Rethink "Affirmative Agreement," but Abandon "Assent"

In the United States the ethical and regulatory regime governing children's participation in nonbeneficial research requires both the permission of the child's parent and, where the child is judged capable, his or her affirmative agreement (45 CFR 46.116d; American Academy of Pediatrics Committee on Bioethics 1995). This system combines a historically-grounded protectionist bias, which defaults to nonparticipation unless these two conditions are met, with a desire to maximize respect for children's developing autonomy. As David Wendler and Seema Shah (2003) point out, however, this scheme has serious problems. First, it fails to specify the ethical or practical conditions for valid assent. It is not clear, for example, whether understanding is necessary for assent to be valid (witness how rarely one comes across the term "informed assent"). Second, it provides no guidance as to what role children who are judged incapable of providing affirmative agreement might play in decisions about research participation. Finally, by conceiving of children and their parents as radically separate moral agents (Crouch and Elliott 1999), it neglects the interconnected nature of healthy decision making within families.

In light of these problems, Wendler and Shah propose substantial revisions to our thinking about children's roles in decisions about research participation. Drawing both on studies of children's cognitive development and on a normative analysis of the principles underlying our obligation to solicit a child's agreement, they suggest that affirmative agreement be required only of those children whose capacity for understanding approaches that of adults. They base their recommendation on a skepticism, which I share, about the moral value of children's affirmative but uninformed agreement. Their reading of the empirical literature leads them presumptively to place this capacity boundary at 14 years. They also recommend that we respect the dissent of all children, regardless of whether or not they are judged capable of assent.

I am not convinced that a sharp distinction at age 14 makes sense. Recent work by Tait, Voepel-Lewis, and Malviya (2003) suggests that while children between seven and ten manifest limited understanding of their research studies, children aged 11-14 perform only slightly worse than older adolescents. Also, age is a poor proxy for children's abilities to comprehend research (Susman, Dorn, and Fletcher 1992; Dorn, Susman, and Fletcher 1995). Nevertheless, Wendler and Shah are correct in their essential point: without attention to understanding, or at least capacity for understanding, the ethical grounds for requiring children's affirmative agreement are shaky at best. Unfortunately, like most other discussions of assent, their analysis leaves out two important factors that we must take into account with respect to children's roles in research decisions. I will argue that, to be complete, we must consider both the healthy developmental trajectory of shared decision making between children and parents, and the reasons that children hold for their preferences. Finally, I will suggest that whatever we decide to do about affirmative agreement and dissent, we must abandon the word assent as a way to describe it.

Rethinking Affirmative Agreement

How do families ordinarily apportion decision-making authority between parents and their children, particularly with regard to issues that affect the child? Figure 1 proposes a model that seems normatively reasonable and likely empirically correct. Parents assume essentially complete responsibility for decisions of any significance regarding infants and toddlers. However, during early childhood parents begin to involve their children in discussions about at least some decisions and under limited circumstances might even take their children's preferences into account. As their maturity and cognitive abilities grow, children play an increasing role in decisions about their lives, and their parents begin actively to solicit their input. As Figure 1 suggests, there is likely to be a time when the relative weights of the parents' and child's preferences shift rapidly. I imagine, tentatively, that this period of rapid change occurs from approximately age eight...