AMA Journal of Ethics^®

Case

The NICU team was called to the trauma bay of the emergency room where obstetricians had just performed a perimortem cesarean section on a pregnant woman who died on arrival to the hospital after a vehicle crash. The NICU team successfully resuscitated the infant after several attempts. The neonatologist evaluated the infant, a male, and estimated him to be 26 weeks' gestation. The infant's condition deteriorated over the next few days, and he required intubation, ventilatory support, and IV medications to maintain his blood pressure. No family members came forward to claim the extremely premature infant.

The infant developed large bilateral intracranial hemorrhages and had daily seizures. He became septic and hemodynamically unstable. On several occasions the NICU team considered withdrawing care, but ultimately the decision was made to continue. Over months he gradually recovered, but it became clear that the baby was severely neurologically impaired; his body took on rigid postures and he had recurrent seizures. He was unable to suck and relied on tube feeds for nutrition. He barely responded to external stimuli. When he was finally discharged as a ward of the state to a long-term care facility for neurologically impaired children, the cost of the infant's care exceeded several hundred thousand dollars.

The executive board of the hospital convened a meeting with the ethics committee and the NICU staff to develop a policy for such cases. A member of the executive board stated that the hospital's budget was too tight to provide hundreds of thousands of dollars of care in medically futile cases when they could help hundreds of other infants with the same funds. The board argued that the social and financial costs of prolonged NICU stays for infants destined to be wards of the state surpassed the care capabilities of the hospital. One particularly blunt executive stated, "No one wanted that baby to live; the mother is gone, there is no family, the baby himself kept trying to die and you all wouldn't let him, and now he's in a lonely institution his whole life as a permanent vegetable, all at taxpayers' cost of a half-a-million dollars—who are we helping here, guys?"

A resident physician proposed a rule-based approach to resuscitation decisions: "Why do we struggle with this same decision with every delivery of a 500-gram infant? We need to set up guidelines that will take the difficulty out of these tough situations. Let's make a standard policy of when to resuscitate and when to withdraw care—we could take some of the agony out of these situations." A more senior physician disagreed: "The practice of medicine is not a cookbook. If a set of fixed rules governed all of our resuscitation decisions then we may as well have the accountants take over. A universal policy would destroy the art and humanity of medicine."

Commentary 2

Before addressing whether it is possible to practice rule-based ethics in neonatal intensive care units, several suppositions made in the case must be examined.

The first supposition is that a rule is needed because there is no process in place to make management decisions under circumstances and a rule would obviate the need for an onerous process. But there is such a process in place: the appointment by the court of a guardian ad litem. How this process is initiated, who or what agency may function as a guardian ad litem, and how decisions are made within this structure vary among jurisdictions. While use of a guardian ad litem is admittedly not as satisfactory as having parent decision makers, the guardian represents the best interests of the infant distinct from the interests of the health care team, hospital, or state. This process grants primacy to the best interest of the patient in health care decisions for those who have never had capacity to do so for themselves; it is a principle espoused by the President's Commission for the Study of Ethical Problems in Medicine and Biomedical and Behavioral Sciences [1]. Rule-based ethics may assist this process, but cannot substitute for it.

The second supposition is that the benefit-burden ratio of intensive care for the infant in this case was low enough to have ethically justified withholding care. Since intensive care was clearly effective in promoting survival of this infant, the questions become: is quality of life ever a sufficient criterion to justify withholding or withdrawing intensive care and, if it is, how severely diminished must that quality of life be? There is an almost universal belief that human life has intrinsic value and ought to be preserved. One extreme of this principle holds that biological human life has intrinsic value and ought to be preserved without regard to the quality of that biological life. Another view holds that only life of some minimum quality to the person ought to be preserved.

Rhoden argues persuasively that "quality of life judgments are appropriate, necessary, and in fact inevitable" in dealing with imperiled newborns [2]. There is, however, no consensus on what constitutes the minimum quality of life that ought to be preserved. The President's Commission concluded:

that a very restrictive standard is appropriate…permanent handicaps justify a decision not to provide life-sustaining treatment only when they are so severe that continued existence would not be a net benefit to the infant…net benefit is absent only if the burdens imposed on the patient by the disability or its treatment would lead a competent decision maker to choose to forego the treatment [3].

The Child Abuse Amendments of 1984 are more specific:

withholding treatment from an infant is permissible only if: (a) the infant is chronically and irreversibly comatose; (b) the provision of such treatment would (i) merely prolong dying, (ii) not be effective in ameliorating or correcting all the infant's life-threatening conditions, or (iii) otherwise be futile in terms of survival of the infant; or (c) the provision of such treatment would be virtually futile in terms of survival of the infant and the treatment itself under such circumstances would be inhumane [4].

Robertson argues that intensive care must be provided only if (in addition to the exceptions specified in the Child Abuse Amendments of 1984) the child possesses or has the potential to possess some threshold level of cognitive ability beyond mere consciousness, specifically "the capacity for symbolic interaction and communication" [5]. Rhoden proposed preliminary guidelines that aggressive treatment is not mandatory if an infant: (1) is in the process of dying; (2) will never be conscious; (3) will suffer unremitting pain; (4) can live only with major, highly restrictive technology which is intended to be temporary (e.g., artificial ventilation); (5) cannot live past infancy (i.e., a few years); or (6) lacks potential for human interaction as a result of profound retardation [6].

Whatever the criteria for withholding intensive care, it is usually presupposed that there is parental concurrence with the decision. Any outcome that would justify the withdrawal of intensive care over parental wishes would certainly require a much lower benefit-to-burden ratio. This suggests that the minimum quality of life that justifies withholding intensive care in the absence of a caring parent who represents the interests of the infant might be poorer than when a parent is available.

Even if agreement could be reached on what minimum quality of life obligates the provision of intensive care, how likely must it be that that minimum quality of life will not be achieved? This is critical in cases like this in which the prognosis can only be estimated when time-management decisions must be made. Is a 5 percent, 10 percent, or 15 percent chance of achieving the minimally acceptable quality of life sufficient grounds to forgo life-sustaining treatment? The likelihood of a major disability, much less the lack of capacity for symbolic interaction and communication, cannot be accurately predicted for individual infants during the NICU course with the data currently available. For example, the positive predictive value of cystic periventricular leukomalacia for major disability has been reported to be 71 percent and 83 percent [7, 8]. In other words, 1 in every 4 to 6 children with cystic periventricular leukomalacia will not have a major disability. And certainly not all major disabilities preclude a quality of life sufficient to justify life-sustaining treatment.

The problems are how to define the best interests of an infant with an ambiguous future and how much to weigh the opinions of the key players—the parents and health care professionals.In this case, let us accept that the outcome of this infant at discharge would ethically have justified forgoing life-sustaining treatment. The issue then is whether this outcome could have been predicted in the infant's course with sufficient reliability to justify withdrawal when withdrawal of life-sustaining treatment was a realistic option. With the information provided we cannot know.

The third supposition made is that the provision of intensive care to this infant is an inappropriate use of limited health care resources. This is an oblique reference to health care rationing—most broadly defined as implicitly or explicitly allowing patients to go without health care services that are of some benefit to them because of cost [9, 10]. The United States has not shown much of an appetite for the explicit rationing of health care resources. If this is to be done, however, it must be at a higher level than that of an individual institution. Distributive justice requires that finite heath care resources be fairly and equitably allocated. These allocation decisions must be made at the community level if they are to be reflective of the range of values within the community and be applicable across the community. Rationing at any level, while unavoidable, is fraught with moral problems that some argue are unavoidable [11]. The authors of the most prominent example of an attempt to ration health care, the Oregon prioritization plan, admitted that there was no perfectly objective, uniquely rational, or indisputably fair way of rationing [12]. It was a "process question to be resolved though open democratic dialogue whose outcome was shaped by both social value judgments and medical information" [13].

The case at hand does not refer to the universal moral rules that underpin utilitarian or Kantian ethical theories, but rather rules for more specific circumstances that are based on one or another ethical theory. We cannot do without some ethical rules; they are essential to understanding what is morally right and wrong. In the best circumstances, they represent a summary of cumulative moral experience and wisdom, but they cannot obviate the need for moral deliberation. To be useful, rules must be general enough to be applicable to a range of circumstances. There cannot be enough rules to cover every ethical dilemma. Moreover, rules may conflict with one another. Thus, rules must be applied to specific circumstances, and application to specific circumstances requires moral deliberation. Focusing on rules emphasizes what we ought to do, rather than the reasons for what we ought to do.

Annas has suggested that adherence to reasonable process for making management decisions for extremely premature infants may be the best we can do "because clear rules seem to be impossible to formulate in this arena" [14]. Today, process includes candid conversation among parents, physicians, and other health care professionals; consideration of all the relevant facts and interests; and, in extremely difficult or refractory cases, consultation with an ethicist or institutional ethics committee. In the case at hand, with no parent available, court involvement may be required as well.