Answers and questions about ethics consultations.

ETHICS CONSULTATIONS ARE TOUTED FOR RESOLVING ETHIcal dilemmas, and most hospitals use them to meet the Joint Commission for the Accreditation of Healthcare Organizations’ requirement for a process to resolve conflicts in patient care. Little rigorous evidence is available, however, about the outcomes of ethics consultations. In this issue of THE JOURNAL, Schneiderman and colleagues report a multisite randomized controlled trial to evaluate ethics consultations. The intervention, which builds on a previous single-center randomized trial, involved ethics consultation in the intensive care unit (ICU) for conflicts and value disagreements either among the health care team or between the team and patients or surrogates. The comparison group received standard care, in which 25% of those patients chose ethics consultation. Schneiderman et al found that patients in the intervention group who died spent 3 fewer days in the hospital, 1.4 fewer days in the ICU, and 1.7 fewer days receiving mechanical ventilation than nonsurvivors in the control group. Almost 90% of surrogates and physicians agreed or strongly agreed that the ethics consultation was helpful. The authors concluded that ethics consultations reduced nonbeneficial care without increasing mortality. The study has many methodological strengths, and overall it raises the standard for evaluating ethics consultations. Randomization by site, or cluster randomization, removed many potential sources of bias. Both process and outcome end points were assessed. Analysis by intention to treat led to a conservative estimate of the true effect of the intervention. However, several methodological concerns should be considered. An important limitation was that the intervention was not standardized. Each institution followed its own procedures, which were characterized as consistent with general guidelines for ethics consultations. Clearly, what is done in an ethics consultation depends on the particular patient and the person doing the consultation. But because a consultation is not standardized like a drug, it is important to have more detail about what was actually done. Without such information, other institutions cannot judge whether the consultations in this study are similar to those at their own, or how they would replicate the intervention. In addition, the intervention group had a slightly higher mortality rate (62.7% vs 57.8%). Although this difference was not statistically significant, it may nonetheless be clinically and ethically meaningful. In a randomized trial, differences in outcomes that are not explained by differences in baseline characteristics or other concomitant interventions are ascribed to the study intervention. In contrast to clinical trials of drugs, it is not clear whether a higher mortality rate in this study would be good or bad. The authors argue that their findings show that ethics consultations do not “simply provide a subterfuge for ‘pulling the plug.’” But this claim requires judgments about the ethical content of the consultations. Recommendations of ethics consultations should be within the “boundaries of morally acceptable solutions.” In some circumstances, an ethics consultation appropriately would lead to forgoing lifesustaining interventions. For example, the consultation might clarify that the patient would not have wanted the interventions or might correct the common misconception that withdrawing life support is synonymous with killing the patient. Conversely, itwouldbewrong for anethics consultation to recommend limiting life-sustaining interventions based on inappropriate claims about futility that physicians sometimes offer. Thus, it would be important to know more about the kinds of disagreements and value conflicts in this study and how they were addressed. It also would be useful to assess whether independent reviewers agreed with the recommendations of the consultations. Finally, the conclusion that patients or surrogates viewed the consultations as helpful rests on somewhat weaker evidence than the objective outcomes. In follow-up interviews, surrogates who did not speak English were excluded. Values conflicts and disagreements with such families may be particularly acute. Many cultures reject core assumptions of the US approach to end-of-life decisions: that it is desirable for patients to anticipate future illness, to give directions for future care, and designate a single person as proxy. Also in light of problems with translation services, it would be important to know whether non–English-speaking families understood the goal and process of ethics consultation. Moreover, 13 surrogates disagreed with the recommendations of the ethics consultation. It would be helpful to know more about the nature of disagreements and recommendations in these cases, provided that such information could be pre-