In this article, I distinguish between 4 models for thinking about how to balance the interests of parents, families, and a sick child: (1) the oxygen mask model; (2) the wide interests model; (3) the family interests model; and (4) the direct model. The oxygen mask model – which takes its name from flight attendants' directives to parents to put on their own oxygen mask before putting on their child's – says that parents should consider their own interests only insofar as doing so is, ultimately, good for the sick child. The wide interests model suggests that in doing well by my child I am at the very same time doing well by myself. My interests can, and plausibly do, encompass the interests of others; they are, to that extent, wide. There is, then, no sharp separation between the interests of the sick child and the interests of other family members. In the family interests model, families themselves are seen as having interests that are neither identical to the sum, nor a simple function, of the interests of individual family members. The family has goals, values, and aspirations that are essentially corporate rather than individual. According to this model, these family interests can explain why sacrifices can sometimes be demanded of some family members for the sake of others in a medical setting. Finally, the direct model takes a simpler view of family members' interests; it claims that these interests matter simply on their own and should be taken into account in making treatment decisions for a sick child. This model openly considers the competing interests that parents and other family members often have when caring for a sick child, and advocates for weighing those interests when making decisions for and about the sick child. While there is room for all four models at the bedside, I argue that the direct model should be highlighted in clinical decision-making.
Consider the following case:
Ten-year-old Marcus has a medulloblastoma. The physicians offer Marcus' parents a choice: they can either give Marcus routine chemotherapy and radiotherapy at their local hospital or they can send him for chemotherapy plus proton beam radiotherapy at a hospital in a major city >500 miles away from where Marcus, his 3 siblings, and his parents live. The physicians say that both treatments are equally effective, but the one farther away has a chance of reducing neurotoxicity and its associated neurocognitive impairments as Marcus grows up.
We might wonder what the parents, as decision-makers, can reasonably decide to do here. To answer that question, we would need to know more details about the case. But without knowing those details, we can still ask: what issues should the parents consider when making their choice?
There is, of course, complete agreement that the interests of the ill child matter. Perhaps more surprising, however, is that there is also widespread agreement that those interests are not all that matter. Most clinicians,1,2 and many bioethicists,3–6 think that the interests of other family members also matter when making a decision about what to do in a case such as this. However, the way in which they matter is unclear. The purpose of the present article is to distinguish among 4 models of the proper place of family members’ interests in clinical decision-making. The 4 models are not mutually exclusive. Instead, all 4 capture important facts about the nature of family interests, and all 4 have a place in clinical decision-making.
Nonetheless, one model (what I will call the “direct model”) should be highlighted in clinical decision-making. The conception of family interests that this model offers is admittedly obvious once articulated. Even so, it is easily overlooked; it is, as it were, hiding in plain sight. It deserves to be seen.
The Oxygen Mask Model
On airplanes, flight attendants tell passengers that in the event of an emergency, parents must place the oxygen mask over their own faces before putting on their child's mask. The flight attendants are not encouraging parents to be selfish. Rather, they are telling parents that, despite initial appearances, the parents will do better by their child if they attend to their own need for an oxygen mask first.
Consider this reasoning in light of our case. Assuming that we think that the interests of Marcus’ parents and their other children matter, why do we think this? Consider what might happen to the patient if, instead of staying local, the parents decide to travel for treatment. One, or perhaps both, parents might need to scale back their working hours. Indeed, perhaps one or both might have to give up their job altogether. This scenario, we might imagine, will cause serious financial strain on the family and ultimately result in Marcus getting worse care than he would if the family remains financially healthy. Or, we might point to the obvious truth that taking care of someone, sick or not, is difficult and that to do it well, caregivers need to sometimes attend to their own needs first to perform suitably as caregivers.
Taking inspiration from what we are told on airplanes, I will call this the oxygen mask model of family interests. According to this model, family interests matter in medical decision-making because it is ultimately good for the patient if the decision-makers consider the interests of themselves and their family members. In other words, things will go better for the patient in the long run if the interests of other family members are taken into account when making important treatment decisions. The oxygen mask model tells us that the interests of other family members are instrumentally important: they matter because taking them seriously will ultimately be good for the patient.
Can this model be used to account for the importance of considering the interests of other children and not just the interests of the parents? It can, once we see that the health of the family, considered as a unit, is important to the flourishing of its members. This idea is argued for forcefully by Hilde Lindemann in her contribution to this Pediatrics supplement.7 As she points out, the family is, for most children, the place where they develop into agents, where they form a notion of what matters in life, and where they acquire the emotional and material support they need to become functioning adults. All of these factors are central to human flourishing. If most children only develop well in the context of a reasonably well-functioning family, then it is in the interest of the ill child that the interests of other family members, including siblings, be taken into account so that the family as a unit can remain healthy.
Therefore, the oxygen mask model states that the interests of other family members matter because the well-being of those other family members is important to making things go well for the patient. This model encourages decision-makers to take the long, or at least the longer, view in determining what is in the best interests of the patient. A focus on the short term might make a parent think that putting his or her child's oxygen mask on first is in the best interests of the child. But a longer term approach reveals that this is not correct. Likewise, a focus on the ill child's situation right now or next week, for example, might lead the parents to a view about what is best for their child that is different than if they were to take a longer view, one that takes into account the effect of various treatment options on the family.
The Wide Interests Model
The second model of family interests is nicely summarized by Lainie Friedman Ross:
Within intimate families, one member's well-being is an integral part of the other member's well-being. Intimates take on one another's goals, even as they retain their independent goals and identity. To the extent that each family member incorporates the other members and the other members' ends into his own ends, the difference between altruism and egoism collapses. Other-regarding activities become self-regarding activities.8
The idea expressed here captures something important regarding intimate relationships, not just intimate families. We can feel its force most clearly, however, by focusing on how loving parents conceive of their interests in relation to their child's interests. One of the central interests of loving parents is that their child flourishes. For a loving parent, part of what makes his or her life go well is that the child's life goes well. In other words, part of the well-being of a loving parent is composed, or consists of, his or her child's well-being. Therefore, when the child does well, it reasonably follows that the parent will also do well.
In this way, “the difference between altruism and egoism collapses. Other-regarding activities become self-regarding activities”: since doing things that help my child flourish also helps me flourish, there's a sense in which actions done for the sake of my child are also done for my own sake. I don't do them to benefit myself (that would be selfish). I do them to benefit my child. But in doing so, I benefit myself since part of my interests consist of my child doing well.9–10
What this means is that actions that might seem to go against my interests really do not once we have in view an appropriately wide conception of my interests. What might seem like a sacrifice for the sake of my child, for example, really is not: in doing well by my child, I am at the very same time doing well by myself. My interests can, and plausibly do, encompass the interests of others. They are, to that extent, “wide.” I thus call this option the “wide interests model.”
Interestingly, the wide interests model has been invoked by the courts when making decisions about whether a child must donate much-needed marrow to an ailing sibling. In a case that June Carbone discusses in greater detail in her article in this Pediatrics supplement,11 a father of a 12-year-boy who needed a bone marrow transplant sought to have the 3.5-year-old twins he had fathered with the now-estranged mother tested for a match. The mother refused on the grounds that donating marrow to their half-brother was not in the twins’ best interests. The court agreed and offered the following as part of the reasoning behind the decision:
There must be an existing, close relationship between the donor and recipient. The evidence clearly shows that there is no physical benefit to a donor child. If there is any benefit to a child who donates bone marrow to a sibling it will be a psychological benefit. According to the evidence, the psychological benefit is not simply one of personal, individual altruism in an abstract theoretical sense, although that may be a factor.
The psychological benefit is grounded firmly in the fact that the donor and recipient are known to each other as family. Only where there is an existing relationship between a healthy child and his or her ill sister or brother may a psychological benefit to the child from donating bone marrow to a sibling realistically be found to exist. The evidence establishes that it is the existing sibling relationship, as well as the potential for a continuing sibling relationship, which forms the context in which it may be determined that it will be in the best interests of the child to undergo a bone marrow harvesting procedure for a sibling.12
One way to understand what the court is stating here is that because the twins lack a psychological connection to their half-brother, the interests of the latter child do not mesh with the interests of the twins in the way they do in a close-knit family. Because there is no meaningful connection between the twins and their sibling, the difference between altruism and egoism does not collapse in this case. Donating marrow (indeed, even testing to see if donation is a possibility) would straightforwardly cut against the interests of the twins.
The corollary to this line of thought is that if the twins and the sibling were sufficiently tight-knit, it might actually then be in the interests of the twins to donate. We could think this scenario is true for purely instrumental reasons: the twins will benefit from having their brother around in the future; they will not suffer the sadness that would accompany their brother dying, for example. But it might also be true if we keep in mind the wide interests model. For example, we might say that if they were part of a tight-knit family, the brother’s good would be part of the twins’ good. Therefore, insofar as the brother does well, so do the twins. In this re-imagined version of the case, “the difference between altruism and egoism” indeed collapses.
How might the wide interests model help us incorporate family interests in the case we started this article with? It is not entirely clear. On the one hand, we might say that Marcus’ parents’ interests are part of his interests, and it is thus not obviously against the child's interests to not seek out proton beam radiotherapy. It is more plausible, however, to deploy the wide interests model in the other direction and say that the child's interests form part of the parents’ interests and therefore doing what is best for the child is actually doing well by the parents, at least to some degree.
The Family Interests Model
Interestingly, Ross thinks that the wide interests model has, at best, limited application. Writing about sibling donation, she noted:
The main risks to the donor are anaesthesia (rare but potentially life-threatening) and pain (which is temporary). The main benefits to the donor are the benefits he gets from his sibling's survival and any benefit he gets from his altruistic behaviour and its related praise. However, the potential donor may dislike his sibling and deny that he will derive any psychological benefit. His parents may argue that he will benefit in the long term. They may be right; but it is also possible that the donor never benefits.13
In response to this line of thought, a proponent of the wide interests model could say that whether the potential donor sibling believes that his interests are deeply, and constitutively, intertwined with his sibling’s is not relevant. What matters is that, in an intimate family at least, they are intertwined (whether he believes it or not). Therefore, it really is in his interest to donate, regardless of whether he believes it. But this gloss on the wide interests model leads to the unattractive conclusion that the donor sibling is wildly off base about his interests. What's more, it suggests that he could lead a life that, while deeply unsatisfying to him, is nonetheless good for him.
A more plausible construal of the wide interests model maintains that the donor sibling's interests are partly constituted by the interests of his siblings only if he takes them to be.3 If that is correct, then Ross' example hits its mark, for here we have a case in which it is not in the donor sibling's interests to donate because he does not consider his sibling’s interests to be part of his. However, it is surely permissible for the parents to insist that the donor sibling actually donate. How do we account for this?
According to Ross, the problem is that we have been interpreting the interests of the family members individualistically. Of course, those interests matter, but they are not all that is at stake; rather, in a close-knit family, another set of interests emerges: the family interests. In the family interests model, families are thought to have interests that, although closely connected to the interests of the individual family members, are not identical to their sum nor are they a simple function of them. More accurately, the family has goals, values, and aspirations that are essentially family goals, values, and aspirations. As Ross states, “Families can have interests that are not reducible to the interests and needs of particular members.”6 It is on the basis of these interests that parents could insist, in the aforementioned case, that the donor sibling actually donates:
From the family's perspective, it is best, all things considered, for the bone marrow transplant to occur. Even if the donor does not procure any direct personal benefit, it is in the interests of the family as a whole [where those interests are not reducible to the interests of the individual family members].13
According to this model, the justification for making a sibling donate marrow is not that it would be in the donor sibling's interests, widely construed or instrumentally. Indeed, the justification would not ultimately be in terms of the donor sibling's interests at all; rather, the justification will be in terms of the family's interests. The parents might say, “For the sake of the family, we need to do what's required to keep your brother alive. That is why you must donate.” Likewise, in the case we began with, the parents might justify choosing the standard treatment for Marcus in terms of the good of the family, and not just their own interests and the interests of their other children.
The Direct Model
The fourth model differs from the first 3 in kind. The first 3 models consider the interests of individual, non-patient family members but in an indirect or attenuated way. The oxygen mask and family interests models consider individual family member's interests in an indirect way: the interests of individual family members matter only to the extent that something else matters. For the oxygen mask model, that “something else” is the patient's interests. For the family interests model, that “something else” is the family's interests. The wide interests model does not consider the interests of individual family members indirectly. But it helps itself to an expansive conception of individual family members' interests so that, for example, donating bone marrow could be said to be part of the donor's interests because the recipient's good is part of the donor's good.
The direct model, as its name indicates, does none of this. Instead, it takes a (relatively) narrow view of family members’ interests and says, simply, that these interests matter and should be taken into account when making treatment decisions for a sick child.
Consider, again, the cost to the parents in our original case of choosing proton beam radiotherapy in a distant city: their work and personal lives will suffer in various ways. Instead of considering how such a cost will ultimately impact the interests of the patient (the oxygen mask model), or how it will impact the interests of the family (the family interests model), or how it might actually be in the interests of the parents to bear such costs because doing so is good for their child (the wide interests model), the direct model says that we should consider the impact of various decisions on the parents' interests considered simply as the parents' interests. Thus, if the mother has a strong interest in advancing in her career, that factor is relevant to deciding what to do considered simply, and ultimately, as an interest of the mother's.
In the donor case, the direct model (like the family interests model) tells us that we should seriously consider the idea that it simply might not be in the best interests of the donor to donate. But unlike with the family interests model, the direct model tells us that the competing interests that weigh against the donor's interests are not “family interests,” where those are construed as something distinct from the interrelation of the interests of individual family members. Instead, the direct model tells us that each member of a family has his or her interests (no matter how intertwined with the interests of other family members they may be) and that they should all be considered in arriving at a decision about what to do.
In the case of the donor sibling, therefore, the direct model tells us that the parents should weigh the needs of the sick child against the needs of the donor sibling. It seems clear that in cases of this kind, such a consideration will often lead reasonable parents to decide that the donor sibling must donate on the grounds that the recipient's need for marrow outweighs the donor's interest in not donating.
Importantly, in saying that the interests of family members should be directly considered, the direct model takes no stand on whether all the competing interests should be weighted equally. It is perfectly consistent with the direct model to think that some interests, perhaps those of children, should weigh more heavily (perhaps substantially so) than others, perhaps those of parents. The central contention of the direct model is just that the interests of other family members matter but not as part of some other set of interests or as part of a nonegoistic conception of interests. Rather, they matter for their own sake.
Which model should parents, or decision-makers in general, use when making clinical decisions for children that will (deeply) impact other family members? As I indicated at the start of this article, the 4 models are not mutually inconsistent (although on particular occasions they may yield different results). It is plausible to think that decision-makers should consider how their own interests (and those of other children) instrumentally relate to the good of the patient; how their own interests (and those of other children) are partly constituted by the patient's interests; how their own interests (and those of other children and the interests of the patient) relate to the family's interests; and, finally, how their own interests (and those of other children and the interests of the patient) matter independently of all these factors.
The interaction between these ways of considering the interests of family members will undoubtedly often be extremely complicated. It is highly doubtful that there is any systematic way of weighing the models against each other or adjudicating disputes among them when they point to different outcomes. This is likely to be true even if we supplement the models with a more rigorous account of how, precisely, competing interests should be adjudicated in cases such as this. The fact that in many cases there will be no clear answer about how to combine the models, however, is not a failure of the models themselves or of the idea that multiple models can simultaneously apply to a single situation. A lack of systematicity is, rather, in the nature of the subject and, as Aristotle reminded us a long time ago:
Our account would be adequate if we achieved a degree of precision appropriate to the underlying material; for precision must not be sought to the same degree in all things. [...] We must want to be content, then, when talking about things of this sort...to show what is true about them roughly and in outline, and when talking about things that are for the most part, and starting from these, to reach conclusions of the same sort. It is in the same way, then, that one must also receive each sort of account; for it is the mark of an educated person to look for precision in each kind of inquiry just to the extent that the nature of the subject allows it.14
The idea that Aristotle then argued for with respect to the whole of ethics is true of clinical ethics as well: in many cases, a proper decision requires the application of a kind of practical wisdom that cannot be reduced to rote application of precise rules.
But even if systematicity of a certain kind will necessarily elude us, I want to conclude by suggesting that the direct model should, if not take precedence over the other models, be openly deployed and accorded at least as weighty a status as the others. I offer this recommendation because there is pressure, so to speak, to downplay this model. The pressure comes from (at least) 2 sources. First, although it has come under fire in recent years,15 the idea that clinicians and decision-makers should act in the best interests of their patients is still largely taken for granted in medical ethics and medical practice. This idea runs directly counter to the direct model, which tells us it is acceptable for decision-makers to consider their own interests (and the interests of other family members) in deciding what to do, without contorting the relevance of those interests into a mold that does not take them seriously on their own terms.
The second source of the pressure to discount the direct model comes from appreciating who the decision-makers usually are in the kinds of cases we have been considering: the parents of a very ill child. Loving parents will naturally subordinate their own interests when dealing with the interests of their child, particularly if the child's needs are great. In some cases at least, they are also likely to subordinate the interests of their other children (although probably to a far lesser extent than their own interests). As I noted earlier, the direct model does not entail that such subordination, particularly when it comes to the parents' interests, is misplaced. But given that there is a natural tendency for parents to subordinate their interests (at least), foregrounding the direct model in clinical decision-making would act as a kind of counterweight to that tendency by reminding parents (or other decision-makers) that it is acceptable to directly consider their own interests and the interests of other children in arriving at a decision about the ill child.
This option would involve openly acknowledging thoughts that parents are likely already having about how various treatment options will affect them (or other family members) and encouraging them to talk about it, without having to filter those interests through the interests of the ill child, “family interests,” or their own interests widely construed. In the case I started with, for example, taking the direct model seriously would involve the parents themselves openly considering the sacrifices various treatment options would mean for them, and the medical team acknowledging the reality of such sacrifices. In some cases, such an acknowledgment might have no effect on the final decision: the interests of the ill patient might be such that none of the parents' interests (or the interests of their other children) tip the scales of the decision away from what it would be if the parents only considered the interests of their ill child. Even if the final decision is not affected by direct consideration of others’ interests, it is surely beneficial for parents to talk openly about the effect that such a decision is having on their own lives and the lives of their other children.
In other cases, openly acknowledging the interests of other family members (including those of the parents) might tip the balance toward a certain treatment or nontreatment decision. In a case in which consideration of the patient's best interests alone yields a highly indeterminate result, the fact that pursuing a particular treatment option might entail serious sacrifice on the part of the parents might tip the balance away from that particular treatment option. The case at the beginning of this article might be such an instance, although it is hard to say without knowing more about the circumstances. It is not hard, however, to imagine the case in such a way that traveling 500 miles with Marcus would involve a significant enough sacrifice on the part of Marcus’ parents and siblings that deciding not to pursue the distant treatment could be reasonable.
Ironically, however, direct consideration of nonpatient interests in these “close cases” might have exactly the opposite effect; that is, considering their own sacrifice might have the effect of making decision-makers more willing to subordinate their own interests. Everyone has had the experience of being more willing to make a sacrifice after that sacrifice has been openly acknowledged. Indeed, oftentimes, what might seem like an unwillingness to make a sacrifice is not that at all, but a kind of resentment that the sacrifice is not being recognized as such. The direct model fosters such recognition.
This article looked very different before receiving feedback from, and reading the other articles by, the participants at the Pediatric Bioethics Workshop in Kansas City in December 2013. I am grateful for all the help I received from all the participants. I am especially indebted to Dan Brudney, John Lantos, and Emily Carroll for all their help.
- Accepted June 10, 2014.
- Address correspondence to: Daniel Groll, PhD, Philosophy Department, Carleton College, 1 North College St, Northfield, MN 55057. E-mail:
FINANCIAL DISCLOSURE: Dr Groll has indicated he has no financial relationships relevant to this article to disclose.
FUNDING: Funding for this research was provided by the Claire Giannini Fund.
POTENTIAL CONFLICT OF INTEREST: Dr Groll has indicated he has no potential conflicts of interest to disclose.
- Lindemann, H, Nelson JL. The romance of the family. Hastings Cent Rep. 2008;38(4):19–21
- Schoeman F
- Ross LF
- Lindemann H
- Ross LF
- Rowe S,
- Broadie CJ
- Carbone J
- ↵Curran v. Bosze, 141 Ill.2d 473, 566 N.E.2d 1319 (Ill.,1990)
- Ross LF
- Rowe S,
- Broadie CJ
- Rhodes R
- Copyright © 2014 by the American Academy of Pediatrics