indicates that the Baby Doe rules have had a substantial chilling effect on treatment decisions in cases where physicians think that further treatment for particular patients will have disproportionately small benefits and extreme burdens for patients and families. Recall that Kopelman’s early study of neonatologists’ attitudes shortly after the effective date of the Baby Doe rules supports this position [14]. In adult critical care settings there is also a range of legal defensiveness but, in contrast, research suggests that legal defensiveness will generally decrease as knowledge of pertinent law increases. The major exception to this trend is that in states like New York, where there are highly restrictive and specific laws, such highvisibility laws will also have a substantial chilling effect on clinical practice, especially decisions to abate life-sustaining treatment. Our research on anticipated conflict may also prove helpful in illuminating these issues – if legal constraints, perceived or real, create in the minds of physicians a posture of anticipated conflict with patients and families, such attitudes may undermine the important clinical goals of minimizing suffering of patients, promoting group process, and facilitating self-determination in decision making. Environments in which critical care is practiced may also be affected comprehensively by legal defensiveness. Disputes among individual physicians and nurses about specific cases may spread and contribute to intra-unit tension and conflict. This may be partially attributable to differences in practice style generally, but in my experience can certainly be exacerbated by varying perceptions of legal risk among critical care providers. Moreover, individual nurses’ and physicians’ attitudes about these issues can affect the culture of an entire unit if practitioners are concerned that staff who disagree with their practice patterns may report activities they believe inappropriate to state regulatory or law enforcement officials. Decisions to abate lifesustaining treatment may be delayed or halted by such conflict between physicians and staff. In this sense, physicians and nurses with the highest legal defensiveness in a particular unit may limit abatement of lifesustaining treatment by default, with little regard for the interests of patients and families (and, in NICUs, the Baby Doe rules provide cover for such practice). In my view, this problem is especially acute in contexts with recognized elevated levels of legal risk, whether they be neonatal or adult settings. My clinical experience also suggests that substantial deficits in legal knowledge among practitioners in both neonatal and adult critical care settings persist. In environments of highest anxiety about legal risk, it seems unlikely that learning more about the actual constraints of the law will ameliorate defensiveness. Such study of law may in fact increase defensiveness in certain contexts, as our New York research suggests. In general, it seems to me that the degree to which the actual law is targeted toward specific clinical activities or particular specialties and sub-specialties (for example, terminating treatment for neonates or removing nutrition for adult patients in New York) the more anxious physicians will be about engaging in those types of activities. This explanation is consistent with both the findings of our research group and with those of Kopelman and colleagues. To summarize, while it may be mostly unfounded and counterproductive for physicians to have excessive anxiety about the law in many other adult critical care contexts, fear of significant legal constraints in neonatal contexts may be quite reasonable because of the Baby Doe rules. As the oft-used saying goes, “it’s not paranoia if they are out to get you.” The trick for practitioners, of course, is determining the reasonableness of one’s fears and acting appropriately. Unfortunately, while this internal debate about the law persists among providers, practicing optimal, ethical medicine may be impeded at the expense of patients’ and families’ well-being. Thus, unnecessary suffering may be the primary legacy of these highly restrictive rules. In neonatal contexts, the high level of clinical uncertainty about mortality and morbidity may also exacerbate the impact of legal defensiveness by causing clinicians 78 S. Van McCrary / Neonatology and the specter of the law Table 1 Multivariable regression analysis1 of effects on legal defensiveness Model 1: Main effects Model 2: Interaction effect Degrees of Parameter Standard Degrees of Parameter Standard freedom estimate error t freedom estimate error t New York compared to Texas 1 −0.12 0.07 −1.78† 1 −0.60 0.23 −2.59∗∗ Lawscore (number correct) 1 0.00 0.00 0.20 1 −0.01 0.01 −1.98∗ New York x lawscore interaction 1 0.01 0.00 2.17∗ Model F = 4.87∗∗∗ Model F = 4.87∗∗∗ Adjusted R2 = 0.04 Adjusted R2 = 0.05 Statistical significance: †p < 0.10; ∗p < 0.05; ∗∗p < 0.01. 1Models are controlled for years experience and medical specialty. to second-guess their decisions. This may explain to some degree the apparent divergence of neonatal practice in the United States when compared with other developed countries that have more uniform policies on the provision of critical care. 6. Toward the future: A research agenda I will now present an overview of a possible research and policy agenda. An important threshold question is: To what degree there really exists an identifiable