New and Lingering Controversies in Pediatric End-of-Life Care
Mildred Z. Solomon, EdD*‡; Deborah E. Sellers, PhD*; Karen S. Heller, PhD*; Deborah L. Dokken, MPA§;Marcia Levetown, MD�; Cynda Rushton, DNsc, RN, FAAN¶; Robert D. Truog, MD‡; and
Alan R. Fleischman, MD#
ABSTRACT. Objectives. Professional societies, ethicsinstitutes, and the courts have recommended principlesto guide the care of children with life-threatening condi-tions; however, little is known about the degree to whichpediatric care providers are aware of or in agreementwith these guidelines. The study’s objectives were todetermine the extent to which physicians and nurses incritical care, hematology/oncology, and other subspecial-ties are in agreement with one another and with widelypublished ethical recommendations regarding the with-holding and withdrawing of life support, the provisionof adequate analgesia, and the role of parents in end-of-life decision-making.
Methods. Three children’s hospitals and 4 generalhospitals with PICUs in eastern, southwestern, andsouthern parts of the United States were surveyed. Thispopulation-based sample was composed of attendingphysicians, house officers, and nurses who cared forchildren (age: 1 month to 18 years) with life-threateningconditions in PICUs or in medical, surgical, or hematol-ogy/oncology units, floors, or departments. Main out-come measures included concerns of conscience, knowl-edge and beliefs, awareness of published guidelines, andagreement or disagreement with guidelines.
Results. A total of 781 clinicians were sampled, in-cluding 209 attending physicians, 116 house officers, and456 nurses. The overall response rate was 64%. Fifty-fourpercent of house officers and substantial proportions ofattending physicians and nurses reported, “At times, Ihave acted against my conscience in providing treatmentto children in my care.” For example, 38% of critical careattending physicians and 25% of hematology/oncologyattending physicians expressed these concerns, whereas48% of critical care nurses and 38% of hematology/oncol-ogy nurses did so. Across specialties, �20 times as manynurses, 15 times as many house officers, and 10 times asmany attending physicians agreed with the statement,“Sometimes I feel we are saving children who should notbe saved,” as agreed with the statement, “Sometimes Ifeel we give up on children too soon.” However, hema-tology/oncology attending physicians (31%) were lesslikely than critical care (56%) and other subspecialty
(66%) attending physicians to report, “Sometimes I feelthe treatments I offer children are overly burdensome.”Many respondents held views that diverged widely frompublished recommendations. Despite a lack of awarenessof key guidelines, across subspecialties the vast majorityof attending physicians (range: 92–98%, depending onspecialty) and nurses (range: 83–85%) rated themselvesas somewhat to very knowledgeable regarding ethicalissues.
Conclusions. There is a need for more hospital-basedethics education and more interdisciplinary and cross-sub-specialty discussion of inherently complex and stressfulpediatric end-of-life cases. Education should focus onestablishing appropriate goals of care, as well as on painmanagement, medically supplied nutrition and hydra-tion, and the appropriate use of paralytic agents. Moreresearch is needed on clinicians’ regard for the dead-donor rule. Pediatrics 2005;116:872–883; pediatric pallia-tive care, end-of-life care, clinical decision-making, painmanagement, ethics.
Since Duff and Campbell1 first brought the issueto public attention, the federal government,2medical3 and nursing4 specialty societies, inde-
pendent ethics institutes,5 and the courts6–10 haveaddressed a number of vexing issues regarding howbest to make decisions about the use of life supportin neonatal and pediatric cases. There is generalagreement that life support may be forgone whenchildren are terminally ill or permanently uncon-scious, in instances of “medical futility,” and whenthe burdens of treatment outweigh the benefits fromthe child’s perspective.11 In the many ambiguouscases, where the burdens of illness and continuedtreatment seem great and the benefits uncertain, pa-rental preferences are granted a great deal of author-ity.2,11 Policy documents also require the provision ofadequate analgesia and offer guidelines related toorgan donation.12 Since the publication of these land-mark reports, new ethical issues in the care ofgravely ill children have arisen, and the Institute ofMedicine has called for more attention to the ethicsof pediatric palliative care.13
A major purpose of the research reported here wasto ascertain physician, house officer, and nurse be-liefs with respect to the range of issues presented inTable 1, which summarizes key national recommen-dations published to date. A second purpose was todetermine the extent to which health care profession-als are concerned with problems of overtreatment. Inprevious research, a member of the study team doc-umented significant concerns of conscience regard-
ing inappropriate use of technology among clinicianscaring for gravely ill adults.14 We hypothesized thatpediatric health care professionals might experiencesimilar concerns. To address these issues, we sur-veyed attending physicians, house officers, andnurses working in PICUs and medical and surgicalservices of children’s hospitals and general hospitalswith pediatric units in the United States.
METHODS
Instrument ConstructionThe survey instrument constructed for this study was based on
the Decisions Near the End of Life Institutional Profile, a tool usedin hundreds of hospitals across the United States to assess clini-cians’ views on a wide range of ethical and legal issues in adultend-of-life care.14 The pediatric version is similar to the adultversion, in that both probe for respondents’ knowledge, attitudes,and self-reported behaviors relevant to published national ethicaland legal guidelines. The pediatric version addresses issues spe-
cific to pediatric decision-making, such as caregivers’ perceptionsof the quality of communication with parents and the extent towhich adolescents should be involved in treatment planning. Thepediatric instrument was developed after a thorough review of theliterature on family-centered care and ethical guidelines for treat-ment of children near the end of life. Then, during a 2-day retreat,national experts in pediatric medicine and nursing, clinical ethics,family-centered care, palliative care, and health law reviewed theitems for content validity.
This article reports on 3 types of items from the survey, ie, (1)opinion items, which probe for respondents’ agreement or dis-agreement with existing national guidelines; (2) knowledge items,which probe for the extent to which respondents are or are notaware of existing guidelines; and (3) belief items, which probe forrespondents’ attitudes and concerns about the care they provide.Five-point Likert response options ranging from strongly disagreeto strongly agree were appropriate for all 3 types of items.
Pilot testing was conducted by administering the instrument to71 subjects in 3 institutions. The extent of variation and the per-centage of missing data were examined for each item. To assesstest-retest repeatability, 49 of these respondents completed theinstrument a second time within a 2- to 3-week period. Responseswere consistent 70% to 90% of the time among respondents withretesting of items.
SampleSeven teaching hospitals with broad geographic representation
participated, in Georgia, Maryland, Massachusetts, New Jersey,Pennsylvania, and Texas (where 2 sites were located). Four of the7 hospitals are children’s hospitals, and 3 are general hospitalswith PICUs. The study protocol was approved by institutionalreview boards at the Education Development Center and partici-pating hospitals. The survey was administered to pediatric phy-sicians and nurses who cared for children (age: 1 month through18 years) with life-threatening conditions in PICUs, medical units,surgical units, and/or hematology/oncology units. Providers whoworked solely in NICUs were excluded, because the survey itemswere relevant mainly to pediatric decision-making and were notdesigned specifically to target special issues arising in the care ofneonates.
AnalysisOur analytical interest was twofold, ie, (1) to determine the
overall pattern of responses to individual survey items and (2) todetermine whether the pattern of responses varied according toprofessional discipline, subspecialty, and/or site. Professional dis-cipline included 3 groups, ie, attending physician, house officer,and nurse. With regard to subspecialty, attending physicians andnurses were grouped into 3 categories, ie, critical care, hematolo-gy/oncology, and other. Among attending physicians, the othercategory included, for example, pulmonology, cardiology, reha-bilitation medicine, orthopedics, and anesthesia. These subspecial-ties were combined into a single category because most wererepresented by �5 respondents. Among nurses, the other categoryincluded those who simply recorded their specialty as pediatricsand those who answered the question by listing the unit or servicein which they worked, such as trauma, emergency transport, ordialysis. Because house officers have not yet completed their train-ing, subspecialty was not categorized for them.
To examine the pattern of differences for each survey itemaccording to discipline, subspecialty, and site, �2 tests of indepen-dence were used. Because a subspecialty was not designated forhouse officers, separate �2 tests of independence were performedto ascertain differences according to subspecialty for attendingphysicians and for nurses. In addition, a �2 test of independencewas performed for each subspecialty group to examine differencesbetween attending physicians and nurses. Finally, site differencesfor each outcome were examined with a �2 test of independencebetween the survey item and the site. To facilitate interpretationand presentation, the 5-point Likert scale was collapsed into 3categories, ie, agree, uncertain, and disagree.
RESULTS
Respondent CharacteristicsQuestionnaires were completed by 211 attending
physicians, 116 house officers, and 469 nurses.
TABLE 1. Summary of Published Guidelines for PediatricEnd-of-Life Decision-Making
1. Withholding versus withdrawing life support. Although it isrecognized that there is a psychologic difference betweenstopping a treatment already begun and not starting it in thefirst place, for nearly 2 decades all nationally prominentethics guidelines have urged clinicians to recognize that,ethically, the distinction between withholding andwithdrawing life support is moot. An assessment, from thechild’s perspective, of the burdens and benefits of offering orcontinuing treatment is what should count ethically, notwhether a treatment has begun or not.2,3,11,27
2. Medically supplied food and fluids. Medically supplied nutritionmay prolong the dying process and, near the end of life, maycause congestion, excessive bodily secretions, andconsiderable discomfort. Medically supplied nutrition andhydration are not ethically and legally required in allcircumstances. They can be stopped or foregone, like all othermedical interventions, when their burdens outweigh theirbenefits.2–4,11,18
3. Use of opioids. The appropriateness of providing adequatelevels of pain medication, even if doing so hastens death, iswell established both ethically2,27 and legally.26 The doctrineof “double effect” holds that it is ethically permissible toprovide adequate analgesia as long as the intention is torelieve suffering and not to cause death. However, recentresearch suggests that, if the dosage has been titratedproperly, then respiratory depression is rare even whenopioids are used at very high levels.25
4. Use of paralytic agents. Ethical principles of palliative care callon clinicians to provide medication at the end of life with theintent of relieving pain and suffering. Althoughneuromuscular blocking agents may make a patient appear tobe comfortable, by rendering the patient incapable ofshowing signs of pain or suffering, they have no analgesic orsedative properties; therefore, clinicians cannot maintain thatthese agents are administered with the intention ofpromoting comfort. Their use should never be initiatedduring the withdrawal of life-sustaining treatments.28,29
5. Brain death. In all 50 states, death can be defined on the basisof neurologic criteria (complete cessation of brain function) aswell as cardiac criteria, and permission of the family is notnecessary for withdrawal of ventilatory support when eithertype of criteria is met.12 Two states have special provisionsfor accommodating religious objections to brain death.Currently, however, no published guidelines address thequestion of how clinicians and institutions should respond toparental objections to the cessation of ventilatory support forchildren who have died according to neurologic criteria.
6. The dead-donor rule. The Uniform Anatomical Gift Act holdsthat organs may be retrieved only from patients who aredead. Permanently unconscious patients are not allowed, bycurrent ethical standards, to serve as organ donors.12
ARTICLES 873 by guest on June 12, 2013pediatrics.aappublications.orgDownloaded from
Across the 7 institutions, the response rates rangedfrom 54% to 71%, but only 1 institution had a re-sponse rate of �60%. The overall response rate was64%, which is similar to response rates for othersurveys of health care professionals.15,16
Table 2 provides details about respondents’ demo-graphic and professional characteristics and abouttheir perceptions of their knowledge of basic princi-ples and issues in medical ethics. The specialty of 2attending physicians and 13 nurses could not bedetermined, and these 15 subjects were excludedfrom analyses. Approximately 12% of attending phy-sicians were from critical care, 25% were from hema-tology/oncology, and 62% were from other special-ties. In contrast, 57% of nurses worked in criticalcare, 19% in hematology/oncology, and 21% in otherspecialties. The vast majority of respondents consid-ered themselves knowledgeable regarding medicalethics. Regardless of subspecialty, attending physi-cians were less likely than house officers or nurses toreport that they were “not very knowledgeable.”
Concerns of ConscienceTable 3 demonstrates concerns of conscience
among all respondents. House officers and nurseswere more likely than attending physicians to report,“At times, I have acted against my conscience inproviding treatment to children in my care.” How-ever, fewer nurses and physicians within the subspe-cialty of hematology/oncology (38% and 25%, re-spectively) reported these concerns of conscience,compared with critical care nurses (48%), other sub-specialty nurses (54%), and critical care physicians(38%).
A full 78% of house officers, 80% of critical carephysicians, 68% of other non-hematology/oncologysubspecialty attending physicians, and more thantwo thirds of critical care and other non-hematolo-gy/oncology nurses agreed with the statement,“Sometimes I feel we are saving children who shouldnot be saved.” As with concerns of conscience, he-
matology/oncology physicians and nurses were lesslikely to hold this view; 36% and 35% of the hema-tology/oncology attending physicians and nurses,respectively, agreed with the statement.
A similar pattern was noted with respect to thestatement, “Sometimes I feel the treatments I offerchildren are overly burdensome.” More than one halfof the critical care attending physicians (56%), twothirds of the other non-hematology/oncology sub-specialists (66%) and house officers (63%), and morethan one half (53%) of other subspecialty nursesagreed with the statement, whereas only one third ofthe hematology/oncology attending physicians(31%) and hematology/oncology nurses (33%) did.Critical care nurses were more similar to their hema-tology/oncology nurse counterparts for this itemthan they were for other items. Across all subspecial-ties and disciplines, very few agreed with the state-ment, “Sometimes I feel we give up on children toosoon.”
Clinicians’ Views Related to PublishedRecommendations
Withholding and Withdrawing Life SupportTable 4 presents the percentages of respondents
who agreed, were uncertain, or disagreed on 3 itemsrelated to withholding and withdrawing life support.Critical care attending physicians were much morelikely than other attending physicians, house offic-ers, or nurses to agree with the frequently promul-gated ethics recommendation, “There is no ethicaldifference between withholding (not starting) a lifesupport measure and stopping it once it has beenstarted.” Similarly, only 14% of critical care attendingphysicians agreed with the false statement, “There isa prevailing consensus among ethicists that with-drawing a treatment is ethically different from with-holding it.” In contrast, at least 40% of hematology/oncology attending physicians, attending physiciansfrom other specialties, house officers, and nurses
agreed with this false statement. In addition, onethird or more of all respondents, except critical careattending physicians, were uncertain about the na-ture of the “prevailing consensus.” With thesegroups added together, �70% of all respondents ex-cept critical care attending physicians were unawareof existing guidelines regarding the ethical equiva-lence of withholding and withdrawing treatments.
The third question asked respondents whetherthey agreed or disagreed with the statement, “Even iflife supports such as mechanical ventilation and di-alysis are stopped, medically supplied food and wa-ter should always be continued.” More than onethird (36%) of critical care attending physicians, onefourth of hematology/oncology attending physi-cians, and �40% of all other respondents agreedwith this false statement.
pain-related items. Two items showed high levels ofagreement with broadly accepted clinical and ethicalguidelines. At least 80% of all respondents agreedwith the true statement, “For a child whose death isinevitable, it is appropriate to provide opioid painmedication before withdrawing a ventilator.” Aneven higher percentage of respondents agreed withthe true statement, “Sometimes it is appropriate togive pain medication to relieve suffering even if itmay hasten a child’s death.”
Nevertheless, one fifth of non-critical care attend-ing physicians and more than one fourth of houseofficers and nurses reported, “At this institution,children are often given inadequate pain medica-tion.” Across subspecialties, approximately one halfof attending physicians and 41% of house officersagreed, “When clinicians give inadequate pain med-ications, they do so most often out of fear of hasten-ing the child’s death.” Among nurses, hematology/oncology nurses (68%) were substantially morelikely than critical care nurses (36%) or other spe-cialty nurses (47%) to agree with that statement.Nurses of all specialties were also substantially morelikely than attending physicians to believe the falsestatement, “It is illegal to give pain medication indoses that risk respiratory depression to the point ofdeath.” In addition to the nurses who held this inac-curate view of the law, substantial proportions ofnurses provided a response of “3,” indicating uncer-tainty about the legality of this practice.
Ethical Aspects of Managing the Dying ProcessTable 6 presents 3 items related to decisions that
clinicians may encounter while they manage the dy-ing process or immediately after a child’s death.Regarding the first question, more hematology/on-cology attending physicians and nurses (32% and48%, respectively) than critical care attending physi-cians (16%), other specialty attending physicians(14%), house officers (19%), critical care nurses (14%),and other specialty nurses (25%) were likely to agreewith the controversial statement, “For a child whosedeath is inevitable, it is appropriate to provide neu-romuscular blocking agents such as Pavulon (pancu-
ronium) with analgesia before withdrawing a venti-lator.”
The remaining 2 items asked clinicians about thedegree to which family wishes should be determina-tive in 2 kinds of cases. One question presented thestatement, “Life supports should be withdrawn froma child who is legally brain dead even if the parentsdo not agree.” Many respondents agreed with thestatement, including two thirds of house officers,critical care attending physicians, and nurses and�45% of hematology/oncology attending physi-cians, other specialty attending physicians, and otherspecialty nurses. However, substantial numbers ofrespondents disagreed or were uncertain.
Answers to the next question were also widelydivergent. Approximately 30% of critical care andother specialty attending physicians and 43% to 47%of all other groups agreed, “The organs of pediatricpatients who are on life support and permanentlyunconscious, but not technically brain dead, shouldbe used for transplantation if requested by the par-ents.” However, substantial numbers of respondentsdisagreed with the statement.
Beliefs That Vary According to SiteTable 7 shows the questions for which site differ-
ences were statistically significant. At 6 of the 7 in-stitutions, two thirds to three fourths of all respon-dents agreed with the statement, “Sometimes I feelwe are saving children who should not be saved.” At1 institution, however, only approximately onefourth of respondents felt that way. The same insti-tution also differed from other institutions in atti-tudes toward medically supplied food and water;there, 60% of respondents agreed, “Medically sup-plied food and water should always be continued.”That view was shared by only 31% to 49% of respon-dents at the other institutions. The proportions ofrespondents who agreed that there was inadequatepain medication for children at their institutionsranged from 11% to 51%.
DISCUSSIONThe survey revealed that a sizeable number of
physicians and nurses felt they had acted againsttheir consciences in providing treatment to the chil-dren in their care. The vast majority of concerns wererelated to the provision of overly burdensome treat-ment, not to giving up on children too soon. Theprevious survey of adult end-of-life care also docu-mented extensive concerns of conscience but, in thatearlier work, 5 times as many of the concerned clini-cians were worried about overtreatment as wereworried about undertreatment.14 In this pediatricsurvey, across all subspecialties, attending physi-cians were �10 times as worried and nurses were�20 times as worried about “saving children whoshould not be saved” as about giving up too soon.These findings suggest that concerns about overlyburdensome treatment are greater in pediatric end-of-life care than in adult end-of-life care.
Interestingly, fewer physicians and nurses in he-matology/oncology had concerns of conscience,compared with their counterparts in critical care.
ARTICLES 877 by guest on June 12, 2013pediatrics.aappublications.orgDownloaded from
Even among hematology/oncology clinicians, how-ever, approximately one third agreed, “Sometimes Ifeel we are saving children who should not besaved” and “Sometimes I feel the treatments I offerchildren are overly burdensome.” The differencesbetween critical care clinicians and hematology/on-cology clinicians warrant additional study. Perhapsmore intensivists have concerns of conscience aboutproviding overly burdensome treatments becausethey care for neurologically devastated patients forwhom restoration to a meaningful existence mayseem impossible. Conversely, fewer hematology/on-cology clinicians may express concerns of consciencebecause prognostic uncertainty for pediatric cancer isso great and therefore cure-oriented interventionsare more likely to be justifiable. Oncologists are alsomore likely to have had longer-term relationshipswith patients and families than intensivists, andthese emotional bonds may impair oncologists’ abil-ity to provide objective counseling and advice. Inter-estingly, Wolfe et al17 reported that oncologists knewa full 6 months before parents that death was virtu-ally certain but cure-oriented treatment plans contin-
ued. The difficulty of holding end-of-life conversa-tions with parents and the understandable desire toensure that everything that could be done was donemay account for the time lag between oncologists’and parents’ understanding of a dire prognosis.However, none of these factors fully explains whythe hematology/oncology clinicians we surveyedwere less likely to see life-sustaining treatments asburdensome to the children in their care than wereintensivists.
Other items in the survey may help explain theseconcerns of conscience. If clinicians do not think theycan stop treatments already underway, then theymay feel compelled to continue disproportionatelyburdensome interventions. In our survey, with theexception of critical care attending physicians, whowere more knowledgeable, �70% of respondentsmischaracterized or were uncertain about the pre-vailing national consensus that holds that there is nomorally relevant distinction between withholdingand withdrawing treatment.
Clearly, withdrawing treatments already begunmay feel different psychologically than never insti-
TABLE 7. Site-Based Variation
A B C D E F G �2 Test
Sometimes I feel we are saving children whoshould not be saved
The number of cases per site, which ranged from 44 to 210, is not provided to protect the confidentiality of the sites. Disagree includesrespondents who answered 1 or 2, uncertain includes respondents who answered 3, and agree includes respondents who answered 4 or5, on a scale ranging from 1 � strongly disagree to 5 � strongly agree.
880 CONTROVERSIES IN PEDIATRIC END-OF-LIFE CARE by guest on June 12, 2013pediatrics.aappublications.orgDownloaded from
tuting a treatment. However, when allowing death isjustifiable, parents and health care professionalsshould know that there is no sanction against with-drawing treatments already underway. If the view isthat any intervention already being offered cannot bewithdrawn, then many children may suffer need-lessly and opportunities for comfort care and forinteractions with parents and siblings may be dimin-ished. Conversely, if health care professionals do notrealize that they can later withdraw treatments thatare no longer beneficial, then they may be reluctantto initiate trials of treatments that could be beneficial.
Medically Supplied Nutrition and HydrationThe American Medical Association,18 the Ameri-
can Academy of Pediatrics,3 and the AmericanNurses Association4 have clarified that medicallysupplied nutrition and hydration are not ethicallyand legally required in all clinical circumstances.When their benefits outweigh their burdens andwhen their use is inconsistent with agreed-on goalsof care, medically supplied food and fluids can beforgone or withdrawn.
In our survey, however, more than one third ofcritical care attending physicians, one fourth of he-matologists/oncologists, and �40% of other subspe-cialty attending physicians, house officers, andnurses believed that medically supplied nutritionand hydration should always be continued. Al-though the survey item asked about “medically sup-plied food and water,” it is possible that, in answer-ing, some respondents did not distinguish betweenthe provision of food and fluids by hand and that bymedical means; indeed, food and water should al-ways be offered to children and adults who can takethem by mouth. Even if there was some confusion inhow the item was interpreted, this finding suggeststhat different members of the same care team mayhold radically different views about the permissibil-ity of withholding these interventions. Such diverseopinions could lead to conflicts among health careprofessionals, confusing messages for parents, andinappropriate prolongation of the dying process.
Given the symbolic nature of food and water,withdrawing or withholding medically supplied nu-trition and hydration for children is one of the mostemotionally wrenching decisions that can be made.Reluctance to forgo these interventions may also beattributable to the mistaken impression that forgoingartificial nutrition and hydration leads to an uncom-fortable death. In fact, medically supplied nutritionmay prolong the dying process and, near the end oflife, can cause congestion, excessive bodily secre-tions, and considerable discomfort.19 In certain in-stances, and with careful attention to symptom man-agement, a child’s interests may best be served byforgoing enteral and parenteral nutrition and hydra-tion.20,21
Pain ManagementOur data suggest that concerns about hastening
death affect clinicians’ pain management behaviors;approximately one half of the attending physiciansand �40% of house officers and nurses agreed,
“When clinicians give inadequate pain medication,they do so most often out of fear of hastening thechild’s death.” More than two thirds of the hematol-ogy/oncology nurses held this view. However, suchfear is misplaced on clinical, ethical, and legalgrounds.
Clinically, this fear is misplaced because datashowed that respiratory depression is extremelyrare, especially among children �2 months ofage.22–24 Indeed, Partridge and Wall25 showed thatpatients who receive effective pain relief may actu-ally live longer than patients whose pain remainsundertreated.
It is important to note, “In some patients, opioidsmust be given with caution. These patients includechildren (and adults) with altered mental status, dis-ordered control of respiration, and abnormal renaland hepatic functioning.”24(p713) However, eventhese patients can receive very high levels of opioidssafely if the drugs are administered gradually andwith titration to effect.24
Although it is rare, it is sometimes the case (par-ticularly in the ICU) that appropriate pain manage-ment shortens the dying process. However, the ap-propriateness of providing adequate levels of painmedication, even if it hastens death, is both ethicallyand legally well established, with approval by the USSupreme Court.26
The preponderance of our respondents seemed tobe in agreement with the longstanding ethical prin-ciple of “double effect,”27 but substantial numberswere uncertain about its legality. For example, mostof the nurses agreed, “Sometimes it is appropriate togive pain medication to relieve suffering even if itmay hasten a child’s death.” However, one third ormore believed that it is illegal to do so, and 17% wereuncertain about its legality. Attending physicianswere the most well informed but, among hematolo-gy/oncology and other specialty attending physi-cians, �17% believed that it is illegal to give painmedication in doses that risk respiratory depressionto the point of death, and another 10% were uncer-tain. Health care institutions, professional medicaland nursing societies, and palliative care expertsneed to be relentless in reinforcing the message thatthe provision of adequate pain medication is appro-priate (indeed mandatory) on clinical, moral, andlegal grounds.
Given that approximately one third of pediatricintensivists and critical care nurses either were un-certain or thought that it was appropriate to provideneuromuscular blocking agents (such as Pavulon)with analgesia before withdrawing a ventilator, PI-CUs should consider auditing their practices anddeveloping policies to guide the use of paralyticagents. These agents have no analgesic or sedativeproperties; therefore, clinicians can never logicallyjustify introducing them at the end of life by claimingthat the intention is to relieve the patient’s pain andsuffering. If patients have been receiving theseagents as part of their therapeutic care, then neuro-muscular function should be restored before life sup-port is withdrawn, to allow for optimal assessmentof the patients’ comfort during the withdrawal pro-
ARTICLES 881 by guest on June 12, 2013pediatrics.aappublications.orgDownloaded from
cess. In rare cases, when the patient’s death is ex-pected to be certain and rapid after withdrawal ofventilatory support even if neuromuscular functionis restored, and if the time required to restore func-tion would subject the patient to unwarranted suf-fering from the continuation of life-sustaining treat-ments, then withdrawal may occur in the presence ofneuromuscular blockade, with careful attention toproviding sufficient analgesia and sedation to ensurethe comfort of the patient despite the presence ofpharmacologic paralysis.28,29
Family Prerogatives With Respect to Artificial SupportAfter Brain Death
The diagnosis of brain death is determinative ofdeath in all 50 states; therefore, physicians have noobligation to continue medical interventions whenpatients have died according to neurologic criteria.12
Two states, namely, New York and New Jersey, havespecial provisions for accommodating religious ob-jections to brain death. But what of parents whosimply may need more time to accept the fact thattheir child has died? Do objections to withdrawal ofartificial support for a brain-dead patient merit ac-commodation if those objections are based on par-ents’ emotional needs, rather than religious objec-tions?
Although there are no legal grounds for families todemand continued treatment of the deceased, it maybe appropriate to allow time for parents to take in thedevastating news, especially because brain death isso often attributable to a sudden unexpected event. Atime-limited approach can help families come togrips with the tragic reality and begin the process ofgrieving, which may remain blocked as long as par-ents are in conflict with health care providers. In oursurvey, approximately one half of hematology/on-cology nurses, one third of hematology/oncologyattending physicians, other specialty attending phy-sicians, and other specialty nurses, and one fifth ofhouse officers, critical care attending physicians, andcritical care nurses seemed willing to make suchaccommodation. Given the substantial number of“uncertain” responses, it is possible that even greaternumbers might be willing to do so.
Family Prerogatives for Organ Donation and the Dead-Donor Rule
The Uniform Anatomical Gift Act clearly estab-lishes that only the organs of deceased persons canbe donated for organ transplantation. However, size-able proportions of respondents agreed with thestatement, “The organs of pediatric patients who areon life support and permanently unconscious, butnot technically brain dead, should be used for trans-plantation if requested by the parents.” One inter-pretation is that respondents are simply confusingpermanent unconsciousness with brain death. A sec-ond alternative is that they know that permanentunconsciousness is distinct from brain death but theybelieve that parental preferences should be determi-native, given the child’s profound degree of neuro-logic damage. Perhaps, in addition to honoring pa-rental choice, these clinicians are taking a basically
consequentialist perspective, hoping to see somegood come from a terrible situation. In this light,responses to this item might be interpreted as prin-cipled objections to the dead-donor rule.
Other researchers have also documented clini-cians’ belief that the organs of permanently uncon-scious patients could be used for transplantation. Forexample, in a 1996 national survey of neurologistsand medical directors, Payne et al30 found that twothirds of respondents believed that it would be eth-ical to use the vital organs of patients in a persistentvegetative state for transplantation. However, nei-ther prior research nor our study helps explain whythese clinicians hold these views. Additional empir-ical studies are needed to discover the reasons whyso many clinicians seem to feel that parental choice,not the dead-donor rule, should be determinative inthese heart-wrenching cases.
LimitationsThere are many other ethical issues regarding pe-
diatric end-of-life decision-making that sensitive cli-nicians and committed institutions will need to ex-amine; in this article, we reported on only some ofthe issues our survey probed. Future papers willaddress other components, such as the discretionadolescents should be granted in making decisionsregarding termination of treatment and the role ofcaregiver suffering in pediatric palliative care.
Our results are also limited in that we did not havea probability-based sample of institutions. This lim-itation is mitigated by the facts that the institutionswere geographically diverse and substantial site dif-ferences were detected, which indicated variation inthe sample.
CONCLUSIONSThis research revealed serious concerns of con-
science about the provision of overly burdensometreatments for gravely ill children, which suggeststhat there are powerful barriers to establishing ap-propriate goals of care. The survey also reveals lackof knowledge regarding basic ethical guidelines forpediatric end-of-life decision-making among a sam-ple of clinicians who rated their knowledge of med-ical ethics quite high. This suggests the need for moreethics education and also more clinical educationregarding pain management, medically supplied nu-trition and hydration, and the use of paralytic agents.
The findings also bring attention to the fact thathealth care professionals (both within and acrossdisciplines and subspecialties) can and do hold dif-ferent points of view on a range of inherently com-plex issues in the care of gravely ill children. Anethically strong culture within hospitals should en-courage diversity of opinions, but the opinionsshould be well informed. Health care professionalsshould be aware of the ethics literature and the pub-lished guidelines, should know why they hold theviews they do, and should be familiar with alterna-tive positions. Surveys such as this one can demon-strate how many different points of view health careprofessionals working within the same institutionbring to the clinical choices they face. Findings can be
882 CONTROVERSIES IN PEDIATRIC END-OF-LIFE CARE by guest on June 12, 2013pediatrics.aappublications.orgDownloaded from
used to prompt individual reflection, interdiscipli-nary dialogue, and institutional self-examination.
ACKNOWLEDGMENTSThis research was supported by the Nathan Cummings Foun-
dation, through a grant to the Center for Applied Ethics andProfessional Practice at Education Development Center, duringthe early needs-assessment phase of the Initiative for PediatricPalliative Care (www.ippcweb.org). The Initiative for PediatricPalliative Care has been supported by the Nathan CummingsFoundation, the Open Society Institute Project on Death in Amer-ica, the Kohlberg Foundation, and the Argosy Foundation.
We thank Erica Jablonski for assistance with survey prepara-tion and data analysis. Thanks also go to Jane Barnsteiner, BrianCarter, Gina Emge, Polly Hesterberg, Javier Kane, Elaine Meyer,Beverly Ryan, Kay Sandor, and Jane Schneck for administering thesurveys. We are grateful to Andrea Kydd of the Nathan Cum-mings Foundation for her original vision and continuing support.We thank all of our funders for enabling us to bring these findingsto the attention of hospitals across the country and for theircombined support of the next training phase of this nationalinitiative.
REFERENCES1. Duff RS, Campbell AGM. Moral and ethical dilemmas in the special-
care nursery. N Engl J Med. 1973;289:890–8942. President’s Commission for the Study of Ethical Problems in Medicine
and Biomedical and Behavioral Research. Deciding to Forego Life-Sustaining Treatment: A Report on the Ethical, Medical, and Legal Issues inTreatment Decisions. Washington, DC: US Government Printing Office;1983
3. American Academy of Pediatrics, Committee on Bioethics. Guidelineson forgoing life-sustaining medical treatment. Pediatrics. 1994;93:532–536
4. American Nurses Association. Position statement on foregoing artificialnutrition and hydration. Ky Nurse. 1993;41:16
6. Annas GJ. Asking the courts to set the standard of emergency care: thecase of Baby K. N Engl J Med. 1994;330:1542–1545
7. Clayton EW. What is really at stake in Baby K? A response to EllenFlannery. J Law Med Ethics. 1995;23:13–14
8. Flannery EJ. One advocate’s viewpoint: conflicts and tension in theBaby K case. J Law Med Ethics. 1995;23:7–12
9. Glover JJ, Rushton CH. From Baby Doe to Baby K: evolving challengesin pediatric ethics. J Law Med Ethics. 1995;23:5–6
10. Paris JJ, Crone RK, Reardon F. Physicians’ refusal of requestedtreatment: the case of Baby L. N Engl J Med. 1990;322:1012–1015
11. Cassidy R. A time to die: ethical, legal and professional grounds forjudgment. In: Cassidy R, Fleischman A, eds. Pediatric Ethics: From Prin-ciples to Practice. Amsterdam, The Netherlands: Harwood AcademicPublishers; 1996:32–33
12. National Conference of Commissioners on Uniform State Laws. UniformAnatomical Gift Act of 1987. Newport Beach, CA: National Conference ofCommissioners on Uniform State Laws; 1987
13. Field M, Behrman R, eds. When Children Die: Improving Palliative andEnd-of-Life Care for Children and Their Families. Washington, DC: NationalAcademies Press; 2003
14. Solomon MZ, O’Donnell L, Jennings B, et al. Decisions near the end oflife: professional views on life-sustaining treatments. Am J Public Health.1993;83:14–23
15. Asch DB, Jedrziewski MK, Christakis NA. Response rates to mail sur-veys published in medical journals. J Clin Epidemiol. 1997;50:1129–1136
16. Lensing SY, Gillaspy SR, Simpson PM, Jones SM, James JM. Encourag-ing physicians to respond to surveys through the use of fax technology.Eval Health Prof. 2000;23:349–360
17. Wolfe J, Klar N, Grier HE, et al. Understanding of prognosis amongparents of children who died of cancer: impact on treatment goals andintegration of palliative care. JAMA. 2000;284:2469–2475
18. American Medical Association, Council on Ethical and Judicial Affairs.Decisions near the end of life. JAMA. 1992;267:2229–2233
19. Miller-Theil J, Glover J, Beliveau E. Caring for the dying child. In:Gallagher C, Amenia M, eds. Nutrition and Hydration in Hospice Care:Needs, Strategies, and Ethics. Binghamton, NY: Haworth Press; 1993:55–72
20. Gemma PB, Penticuff JH, Rushton CH, Will J, Dokken D, Delo B.End-of-Life Care for Children and Their Families: Ethical Dimensions. Re-search Triangle Park, NC: Glaxo Wellcome; 1997
21. Nelson LJ, Rushton C, Cranford RE, et al. Forgoing medically providednutrition and hydration in pediatric patients. J Law Med Ethics. 1995;23:33–46
22. Hertzka RE, Gauntlett, JS, Fisher DM. Fentanyl-induced ventilatorydepression: effects of age. Anesthesiology. 1989;70:213–218
23. Lynn AM, Nespeca MK, Opheim KE, et al. Respiratory effects of intra-venous morphine infusions in neonates, infants, and children aftercardiac surgery. Anesth Analg. 1993;77:695–701
24. Galloway KS, Yaster M. Pain and symptom control in terminally illchildren. Pediatr Clin North Am. 2000;47:711–746
25. Partridge JC, Wall SN. Analgesia for dying infants whose life support iswithdrawn or withheld. Pediatrics. 1997;99:76–79
26. Vacco, Attorney General of New York, et al vs Quill et al: certiorari tothe United States Court of Appeals for the Second Circuit, No. 95–1858.Argued US Supreme Court January 8, 1997; decided June 26, 1997. Avail-able at: http://straylight.law.cornell.edu/supct/html/95-1858.zs.html.Accessed August 22, 2005
27. Hastings Center. Guidelines on Termination of Treatment and Care of theDying. Bloomington, IN: Indiana University Press; 1987
28. Truog R, Burns J, Mitchell C, Johnson J, Robinson W. Pharmacologicparalysis and withdrawal of mechanical ventilation at the end of life.N Engl J Med. 2000;342:508–511
29. Rushton CH, Terry P. Neuromuscular blockade and ventilatorwithdrawal: ethical controversies. Am J Crit Care. 1995;4:112–115
30. Payne K, Taylor RM, Stocking C, Sachs GA. Physicians’ attitudes aboutthe care of patients in the persistent vegetative state: a national survey.Ann Intern Med. 1996;125:104–110
USA—DOING OK?
“Americans work harder than almost anyone else—300 hours a year more thanEuropeans. They switch jobs more often—about once every seven years, comparedwith once every 11 years in German and Japan. . . . [T]he quality and quantity ofintellectual life in America is still the highest in the world. A quarter of Americanadults have a university education. The country produces one-third of the world’sscientific papers, employs two-thirds of the world’s Nobel-prize winners, has 17of the top 20 universities (as ranked by Shanghai’s Jiao Tong University) and hasmore ideas-based workers than anyone else.”
The Economist. July 16, 2005
Noted by JFL, MD
ARTICLES 883 by guest on June 12, 2013pediatrics.aappublications.orgDownloaded from
f_care_subhttp://pediatrics.aappublications.org/cgi/collection/standard_oStandard of Carethe following collection(s):This article, along with others on similar topics, appears in
Permissions & Licensing
mlhttp://pediatrics.aappublications.org/site/misc/Permissions.xhttables) or in its entirety can be found online at: Information about reproducing this article in parts (figures,
