OBJECTIVE. Legal and ethical standards require resuscitation when it is considered to be in the patient's best interest. We hypothesized that newborn infants might be dealt with according to different standards, compared with older patients.
METHODS. An anonymous questionnaire describing 8 currently incompetent patients with potential neurologic sequelae who required resuscitation was administered to groups of physicians and students. Survival and morbidity rates were explicitly described; a very preterm infant, a full term infant and a 2-month-old infant had identical outcomes. Two patients, a 7-month-old infant and an 80-year-old patient, were previously significantly impaired. Respondents were asked whether resuscitation was in the patient's best interest and whether they would comply with the families' wishes if resuscitation was refused.
RESULTS. There was an 85% response rate (n = 524). The largest proportions of respondents stated that it was in the best interests of the 2-month-old infant and the 7-year-old child to be resuscitated (97% and 94%, respectively), followed by the 50-year-old patient and the term infant (87%), the 2 patients with 5% chance of survival (76% and 80%), the premature infant (69%), and finally the 80-year-old patient (32%). Approximately one fifth of the respondents who thought that it was in a patient's best interests to be resuscitated would nevertheless accept the family's refusal of resuscitation for all scenarios except the 80-year-old patient (72% acceptance) and the preterm infant (54% acceptance).
CONCLUSIONS. Whether resuscitation is considered in a patient's best interests is not closely related to survival rates or disability. Newborn infants and particularly preterm infants are systematically devalued, in comparison with older patients whose outcomes are the same or worse. Accepting a family's refusal of resuscitation, even among respondents who thought that resuscitation was in the patient's best interest, was much more common for the newborns.
Technologic progress has increased the limits of medicine and enables many patients to survive or to prolong their lives. Some life-threatening events, for example, meningitis, trauma, and asphyxia, can lead to potentially serious neurologic sequelae and to complex decision-making involving withholding or withdrawal of care. For incompetent patients, these decisions are often made by a surrogate, most often a family member.
Ideally, the decision-making process involving the caregivers and the family of the incompetent patient proceeds smoothly. In reality, individual interpretations of risks, benefits, and outcomes can vary depending on religious beliefs, cultural factors, and ages. When controversies and disagreements exist and the patient has left neither advanced directives nor evidence of any preferences, the patient's surrogate, usually a family member, is consulted. Surrogates are expected to use the patient's best interest as the guiding principle for decision-making.1,2 Treatments are considered to be in patients' best interests when the benefits of treatment outweigh the risks and burdens. Indirectly, the value of the life of the patient is judged during decision-making. If the caregivers think “that the designated surrogate threatens the patient's best interests, the decision should be overridden,”1 and such overriding of family decision-making has been supported in the courts.1,3,4 If resuscitation is considered to be in an incompetent patient's best interest and significant harm would be caused by not intervening, then refusal of care generally is not accepted, either ethically or legally. We wished to determine whether the same decision-making process would be followed for preterm and term infants as for older patients and whether decisions are made on the basis of outcomes or according to other criteria. The goals of this study were to describe how different groups of educated individuals valued patients' best interests, to determine whether survival, disability, poor neurologic prognosis, and age affected these estimates, and to determine whether the best-interest principle is generally followed (ie, if resuscitation is considered to be in the best interest of an incompetent patient, then how often would family demands to withhold care be accepted?).
Between February 2005 and January 2006, we administered an anonymous questionnaire to physicians at McGill University Health Center who were involved in resuscitation decisions. The same questionnaire was administered to university students in different disciplines, including law, anthropology, bioethics, and medicine. Students in law, medicine, and anthropology were first-year students at McGill University. The bioethics students were postgraduate students at the University of Montreal. For the target groups of respondents, our goal was to have >80% participation from the full-time physicians and from the students.
Questionnaires were administered during a group activity where a maximal number of respondents from each target group could be reached, such as a resident/student teaching session or a staff meeting for physicians. We obtained the authorization of the group leader before distributing the questionnaire, and we attended a group activity that was most convenient for all respondents. The first question was a request for consent; if consent was refused, then the questionnaire was collected blank with the others at the end of the period. The first page of the questionnaire contained demographic information (ie, gender, work environment, occupation, residency years, and having children or not). The questionnaires were completed individually and could not be taken home.
Eight scenarios of currently incompetent, critically ill patients with potential neurologic sequelae were presented. All arrived in the emergency department of a university health center when a family member cannot be immediately consulted. The patients were of different ages, and their outcomes were explicitly described; gender or other social information (such as marital status) was not provided (Table 1).
There were 4 patients with a 50% chance of survival, that is, a premature infant at 24 weeks of gestation who has just been delivered, an infant just born at term with a known brain malformation, a 2-month-old infant with bacterial meningitis, and a 50-year-old patient with severe trauma, including head injury, resulting from a car accident. If these patients survived, 50% would be without disability and 25% severely impaired. Two other patients were previously disabled, namely, a 7-year-old patient with multiple disabilities (cerebral palsy, deafness, learning disability, and hyperactivity) and new head trauma and an 80-year-old patient with substantial disability from dementia and a new stroke. Both patients were noted to have a 50% predicted chance of survival and, if they survived, a 50% chance of having additional disability. Two patients were described as having only a 5% chance of survival, that is, a 14-year-old patient with acute myeloid leukemia with central nervous system involvement, with a 20% risk of disability in case of survival, and a 35-year-old patient with brain cancer with a 100% risk of disability with treatment.
The patients were presented in order from the youngest to the oldest. After each patient description, the following questions were asked. (1) Do you think intubating, resuscitating, and consulting intensive care for admission is in the patient's best interest? (2) If the parents/family asked you not to resuscitate, would you respect their decision? (3) If the patient was your child/partner, would you wish the physician to intubate, to resuscitate, and to consult intensive care for admission? (4) If the patient was your sibling's or good friend's child/partner and he or she had a few minutes to consider the decision and asked for your opinion, would you recommend that the physician intubate, resuscitate, and consult intensive care for admission? (5) If the patient were you and you were able to decide, would you want the physician to intubate, to resuscitate, and to consult intensive care for admission? (The last question was asked only for scenarios in which the hypothetical patients were adults.) For each of these questions, respondents could respond on a Likert scale, with the following options: always, generally, exceptionally, or never. Always and generally were counted positive answers.
Proportions were compared by using the χ2 test with Yates correction; P values of >.01 are reported as not significant. To limit the number of comparisons and to protect against type I errors, in addition to limiting the critical P value for each comparison to .01, the following analyses were performed. For each of the questions, we compared the proportions answering always or generally for the 8 patients with a single χ2 test and examined comparisons between scenarios only when the overall χ2 was significant at P < .01. Similarly, in comparisons of responses between groups of respondents for a particular question, the overall χ2 had to be significant at P < .01 before individual groups were compared. This study was approved by the McGill University institutional review board.
A total of 615 individuals were in our target groups; 527 received the questionnaire, with 524 respondents (response rate: 85% of the target groups and 99% of participants) (Table 2). All groups had similar response rates. Fewer students than physicians had children (P < .0001). Of the 357 student respondents, 107 were in law, 88 in anthropology, 139 in medicine, and 23 in bioethics. Of the 95 attending staff members, 32 were in family medicine, 23 in obstetrics, 12 in neonatology, and 28 in emergency medicine (adult and pediatric). There were 86 residents, 20 in obstetrics, 33 in pediatrics, and 19 in family medicine.
The 2-month-old infant and the 7-year-old child with multiple disabilities had the largest proportions of respondents stating that it was always or generally in their best interest to be resuscitated (97% and 94%, respectively; not significant in comparison with each other; P < .001 in comparison with all other patients), followed by the 50-year-old patient and the term infant (both 87%; P < .01 in comparison with all other patients). For the 2 patients with a 5% chance of survival, 76% and 80% of responses were always or generally (not significant in comparison with each other; P < .005 in comparison with all other patients). For the preterm infant, 69% estimated that it was in the patient's best interests to be resuscitated (P < .001 in comparison with all other patients). Finally, the smallest proportion of respondents estimated that it was in the best interests of the 80-year-old patient to be resuscitated (32%; P < .001 in comparison with all other patients) (Fig 1).
If It Were Your Own Child, Partner, or Sibling
The proportions desiring resuscitation for their own child, partner, and sibling were almost identical and were not significantly different from the proportions thinking it to be in the patient's best interest to be resuscitated. Most groups of respondents wanted resuscitation for their preterm infant significantly less than for all other children, with neonatologists being exceptions (92%, same as for other infants). Ethics students wanted to resuscitate their own child, partner, or sibling less than all other groups for all scenarios (P < .01) except for the 80-year-old patient, for whom the responses were not significantly different.
If It Were You
The respondents wanted resuscitation for themselves in smaller proportions than for their partner or sibling for all 3 adult scenarios (P < .01 for all comparisons). For example, for the 80-year-old scenario, 36% would never want resuscitation for themselves, compared with 15% for a sibling or partner (P < .001). For these 3 questions, there was no significant difference between the students' answers and those given by physicians. The answers were not different between respondents who had children and those who did not.
Best Interests Versus Accepting Nonresuscitation
The 80-year-old patient was the patient with the most respondents accepting the family's wish not to resuscitate (88%; P < .001 in comparison with all other patients), followed by the preterm infant (66%; P < .001 in comparison with all other patients) and then by the 2 patients with a 5% chance of survival (32% and 35%; P < .01 in comparison with the 50-year-old patient) and by the 50-year-old patient. In contrast, the 2-month-old infant and the 7-year-old child were the patients who had the largest proportions of respondents not accepting the withholding of resuscitation (15% and 18%, respectively; not significant in comparison with each other; P < .005 in comparison with all other patients).
Not surprisingly, the large majority of respondents who thought it was exceptionally or never in the patient's best interests to be resuscitated would accept family refusal of intervention for each of the scenarios (>95% would always or generally accept refusal, for all scenarios). Among those who stated that it was always or generally in the patient's best interest to be resuscitated, between 15% and 22% would accept family refusal of intervention for each of the scenarios except for the newborn infants and the 80-year-old patient; significantly more would accept withholding care for the newborns and the 80-year-old patient (P < .0001) (Table 3).
When findings were analyzed only among the smaller number of respondents who thought that resuscitation was always in the patient's best interest, there remained a statistically significant variation in the numbers of those who would accept the family's refusal to resuscitate (P < .001). The individual comparisons that were statistically significant were the 80-year-old patient in comparison with all other patients (P < .0001) and the preterm infant in comparison with the 2-month-old infant and the 7-year-old child (P < .001) (Table 3).
Among those who thought that it was in the patient's best interest to be resuscitated, physicians accepted refusal of care statistically more often than students for all patients (P < .01) except the 80-year-old patient, for whom there was no difference. Among the students, the bioethics students stood out; they were much more likely to accept family refusal for all cases, even when they thought that resuscitation was in the patient's best interests (P < .01) (Fig 2). The answers were not different between respondents who had children and those who did not.
Decisions about initiating or withdrawing intensive care often are based on the best-interest principle, which takes risks and benefits in consideration. In this study, the majority (69%) thought that resuscitation was in the best interest of a 24-week preterm infant, but a significantly larger majority thought that resuscitation was in the best interests of a term infant and a 2-month-old infant with identical outcomes, an already seriously impaired child, an older child with 5% chance of survival, and even an adult with 5% chance of survival and 100% chance of permanent disability. The 80-year-old patient who already had moderately severe dementia was the only case with fewer respondents than for the preterm infant considering it in the patient's best interest to be resuscitated. Therefore, there seems to be a different value placed on life at its extremes, which, with the potential for many years of life to be gained, is particularly surprising for those just born. Survival rates seemed to have some influence on estimates of best interests but, for the 2 patients with 5% predicted chance of survival, including the patient with no chance of intact survival, significantly more respondents than for the premature infant thought it was in the patient's best interest to be resuscitated. Disability did not seem to influence respondents, because the 7-year-old child who already had multiple disabilities had the largest proportion of respondents estimating that it was in the patient's best interests to be resuscitated.
We found that a judgment of best interests was closely related to a desire to treat one's own child or partner; this may be a sign that what we want for our own families becomes the approach that we consider to be in the best interests of patients in general. It was also interesting that, for all of the adult scenarios, more respondents would want resuscitation for their partner or sibling than for themselves. Does this reflect a sense of duty to those to whom we are closest?
It is generally accepted that a treatment that is considered to be in a patients' best interest should be pursued.1–4 Several authors contest this, however, especially in the field of incompetent or pediatric patients, and give more importance to minimizing harm.5,6 For example, although vaccination is clearly in a child's best interests, one does not remove a child from his or her parents' custody solely because of refusal of basic vaccinations, because removing a child from the family may do much more harm.5,6 For the patients in our scenarios, the harm risked through withholding care is death, and how respondents balance that potential harm with the “harm” of either intact survival or the possibility of lifelong handicap warrants additional investigation. Others have suggested that a standard of “reasonableness” would be preferable7; however, such a standard also would require a subjective assessment of whether it is reasonable to allow the infant to die, rather than having a chance of survival with an uncertain future.
Our results showed that, for all patients, ∼1 of 5 respondents would be prepared to withhold resuscitation at the families' request, even when the respondents considered resuscitation to be in the best interests of the patient. We were surprised that the proportions were identical among all subgroups, including anthropology and law students, who had not received any formal training in this sort of decision-making, and staff physicians, who presumably had been exposed to bioethical reasoning. The exception was the bioethics students, who were prepared in a very large proportion to withhold a life-saving emergency intervention that they considered to be in an incompetent patient's best interest at the request of the family; this could point to the importance that autonomy has in the bioethics curriculum or the underlying values of those who self-select to enter the field of bioethics.
This abrogation of the best-interest principle was most marked for the 2 patients at the extremes of life, for whom more than one half of respondents who thought that resuscitation was in the patient's best interests would accept family refusal of intervention. The differences between patients were also significant even among those who were more definite and answered that it was always in the patient's best interest to be resuscitated; the 80-year-old patient and the preterm infant still elicited a greater willingness to withhold resuscitation than did the other patients. It seems that accepting withholding of resuscitation is somehow easier for the very old and the very young. Waring,8 a philosopher and bioethicist, remarked that the value of persons is indicated by how we react to their deaths, “feelings of tragedy, evil, loss and sharp regret are supposedly more appropriate responses to the deaths of younger people.” One might view the deaths of older people as tolerable. Indeed, it is not rare to hear that it is “better this way, nature took its course” or “he lived long enough” for an older individual.8 Similar statements (“it is better this way, nature took its course” or “at least she didn't suffer”) are also made for premature infants by staff members and by families. Perhaps the premature infant has not yet lived long enough?
For resuscitation decisions, some authors think that age should have an impact in decision-making and have proposed setting an upper age limit for resuscitation9,10; this is very controversial.11–13 There seem to be no official policies or professional association guidelines that suggest an age limit for resuscitation, with the exception of premature infants.14–18 At 24 weeks of gestation, many national associations deem prognosis to be so poor that life-saving interventions are considered optional and performed only with explicit family consent.19 The survival chance and the probabilities of long-term disability for the 24-week patient in our survey were designed to reflect conservatively the actual chances of an infant delivered at 24 weeks of gestation in a Canadian tertiary care center.20–25 We can find no other population in the literature for which a 50% survival rate and a 50% rate of normal outcomes among survivors would be seen as dismal enough to defer life and death decisions entirely to the family. Policy statements for preterm infants often state survival and disability rates as justification for optional intervention,13–17 but our results suggest that underlying, widely held beliefs about the value of the life of a newborn may have a major influence on such recommendations.
The reduced value placed on the life of newborns, and particularly the preterm infant, by our respondents is less than expected on the basis of any objective medical data. In fact, this limitation of their value is probably not related to medical factors, because physicians and students had almost identical answers. What does explain these differences? It has been suggested that the proximity of gestational ages at which abortions are performed is relevant to resuscitation decisions.26 This suggests that decisions regarding life and death for newborns, especially premature infants, may still be in the realm of reproductive choice and essentially considered a family decision. Perhaps there is a diminished sense of duty toward the premature infant. This would explain why so many respondents would be ready to accept withholding resuscitation even if they thought that resuscitation was in the patient's best interest. Several authors also suggested that newborns do not have the same status as older individuals, because they lack personhood.27–31 Although there is no general agreement on this issue, our responses fit a general assessment that newborns were of less value. Even the respondents who had children and had personally experienced the gradual development of personhood in an infant had the same answers. Is the systematic devaluation of newborns attributable to more-deeply rooted anthropologic, cultural, social, and evolutionary factors? Until recently, most parents had experienced the death of a newborn or an infant, often several times; perhaps the necessary protective mechanisms to avoid continual grief over the loss of newborns have caused us to devalue them.
Because of our high response rate, we know that the results accurately reflect the opinions of our target population, but these cannot be directly extrapolated to other groups or cultures. Also, all questionnaire studies have their inherent limitations. We do not know whether opinions would translate into actions if the respondents were in the specific situations. In this study, the scenarios were variably unrealistic; very few preterm infants are delivered in an emergency department, for example. Even for this scenario, however, it does occasionally happen that a mother delivers in the emergency department and we must decide whether to initiate resuscitation without discussing the issue at length with the mother. More importantly, this was a conscious decision to make the scenarios as consistent as possible except for the ages and outcomes of the patients. Also, the order of presentation of the cases was the same for all respondents in this study. We considered trying to randomize the order, but this was not feasible logistically. We decided to present the scenarios in the order of the patient's age because this would seem logical and unbiased to the participants, but we cannot rule out the possibility that a different order of presentation might have produced some differences in the results.
The range of caregivers and students we approached reflected the groups for which we could use a method that has provided us with very high response rates in the past and therefore eliminates the common concern regarding response bias in questionnaire studies. We were able to obtain agreement from directors of programs and courses to approach their students, residents, or staff members. This led to the choices we made regarding groups of respondents, which we tried to make as varied as possible to obtain a wide representation. We recognize that some respondents are rarely or never involved in these decisions during their working lives. However, the remarkable consistency between most groups of physicians and students suggests that these attitudes have little to do with training or experience. The groups of physicians and students were very heterogeneous but, to avoid inflating the number of exploratory analyses, we did not search out differences between some of the possible subgroups, Therefore, it may be that important systematic differences in responses were masked. Additional research is needed to explore the underlying reasons for these responses. Neonatologists in particular might respond differently to some of these ethical dilemmas in very preterm infants, because they might have had frequent actual experiences with such decisions.32 Our initial sample of neonatologists was too small to draw any conclusions. Estimating an individuals' best interest indirectly demands placing a value on his or her life; the results of our study show that the lives of newborn infants, particularly preterm infants, are systematically devalued in comparison with others, both by physicians and by a nonmedical educated population.
- Accepted September 20, 2007.
- Address correspondence to Keith James Barrington, MB ChB, Department of Pediatrics, McGill University, Room C7.68, Royal Victoria Hospital, 687 Pine Ave W, Montreal, Quebec H3A 1A1, Canada. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
What's Known on This Subject
If resuscitation is considered to be in an incompetent patient's best interest and significant harm would be caused by not intervening, then refusal of care is generally not accepted, either ethically or legally.
What This Study Adds
Whether resuscitation is considered in a patient's best interests is not closely related to survival rates or disability. Newborn infants and particularly preterm infants are systematically devalued, compared with older patients whose outcomes are the same or worse.
- ↵Beauchamp TL, Childress JF. Principles of Biomedical Ethics. 5th ed. New York, NY: Oxford University Press; 2001:102– 103
- ↵Sneiderman B, Irvine JC, Osborne PH. The mentally incompetent patient. In: Canadian Medical Law: An Introduction for Physicians, Nurses and Other Health Professionals. 3rd ed. Toronto, Canada: Thomson-Carswell; 2003:501– 577
- ↵American Academy of Pediatrics, Committee on Bioethics. Guidelines on forgoing life-sustaining medical treatment. Pediatrics.1994;93 (3):532– 536
- ↵President's Commission for the Study of Ethical Problems in Medicine and Biomedical and Behavioral Research. Deciding to Forego Life-Sustaining Treatment. Washington, DC: US Government Printing Office; 1983
- ↵American Academy of Pediatrics, Committee on Bioethics. Religious objections to medical care. Pediatrics.1997;99 (2):279– 281
- ↵Silverman WA. Medical decisions: an appeal for reasonableness. Pediatrics.1996;98 (6):1182– 1184
- ↵Waring D. Fair innings and need over lifetime. In: Medical Benefit and the Human Lottery: An Egalitarian Approach to Patient Selection. Toronto, Canada: Springer; 2004:89– 97
- ↵Callahan D. Setting Limits: Medical Goals for an Aging Society. New York, NY: Simon and Schuster; 1987:17
- ↵Jahnigen DW, Binstock RH. Economic and clinical realities: heath care for elderly people. In: Binstock RH, Post SG, eds. Too Old for Health Care? MD: Johns Hopkins University Press; 1991:17 Controversies in Medicine, Law, Economics, and Ethics. Baltimore: 23–30
- Griffin J. Well-Being: Its Meaning, Measurement, and Moral Importance. Oxford, England: Clarendon Press; 1986
- ↵Waring D. An egalitarian ethos. In: Medical Benefit and the Human Lottery: An Egalitarian Approach to Patient Selection. Toronto, Canada: Springer; 2004:115– 131
- ↵Danish Council of Ethics. Extreme Prematurity, Ethical Aspects. Copenhagen, Denmark: Danish Council of Ethics; 1995
- Swiss Society of Neonatologists. Guidelines: recommendation for the care of infants born at the limit of viability (gestational age 22–26 weeks). Available at: www.neonet.ch/assets/doc/Infants_born_at_the_limit_of_viability_-_english_final.pdf. Accessed April 3, 2007
- Salle B, Sureau C. The premature of less than 28 weeks, resuscitation and outcomes: report of the academy of medicine, 2006 [in French]. Available at: www.academie-medecine.fr/upload/anciens/rapports_289_fichier_lie.rtf. Accessed March 18, 2008
- ↵Nuffield Council on Bioethics. Critical care decisions in fetal and neonatal medicine: ethical issues. Available at: www.nuffieldbioethics.org/go/ourwork/prolonginglife/publication_406.html. Accessed April 3, 2007
- ↵Canadian Paediatric Society, Fetus and Newborn Committee, Society of Obstetricians and Gynaecologists of Canada, Maternal-Fetal Medicine Committee. Management of the woman with threatened birth of an infant of extremely low gestational age. CMAJ.1994;151 (5):547– 553
- ↵American Academy of Pediatrics. Textbook of Neonatal Resuscitation. 5th ed. Elk Grove Village, IL: American Academy of Pediatrics; 2006
- Lorenz JM, Wooliever DE, Jetton JR, Paneth N. A quantitative review of mortality and developmental disability in extremely premature newborns. Arch Pediatr Adeolesc Med.1998;152 (5):425– 435
- ↵Hintz SR, Kendrick DE, Vohr BR, Poole WK, Higgins RD, National Institute of Child Health and Human Development Neonatal Research Network. Changes in neurodevelopmental outcomes at 18 to 22 months' corrected age among infants of less than 25 weeks' gestational age born in 1993–1999. Pediatrics.2005;115 (6):1645– 1651
- ↵Boyle RJ, Salter R, Arnander MW. Ethics of refusing parental requests to withhold or withdraw treatment from their premature baby. J Med Ethics.2004;30 (4):402– 405
- Singer P. Practical Ethics. Cambridge, England: Cambridge University Press; 1979:137
- Engelhardt HT. Ethical issues in aiding the death of young children. In: Kohl M, ed. Beneficent Euthanasia. Buffalo, NY: Prometheus; 1975:185– 201
- ↵Tooley M. Personhood. In: Singer P, Kuhse H, eds. A Companion to Bioethics. Oxford, England: Blackwell Publishers; 1998:117– 126
- Copyright © 2008 by the American Academy of Pediatrics