The Future of Children, Princeton - Brookings: Providing research and analysis to promote effective policies and programs for children.

Legal questions aside, what care should be given to the most immature, malformed, or sick premature infants? The important relevant ethical principles include preservation of life, alleviation of suffering, nonmaleficence ("do no harm"), patient autonomy, and justice (fair allocation of medical resources).²¹ When ethical problems arise in the care of premature infants, there is usually conflict between two or more of these principles. How best to address the conflict between these abstract principles while dealing with the concrete problems of premature infants has been debated for years. The eight major questions discussed below are of fundamental importance in addressing the ethical questions involved in using intensive care for patients of any age.

What Outcomes Are Too Dismal to Justify Aggressive Care?

Without neonatal intensive care, virtually all extremely ill, malformed, or immature infants die within a few hours, if not a few minutes, after birth. Yet, neonatal intensive care can make the outcome worse—death occurring after days or weeks of suffering. Moreover, some, though not many, survivors have severe lifelong handicaps considered to be worse than death by many parents.²² Thus, neonatal intensive care is not necessarily beneficial or justified simply because it affords a modest chance of survival.

Most religious leaders, ethicists, physicians, and members of the lay public agree that a variety of factors beside the risk of dying should be considered in deciding how to treat marginally viable patients. Contrary to the Baby Doe regulations, these factors include pain and suffering, future quality of life, and sometimes the cost of care.^4,23,24

Should the Effect on the Family Be Considered?

The effect of neonatal treatment decisions on the family are considered to be irrelevant under the Baby Doe regulations; however, this view has not gone unchallenged.^25-27 Strong asserts that the principle of nonmaleficence should be applied to the effect of treatment decisions on the family as well as the infant. "Just as the interests of the infant limit parental authority, so the interests of the family limit what can be required of the family for the sake of the infant."²⁶ Silverman argues for balancing the interests of the infant, the family, and society, noting that "parents of a badly damaged baby often resent the implied demand that their family is required to pass a `sacrifice test' to satisfy the moral expectations of those who do not have to live, day by day, with the consequences of diffuse idealism. It is easy, some parents say, to demand prolongation of each and every new life that requires none of one's own . . . resources to maintain that life later."²⁵

Will Aggressive Care Improve or Worsen the Outcome?

Unfortunately, the effect of intensive care on the outcome of marginally viable premature infants is not easily predicted. This problem results in part from uncertainty about gestational age. A one- or two-week error can change whether neonatal intensive care is considered to be justified. Yet, the error is often several weeks or more whether gestational age is estimated from the infant's physical examination or the mother's menstrual history.^28,29 Many mothers of premature infants have little or no prenatal care, irregular menstrual cycles, or an uncertain menstrual history which makes the determination of an exact gestational age difficult.³⁰

The effect of neonatal intensive care on outcome is also not well predicted from birth weight. Outcome is best predicted using multiple factors—birth weight, gestational age, gender, and condition at birth. However, even when all of these factors were used, it was impossible to predict with great accuracy whether infants whose birth weights were less than 800 grams (1 pound, 11 ounces) would survive or die when given neonatal intensive care.⁵

Severe handicap is also difficult to predict. The use of sonography has been a major advance in identifying brain damage. Yet, such damage is often not detectable on sonograms until after decisions to administer, forego, or withdraw mechanical ventilation must be made. Moreover, the relation of clinical findings in the neonatal period to long-term outcome is often not clear in the results from the available follow-up studies.

The Need for Rigorous Follow-Up Studies

It is difficult to assess the relationship between neonatal events and longer-term outcomes because it is difficult to conduct valid follow-up studies. This problem results, in part, because follow-up evaluation and home care have been woefully underfunded. Informed ethical decision making requires an understanding of all important consequences to the infant and family which might result from decisions to administer or forego neonatal intensive care. Yet, few follow-up studies have included families of infants who died, and no follow-up study has related the long-term outcome of the family to the decisions made to administer or forego neonatal intensive care to the infant. This lack of information is a fundamental problem for ethical decision making in the care of newborn infants.²⁵

What Care Is Mandatory?

Should intensive care be routinely administered because outcome cannot be well predicted? Some believe that intensive care is obligatory unless it has been shown to be futile. However, futility is a meaningful and unambiguous term only when quantitatively defined.^31,32 As recommended by one group, intensive care would be considered futile only when the likelihood of treatment failure had been shown to be at least 97%.³¹ This approach appears to be plausible. Yet, at least for neonatal intensive care, futility is virtually impossible to demonstrate.

Futility—An Unattainable and Unsatisfactory Criterion

Hospitals vary in the outcomes that can be achieved. Thus, a neonatologist cannot use results from other hospitals to demonstrate futility in his or her own hospital. To show that neonatal intensive care is futile for preventing death among infants with a birth weight of less than, say, 500 grams (1 pound, 2 ounces), it would be necessary to administer neonatal intensive care to at least 100 such infants.³¹ Moreover, because the birth of these very tiny infants is so rare, it would require five or more years in almost all U.S. hospitals to complete such a study. Many of the parents might not agree to neonatal intensive care, and serious ethical objections might be raised by any of a variety of people. Moreover, the neonatologist could not be confident that the likelihood of death was at least 97% unless all 100 of the infants had such an outcome.³¹ Even then, such certainty might not be warranted because of ongoing improvements in obstetric or neonatal care.

If even one infant survived, neonatal intensive care would be considered to be mandatory under the above approach. The suffering of all the infants who died (as well as the costs of their care) would be ignored. Use of a more stringent requirement for futility (for example, demonstration that the likelihood of death is at least 99%) would be even more problematic. Use of a less stringent definition (for example, a 95% likelihood of death or severe handicap) would still pose major problems and mandate care that might produce much more harm than good.

Evidence-Based Ethics and Medical Reasonableness

Under the usual rules of evidence in medicine, the only mandatory interventions are those for which there is credible evidence that the benefits outweigh the hazards and burdens.³³ Evidence-based decision making^34,35 allows consideration of the level of care that is reasonable, based on the quality of evidence available; the identified benefits, hazards, and costs of treatment; and as is crucial in a pluralistic society, the values and preferences of the patient or surrogate.³⁶ Thus, medical reasonableness is a more broadly acceptable criterion than is futility for deciding whether to forego or administer neonatal intensive care.

What Care Is Optional? What Care Is Investigational?

The concept that neonatal intensive care may be an ethically optional therapy for some infants has been widely discussed. Little attention has been focused on whether neonatal intensive care in some circumstances should be considered an investigational therapy, a therapy for which the value is sufficiently uncertain that there should be special care to obtain consent and to evaluate its effects. In an evidence-based approach to ethical decision making, neonatal intensive care should be considered to be an investigational therapy for infants for whom its benefits are highly uncertain. Whether a traditional institutional review board or some other body should oversee the use of neonatal intensive care as an investigational therapy should be carefully considered.

For both optional and investigational use, the neonatologist and the parents should jointly decide whether to administer neonatal intensive care. However, when neonatal intensive care is used as investigational therapy, parents should be clearly informed that neonatal intensive care has dubious value and should explicitly consent to its use, and to the extent feasible, it should be provided—as for organ transplantation^37,38—only in carefully selected units. In the author's view, these units should have (1) exemplary neonatal outcomes documented for high-risk infants born in their center, (2) carefully standardized follow-up evaluations through at least 18 months of age with no more than a 20% loss to follow-up at that age among infants for whom intensive care was optional or investigational, and (3) sufficient resources to provide investigative neonatal intensive care without compromising outcome for infants for whom neonatal intensive care is mandatory or optional. Success in publishing clinical research in peer-reviewed journals would also be highly desirable.

What Cost Is Too High?

Unfortunately, the costs as well as the benefits must be considered in the use of neonatal intensive care. While some people find it abhorrent to consider these costs, expenditures to preserve life are limited throughout our society. For example, we limit what we spend to purchase safer automobiles, build safer roads, and enforce traffic safety laws. (Indeed, highway speed limits were increased despite the predictable increase in traffic deaths as well as costs.) In proclaiming the importance of medical care, physicians sometimes lose sight of the opportunity costs of medicine—the lost opportunities to devote the same resources to improve education, employment, housing, or other social goods. Silverman notes, "It is easy to see how senseless it would be to send a fully staffed neonatal intensive care unit to rescue very small or severely malformed neonates to an impoverished country like Ethiopia. But we also need to see the folly of ignoring the disparity between unlimited medical life-prolongation in the days and weeks immediately after birth and, later, the total lack of supporting care of vulnerable, often unwanted infants living in unsanitary housing amidst the social chaos of violence and crime in inner city slums."³⁹

Assessing Medical Costs Relative to Their Value

To aid in deciding how to use limited health care resources, costs should be expressed relative to some measure of value.⁴⁰ The cost-effectiveness of neonatal intensive care can be expressed as the cost per extra survivor who would have died without neonatal intensive care or as the cost per additional life-year gained by survivors who would have died without neonatal intensive care. Assessments of the cost utility of neonatal intensive care can be expressed as the cost per quality-adjusted life-year (QALY) gained as a result of neonatal intensive care. In such assessments, the number of life-years gained from intensive care is reduced according to the number of handicapped survivors and the severity of the handicaps in those survivors.⁴⁰

If two medical programs result in the same expenses and the same number of life-years gained, the program that results in the fewest handicapped survivors will have the lowest cost per quality-adjusted life-year gained. Because a goal of medicine is to maximize the quality as well as the length of life, this program could be considered to have greater utility for the cost. The use of quality-adjusted life-years for this purpose is similar to that of "well-years" or of "disability-adjusted life-years" (DALYs), a measure recently hailed as a major advance in assessing the global burden of different diseases.⁴¹

However, the use of quality-adjusted life-years is not without controversy or problems. Such use has been criticized for discriminating unfairly against the disabled. However, use of a measure like quality-adjusted life-years which favors therapies that prevent handicap would seem to have no inherent conflict with concern for the welfare of handicapped persons. However, it is difficult to know how to adjust appropriately for disability and disease, in part because quality of life in the presence of handicaps and chronic illnesses may be rated higher by those affected than by other persons.⁴²

It is also not clear that a life-year should be valued the same for all persons, for example, the old versus the young; persons who are responsible for only themselves versus pregnant women or persons responsible for young children, elderly parents, or a disabled spouse.

In defining priorities for Medicaid expenditures, an attempt was made in Oregon to use a measure comparable to quality-adjusted life-years to rank medical and surgical therapies. This ranking was abandoned, partly because of sharp criticism of inconsistent or irrational results.⁴³ Such results were at least partly due to inadequate information to properly rank different therapies.

In the revised and less controversial Oregon ranking, clear priority was given to the care of pregnant women and children. Seventeen categories of care were ranked according to their importance. The rankings were defined by using scientific data, political importance, and other unspecified criteria. Within categories, therapies were ranked according to perceived effectiveness. Maternity care and care of the newborn—including low birth weight infants—were ranked second only to the category of life-saving therapies which result in full recovery. However, it remains to be determined whether our society as a whole would choose to accord a higher priority for newborn infants and pregnant women than for other citizens and, if so, whether this would be on the basis of an equal or higher value per life-year or quality-adjusted life-year gained.

The Ratio of Cost to Value: Better for Intensive Care of Premature Infants Than for Adults?

A fundamental problem in comparing the costs of care for different interventions or different patients in the United States is that hospital charges are designed to maximize revenues and are an unreliable indicator of true costs. The most comprehensive assessment of the short- and long-term costs of neonatal intensive care comes from a Canadian study.⁴⁴ This information is the basis for official estimates of the cost of neonatal intensive care in the United States.⁴⁵

For at least the great majority of premature infants, this study suggests that the cost of neonatal intensive care is not excessive relative to its utility. Indeed, the estimated cost per quality-adjusted life-year gained for infants with birth weights of 1,000 to 1,500 grams (2 pounds, 3 ounces to 3 pounds, 5 ounces) was $4,500, less than that for treating severe or even mild hypertension in 40-year-old males ($9,400 to $19,100).⁴⁶ Cost increased with decreasing birth weight. Yet, the estimated cost per quality-adjusted life-year gained for infants with birth weights of 500 to 1,000 grams (1 pound, 2 ounces to 2 pounds, 3 ounces) was $31,800, still less than the estimates for coronary artery bypass surgery ($36,300) or renal dialysis ($47,100 to $54,000).

With advances in neonatal intensive care since this study, the cost utility is likely to have improved for infants whose birth weight exceeds 800 to 900 grams (1 pound, 11 ounces to 2 pounds, 0 ounces). However, cost per quality-adjusted life-year gained for smaller infants is uncertain and may have increased substantially. New comprehensive economic analyses are needed. Nevertheless, analyses of hospital stays performed using data from the Neonatal Research Network⁵ suggest that the cost per life-year gained or per quality-adjusted life-year gained from intensive care may still be less—perhaps much less—for most infants with birth weights of 501 to 800 grams (1 pound, 2 ounces to l pound, 11 ounces) than for many adults. Hospital charges for intensive care of adults expressed per life-year gained at home were recently reported to be $181,308 for the highest-risk elderly patients,⁴⁷ $95,142 for adults with solid tumors, and $449,544 for adults with hematologic cancers.⁴⁸ Total costs may be even higher for resuscitative efforts of adults outside the hospital who do not initially respond⁴⁹ and such programs as cholesterol screening in young adults.⁵⁰

Is Rationing of Care at the Bedside an Unavoidable Reality?

Findings like those noted above indicate the importance of better assessing the true costs of expensive medical therapies and defining societal priorities for expensive medical care for patients of all ages. It is also important that physicians use the available resources wisely. Two reports have noted a worsening outcome among infants with birth weights greater than 1,000 grams (2 pounds, 3 ounces) as use of neonatal intensive care for infants with birth weights between 500 and 1,000 grams (1 pound, 2 ounces and 2 pounds, 3 ounces) increased.^51,52 Even in the United States, inadequate resources for neonatal intensive care may now be or may soon become a serious unrecognized problem.^52,53 Stringent measures to reduce costs are now being implemented in many neonatal units, and pressure to ration is intensifying throughout the health care system.^52,53 The often-repeated dogma that rationing should not occur at the bedside ignores the clinical realities of busy intensive care units with limited beds, staff, or support services.^54-56 To minimize preventable deaths and handicaps, resources must be used in the most effective manner feasible.

Who Should Decide?

A conceptual analysis of the role that society, parents, and physicians should play in making treatment decisions for neonatal intensive care is shown in Table 1 (based on the work of Fleck, Lorenz, and Tyson⁵⁷). In principle, the designation of neonatal intensive care as unreasonable, investigational, optional, or mandatory should be a societal judgment based on a painstaking analysis of the best available evidence concerning the benefits and the cost. Describing the steps that would constitute an appropriate process to achieve a societal judgment is beyond the scope of this article; however, the process would necessarily involve a broad variety of persons, including neonatologists, parents, economists, policy analysts, and members of the lay public.

When Parents Should Not Be Involved

Neonatal intensive care would be mandatory (last category in Table 1) when the cost is judged to be acceptable and the evidence of benefit is strong, as for a 1,500-gram infant born at 31 weeks gestational age with respiratory distress syndrome in the United States. In this situation, the parents do not participate in decision making; they have no right to demand that neonatal intensive care be withheld, and the physician is obligated to administer neonatal intensive care. (If the resources within the hospital are not adequate, the infant should be transferred to an appropriate facility whenever feasible.) Neonatal intensive care would be unreasonable (first category in Table 1) when there is little or no evidence that it would be beneficial and when the expected value of that care would not justify the costs. Barring major advances in neonatal intensive care, care of this type would be unreasonable for an infant weighing 250 grams (0 pounds, 9 ounces) born at 20 weeks gestational age with a slow heart rate. In this situation,⁵⁸ parents would not participate in decision making and would have no right to demand neonatal intensive care, and physicians ought not to administer neonatal intensive care.

When Parents Should Be Involved

As indicated in Table 1, parents have a right to be involved in decision making whenever neonatal intensive care is optional or investigational. The neonatologist needs the freedom to initiate neonatal intensive care if the parents need more time to participate effectively in treatment decisions. However, well-prepared parents may be able to participate in decision making at or before birth. In an important article, Harrison²⁷ has urged the development and evaluation of new strategies to improve decision making, including the use of advance directives written during pregnancy to advise physicians of parental preferences if the appropriateness of resuscitation or aggressive treatment at birth is questioned.

Classification of Infants

How should infants be classified into the different categories? Any specific clinical criteria (birth weight, gestational age, and so on) would have to be periodically revised as neonatal care and prognosis changed. However, this problem might be avoided by developing criteria based on specific levels of risk for the occurrence of either death or severe neonatal morbidity. (Severe neonatal morbidity among infants with birth weights under 1,500 grams [3 pounds, 5 ounces] might be defined as Grade III or IV intracranial hemorrhage; cystic white matter disease; necrotizing enterocolitis requiring surgery; and a need for oxygen therapy at 36 weeks postmenstrual age. These disorders are associated with a prolonged need for intensive care and painful procedures or complications.)^5,27,59

Neonatologists are likely to differ considerably in the risk levels that they would select to classify individual babies.⁶⁰ As noted above, a societal judgment needs to be made. However, for the sake of discussion, the author's view is that the initiation of neonatal intensive care might be considered mandatory when the risk of death or severe neonatal morbidity (estimated as described below) is 50% or less, optional when the risk is 51% to 75%, and investigational when the risk is 76% to 95%. Neonatal intensive care might be considered unreasonable when the risk is greater than 95% (with the possible exception of infants who are born in centers that are qualified to give investigational neonatal intensive care and whose attending neonatologist and parents desire to initiate such therapy). These risk levels refer to those for infants born in centers with neonatal intensive care units.

Large numbers of infants are required to develop precise estimates of risk for a given birth weight, gestational age, and sex. However, outcome data are published periodically from the Neonatal Research Network⁶¹ and the Vermont-Oxford Trials Network.⁶² (Neonatologists working in large neonatal units could use data for infants in their unit during the previous three to five years.) If recent Network data were used, the initiation of neonatal intensive care might be designated as investigational for small-for-gestational-age females with birth weights below 500 grams (1 pound, 2 ounces), and appropriate for both small-for-gestational-age males with birth weights below 700 grams (1 pound, 8 1/2 ounces), and all other infants with birth weights below 600 grams (1 pound, 5 ounces). Similarly, initiation of neonatal intensive care could be considered as mandatory for small-for-gestational-age females with birth weights above 700 grams and appropriate for both small-for-gestational-age males with birth weights above 900 grams and all other infants with birth weights above 800 grams. After neonatal intensive care is initiated, the infant's risk would change depending on his condition, response to therapy, and age. Thus, separate criteria would need to be developed for decisions to withdraw neonatal intensive care at various ages after it has been started.⁶³

The Uncertain Role of Hospital Bioethics Committees

The number and influence of hospital bioethics committees appears to have increased considerably after the Child Abuse and Treatment Act. These committees may reduce the likelihood of hasty or unreasonable decisions. They may also be very useful as advisory bodies to address ethical questions raised by clinical staff or disagreement between parents and physicians. However, these committees pose important potential problems, including the possibility that the members might not know enough about the patient or the medical problem, or might be biased by institutional concerns.^64,65 These committees will require further evaluation before their proper role is clear. There seems to be no information to indicate that they should routinely review or participate in treatment decisions for high-risk newborns.

Seeking Consent in the Early Neonatal Period

The goal of informing and involving parents appropriately in treatment decisions is of indisputable importance. Yet, this goal is often difficult to achieve, particularly during labor or shortly after delivery when decisions to forego or initiate neonatal intensive care are often made. If the baby was transported to a neonatal intensive care unit, the parents may be unavailable because they live in a distant community. Moreover, fathers are often not immediately available. About half of the mothers of infants weighing less than 1,500 grams (3 pounds, 5 ounces) at birth have undergone a cesarean section. Many of these mothers have major illnesses or complications. Most receive potent medications. Virtually all experience considerable physical and emotional distress. Moreover, as discussed below, "adequately" or "fully" informed consent may be impossible even if the parents—who are often young or poorly educated—are told everything known by the physician.⁶⁶