranked prob¬lem by attendings); (4) uncertainty about who is the sur¬rogate; and (5) potential surrogates disagree as to the pa¬tient's DNR status.

COMMENT

Only about half of the distributed questionnaires werereturned, most likely a result of both the method of on-site collection at conferences and lack of follow-up com¬munication with those who had questionnaires placed in

Downloaded From: http://archinte.jamanetwork.com/ by a STANFORD Univ Med Center User on 03/12/2014

Page 4: Login | Login to Stanford Medicine web | Stanford Medicine

their mailboxes. It is unclear from this study whether otherfactors such as discomfort with the issues and lack of in¬terest contributed to the low return rate or whether a pre¬selection bias may have occurred simply by virtue of thesamplingmethod. Yet, assuming that the respondents likelyrepresent physicians who are both more favorably dis¬posed to DNR issues and more inclined to be abreast ofnew issues (eg, attending hospital conferences), this studyis valuable in that it points out serious obstacles even forthose physicians seemingly willing to address DNR issueswith their patients.

The results of this study represent universal dilem¬mas in communication among physicians, patients, andsurrogates, as previous studies have acknowledged.7 Forexample, although it is the attending physician who mustsign the DNR order, in clinical practice the house staff isoften expected to initiate a discussion of these issues withpatients and/or surrogates. Another study has found in¬terns initiating 90% of the DNR discussions with eitherthe patient or the family, even though policies and guide¬lines give attending physicians distinct responsibilities con¬cerning the DNR order.8 In American teaching hospitals,this practice more than likely results from several factors,including attending physician's time constraints on rounds,difficulty in pursuing surrogates not present at the bed¬side, as well as discomfort in confronting the issue of death,particularly with patients for whom they have been car¬ing over a period of years. In contrast, house staff are lesslikely to have a prior relationship with the patient andthereby may be better prepared psychologically to obtainsuch consent. However, since many hospitals do not have

Table 2. Characteristics of Physician Respondents

Total responsesAttending physicians, No. (%)AgeMedianRange

Gender, No. (%)fMateFemale

Years in practicefMedian yearsRange years

Prior DNR instruction, No. (%)House staff officers, No. (%)AgeMedianRange

Gender, No. (%)tMaleFemale

FellowsPGY-3PGY-2PGY-1Prior DNR instruction, No. (%)

Beth Israel Mount Sinai Totals

55 50 10525 (45) 20 (40) 45 (43)

45 45 4529-76 32-66 29-76

24 (96) 16 (80) 40 (89)1 (4) 3 (15) 4 (9)

15 14 141-47 1-35 1-479 (36) 4 (20) 13 (29)30 (55) 30 (60) 60 (57)

26 28 2823-36 25-33 23-361 NR

...

1 NR

20 (67) 20 (67) 40 (67)9 (30) 9 (30) 18 (30)

1 NR (3) 1 NR (3) 2 NR (3)2 2 44 14 189 4 1315 10 25

15(50) 11(37) 26(43)*NR indicates no response, DNR, do not resuscitate.\Not all respondents answered these questions.

Table 3. Mean Rank Score of Responses and 95% Confidence Interval*

ConfidenceRank Question Mean Interval Condensed Statement1 3 3.78 (3.52-4.02) Attending physician does not discuss DNR with patient2 12 3.63 (3.41-3.85) Surrogate not contacted before patient arrests3 10 3.40 (3.16-3.64) Surrogate does not agree to DNR4 14 3.31 (3.07-3.55) Disagreement between potential surrogates about patient's DNR status5 13 3.22 (2.97-3.47) Uncertainty about who is surrogate6 1 3.01 (2.77-3.25) Patient refuses DNR status7 7 2.91 (2.66-3.16) Communication failure: attending physician and house staff8 8 2.67 (2.40-2.94) Attending physician fails to write DNR order9 6 2.66 (2.41-2.91) Disagreement between attending physician and house staff about patient's DNR status10 5 2.48 (2.23-2.73) House staff uncomfortable discussing DNR with patient11 15 2.40 (2.15-2.65) Surrogate objects to DNR; patient previously agreed12 11 2.36 (2.11-2.61) Surrogate agrees verbally; does not sign DNR13 4 2.21 (1.99-2.43) House staff too busy to discuss DNR with patient14 16 2.15 (1.91-2.39) Disagreement between attending physician and house staff about patient's capacity15 2 1.95 (1.71-2.19) Patient agrees verbally; does not sign DNR16 9 1.82 (1.62-2.02) House staff too busy to complete DNR forms17 17 1.39 (1.23-1.55) Witness is unavailable18 18 1.32 (1.18-1.46) Witness refuses to sign DNR

"DNR indicates do not resuscitate.

Downloaded From: http://archinte.jamanetwork.com/ by a STANFORD Univ Med Center User on 03/12/2014

Page 5: Login | Login to Stanford Medicine web | Stanford Medicine

Table 4. Rank Order of Responses by Physician Status*

Attending Physician House Staff

Rank QuestionOrder No. Statement

RankOrder

QuestionNo. Statement

1 12 Surrogate not contacted before patient arrests 12 14 Disagreement between potential surrogates about patient's 2

DNR status3 13 Uncertainty about who is surrogate 34 3 Attending physician does not discuss DNR with patient 4

5 10 Surrogate does not agree to DNR 56 5 House staff uncomfortable discussing DNR with patient 67 1 Patient refuses DNR status 78 7 Communication failure: attending physician and house staff 8

9 4 House staff too busy to discuss DNR with patient 910 11 Surrogate agrees verbally, does not sign DNR 1011 15 Surrogate objects to DNR; patient previously agreed 1112 8 Attending physician fails to write DNR order 12

13 6 Disagreement between attending physician and house staff 13about patient's DNR status

14 9 House staff too busy to complete DNR forms 1415 16 Disagreement between attending physician and house staff 15

about patient's capacity16 2 Patient agrees verbally; does not sign DNR 1617 18 Witness refuses to sign DNR 1718 17 Witness is unavailable 18

3 Attending physician does not discuss DNR with patient10 Surrogate does not agree to DNR

12 Surrogate not contacted before patient arrests14 Disagreement between potential surrogates about patient's

DNR status13 Uncertainty about who is surrogate7 Communication failure: attending physician and house staff1 Patient refuses DNR status6 Disagreement between attending physician and house staff

about patient's DNR status8 Attending physician fails to write DNR order15 Surrogate objects to DNR; patient previously agreed11 Surrogate agrees verbally, does not sign DNR16 Disagreement between attending physician and house staff

about patient's capacity5 House staff uncomfortable discussing DNR with patient

4 House staff too busy to discuss DNR with patient2 Patient agrees verbally; does not sign DNR

9 House staff too busy to complete DNR forms17 Witness is unavailable18 Witness refuses to sign DNR

"DNR indicates do not resuscitate.

house staff, further studies need to be conducted to de¬termine the extent to which time and reimbursement af¬fect attending physicians' initiation of DNR orders. It isinteresting to note, as indicated in Table 4, that althoughthe roles of the house staff and attending physicians differin the implementation of a DNR order, both groups ofphysicians identified the same five problems as key ob¬stacles, four of which focus on surrogate issues.

This study points out that lack of previous designa¬tion of a surrogate by the patient may delay implementa¬tion of a DNR order. According to New York State Law, ifthere is no previously designated surrogate or the previ¬ously designated surrogate is not readily available or doesnot wish to participate in the decision regarding CPR, thepriority for choosinganother surrogate, respectively, is spouse,adult child, parent, adult sibling, and, finally, close friend—even if a close friend may be more familiar with the pa¬tient's wishes than a family member. Although the law isseemingly clear, uncertainty about who is the proper sur¬rogate is among the five most commonly identified obsta¬cles cited by both house staff and attending physicians atboth sites. Furthermore, according to NewYork State Law,if the patient's wishes for care at the end of life are not known,the surrogate should make a decision based on the patient'sprior expressed wishes and religious and moral beliefs. If

the patient's wishes are unknown, the surrogate should basethe decision on the patient's best interests.

One study examining the utility of the principal of sub¬stituted judgment in surrogate decisionmaking in predict¬ing a patient's resuscitation preferences found that althoughspouses believed their predictions were accurate, theywereno better than chance alone and they overestimated pa¬tients' preferences for resuscitation.9 In our study, surro-

lack of previous designation of asurrogate by the patient may delayimplementation of a DNR order

gates were identified more often as obstacles to obtaininga DNR order than were patients themselves. One possibleexplanation for this is that without a clear understandingof the patient's viewpoint, it is likely to be psychologicallyless burdensome for the surrogate to err on the side of life-sustainingmeasures than to sign aDNR form. Another study,examining the accuracy of family members in predictingresuscitation preferences in current state of health and di¬minished mental capacity scenarios comparedwith patient'sstated preferences, also confirms this phenomenon. Thestudy showed that where discrepant responses occurred,

Downloaded From: http://archinte.jamanetwork.com/ by a STANFORD Univ Med Center User on 03/12/2014

Page 6: Login | Login to Stanford Medicine web | Stanford Medicine

family members tended to err on the side of providing re¬suscitation for the patient.5

Clearly, improved communication, prior to hospi-talization, among physicians, patients, and surrogates inan attempt to understand patients' wishes for care at theend of life could mitigate many major potential obstaclesto DNR implementation identified in this study. Just whena patient should be included in a DNR discussion hasbeen an ongoing subject of debate.

Youngner,10 for example, supports early discussion,noting that as patients become more aware of resuscita¬tion issues, they are increasingly drawing up living willsand durable power-of-attomey agreements. Brennan11 sug¬gests that many physicians are resistant to discussing DNRstatus with a patient because of an inability to predictwhen a patient will die or to determine when he/she ishopelessly ill. Some argue that patients who slowly andprogressively deteriorate would have undue anxiety if thesubject of resuscitation were raised.12

Physicians may also fear legal repercussions ofhaving a patient sign a DNR order and thus not offerthe patient all possible available medical technologies.In fact, one recent study on the effect of the New YorkState Law on in-hospital CPR practice showed no sig¬nificant decline in the resuscitation attempts with theadvent of the law.13

Reasons why physicians tend not to discuss these

issues with patients or encourage patients to choose a sur¬rogate are likely to be multifactorial. Most physicians inthis study have had no prior instruction in developingcommunication skills with patients and surrogates aboutDNR issues, despite the enactment of the New York StateLaw several years ago requiring that such discussions takeplace.14 In fact, fewer than half of the physicians in thisstudy reported any formal or informal instruction in theimplementation of this law. Since physicians traditionallyview themselves as healers of the sick, not planners forthe future death of their patients, it is understandable thatthey may feel some discomfort in discussing death withtheir patients. A training program stressing the value ofthese discussions with patients in advance of hospitaliza-tion and suggesting practical approaches to this topic mightalleviate some of the potential obstacles outlined in thisstudy.

Problems such as patient or surrogate verballyagreeing but not signing the forms, house staff toobusy to complete the DNR forms, or witness refusingto sign a DNR form did not receive high scores, sug¬gesting that an overabundance of paperwork, althoughburdensome, is not a strong obstacle to implementinga DNR order.

The questionnaire used in this study only superfi¬cially touches on problems one may encounter in imple¬menting a DNR order, and the responses only suggest

Table 5. Rank Order of Responses by Institution*

Rank QuestionOrder No. Mount Sinai

Rank QuestionOrder No. Beth Israel

123.53.5

10

1112

1314

15161718

3 Attending physician does not discuss DNR with patient12 Surrogate not contacted before patient arrests13 Uncertainty about who is surrogate10 Surrogate does not agree to DNR

14

716

5

Disagreement between potential surrogates about patient'sDNR status

Communication failure: attending physician and house staffPatient refuses DNR statusDisagreement between attending physician and house staffabout patient's DNR status

Attending physician fails to write DNR orderHouse staff uncomfortable discussing DNR with patient

11 Surrogate agrees verbally; does not sign DNR16 Disagreement between attending physician and house staff

about patient's capacity15 Surrogate objects to DNR; patient previously agreed2 Patient agrees verbally; does not sign DNR

4 House staff too busy to discuss DNR with patient9 House staff too busy to complete DNR forms17 Witness is unavailable18 Witness refuses to sign DNR

910

1112

1314

15161718

12 Surrogate not contacted before patient arrests3 Attending physician does not discuss DNR with patient10 Surrogate does not agree to DNR14 Disagreement between potential surrogates about patient's

DNR status13 Uncertainty about who is surrogate

1 Patient refuses DNR status7 Communication failure: attending physician and house staff15 Surrogate objects to DNR; patient previously agreed

8 Attending physician fails to write DNR order6 Disagreement between attending physician and house staff

about patient's DNR status5 House staff uncomfortable discussing DNR with patient4 House staff too busy to discuss DNR with patient

11 Surrogate agrees verbally; does not sign DNR16 Disagreement between attending physician and house staff

about patient's capacity9 House staff too busy to complete DNR forms2 Patient agrees verbally; does not sign DNR18 Witness refuses to sign DNR17 Witness is unavailable

"DNR indicates do not resuscitate.

Downloaded From: http://archinte.jamanetwork.com/ by a STANFORD Univ Med Center User on 03/12/2014

Page 7: Login | Login to Stanford Medicine web | Stanford Medicine

where problems may exist. Additional data regarding phy¬sicians' clinical exposure to DNR issues would be helpfulin interpreting the results. Further studies might includequestions focusing on why attending physicians do notdiscuss DNR issues with patients, examining the level ofcomfort in initiating such discussions with a surrogate orpatient and possible distinction between situations wherecardiac arrest is expected and unexpected. Similarly, fur¬ther assessment of surrogates' comfort in signing the pa¬pers and knowledge of the patient's wishes for resusci¬tation would be useful.

Guidelines and policies regarding DNR orders needto be reassessed in an effort to facilitate obtaining a DNRorder before the patient arrests, and in so doing avoid apotentially unwanted, invasive, and costly procedure. Im¬proved communication between physicians, patients, andsurrogates in advance of acute hospitalization could po¬tentially facilitate the implementation of an appropriateDNR order, respecting the patient's wishes for care at theend of life.

Accepted for publication February 16, 1993.Reprint requests to the Department of Geriatrics and

Adult Development, Box 1070, Mount Sinai Medical Center,One Gustave L. Levy Place, New York, NY 10029 (Dr Paris).

M REFERENCES M1. Swidler RN. The presumption of consent in New York State's do-not\x=req-\resuscitate law. N Y State J Med. 1989;89:69-72.

2. Rabkin MT, Gillerman JD, Rice NR. Orders not to resuscitate. N Engl J Med.1976;295:364-366.

3. Article 29-B, Statute 413-A. The State of New York Public Health Law, effectiveApril 1, 1988.

4. Emanuel LL. Does the DNR order need life-sustaining intervention? time forcomprehensive advance directives. Am J Med. 1989;86:97-90.

5. Seckler AB, Meier DE, Mulvihill MN, Paris BE. Substituted judgement: howaccurate are proxy predictions? Ann Intern Med. 1991;115:92-98.

6. Bedell SE, Delbanco TL. Choices about cardiopulmonary resuscitation in thehospital: when do physicians talk with patients? N Engl J Med. 1984;310:1089-1093.

7. Bernard LO, Steinbrook RL. Deciding whether to resuscitate. Arch Intern Med.1983;143:1561-1563.

8. Schwartz DA, Reilly P. The choice not to be resuscitated. J Am Geriatr Soc.1986;34:807-811.

9. Uhlmann RF, Pearlman RA, Cain KC. Physicians' and spouses' predictions ofelderly patients' resuscitation preferences. Gerontology. 1988;43:M115-121.

10. Youngner SJ. Do-not-resuscitate orders: no longer a secret, but still a prob-lem. Hastings Cent Rep. 1987;17:24-33.

11. Brennan TA. Do-not-resuscitate orders for the incompetent patient in the ab-sence of family consent. Law Med Health Care. 1986;14:13-19.

12. Charlson ME, Sax FL, MacKenzie CR, Fields SD, Braham RL, Douglas D Jr.Resuscitation: how do we decide? J Am Geriatr Soc. 1986;255:1316-1322.

13. Kamer RS, Dieck EM, McClung JA, White PA, Sivak SL. Effect of New YorkState's do-not-resuscitate legislation on in-hospital cardiopulmonary resusci-tation practice. Am J Med. 1990;88:108-111.

14. McClung JH, Kamer RS. Legislating ethics: implications of New York's do\x=req-\not-resuscitate law. N Engl J Med. 1990;323:270-272.

Downloaded From: http://archinte.jamanetwork.com/ by a STANFORD Univ Med Center User on 03/12/2014