As palliative care practitioners have struggled to find ways to respond to intractable end-of-life suffering without resorting to physician-assisted suicide, terminal sedation and voluntarily stopping eating and drinking have been suggested as morally acceptable alternatives that do not require changes in the law (1, 2). We agree with Sulmasy and colleagues that these practices are much more morally and clinically complex than is ordinarily acknowledged (3). For this reason, we support guidance for those who might participate in these practices even though it may not be legally required (4); the University of Pennsylvania Center for Bioethics Assisted Suicide Consensus Panel also supports providing this guidance. We agree that the practices should be rare and restricted to highly selected patients and that the primary intention of participating clinicians should be relief of severe suffering. However, requiring that the intent must never also be to hasten death may at times create an impossible dilemma for the care of patients who are near death, are suffering terribly, and are psychologically prepared to die. Under the rule suggested by Sulmasy and colleagues, if the patient consents to terminal sedation exclusively to escape suffering, then the intervention may be permitted. If a patient's intent is also to hasten inevitable death, then suddenly the same intervention would become impermissible.

Guidelines can reinforce careful decision making and ensure access to standard palliative care before terminal sedation and the voluntary refusal of food and fluids are seriously considered. Without such guidance, "inexperienced hands"—the term used by Sulmasy and associates—might consider these options of last resort to be part of standard palliative care. With guidelines, the primary physician must document that standard palliative care has failed, that the patient is competent, and that all meaningful alternatives have been considered. Furthermore, these findings would have to be corroborated by a specialist in palliative care and by a mental health specialist if there is uncertainty about the patient's mental status. The question posed is whether appropriate access and oversight are promoted by formally accepting such guidelines, or whether the practices should remain solely at the discretion of individual providers.

In the current environment, we know that the use of terminal sedation varies from 0% to 50% of deaths in palliative care programs (5). Such variation by program suggests that current practices are driven more by the values of practitioners and programs than of patients and families, when just the opposite should be true.