One of the major mile posts in biomedical ethics was the passage of legislation that today’s medical residents—and I daresay most Americans—have never heard of, the Patient Self-Determination Act of 1990. Certainly all those who inveighed against “death panels” and who balked at the idea that the Affordable Care Act might include provisions allowing Medicare to reimburse for advance planning conversations never heard of it. This act, as its title indicates, was intended to put patients in the driver’s seat, to allow them to weigh in on the approach to medical care they would get at the end of life, even if they were unable to participate in decisions at the time those decisions need to be made. It officially sanctioned advance directives by enshrining them in federal law. 

What the PSDA says is that any health care institution receiving government funds, which is to say virtually all health care institutions, is obligated to ask patients if they have an advance directive, to offer them the opportunity to complete one if they don't, and to prominently display a copy of that directive in the medical record if they do. The PSDA put advance care planning on the map. It also put advance care planning squarely in the legal domain and that, as the authors of a new article in the New England Journal assert, was a big mistake.

What “Delegalizing Advance Directives—Facilitating Advance Care Planning” argues is that a major reason that advance directives haven’t caught on is that they typically have to be signed by two witnesses (or a notary), and in some states (North Carolina, South Carolina Virginia, and Missouri) a notary; some states also require use of a specified form. I agree that these requirements are an impediment to widespread use of advance directives. I agree that the POLST (Physician Orders for Life-Sustaining Treatment) model, which uses a medical order rather than a legal document and just requires the signature of the patient and the physician, puts advance care planning unambiguously within the medical sphere. But I don’t think that simply allowing patients or prospective patients to designate a health care agent, or surrogate decision-maker, without use of a mandated form and without witnesses would solve the problem. 

The real problem is not just that people don’t bother with the forms and that the forms don’t always make it into the hands of clinicians. The real problem, as Charlie Sabatino of the American Bar Association put it in a 2010 article, is that advance directives are based on a transactional view of advance care planning rather than a communications model. And what we now understand is that advance care planning has to be founded on dialogue between a clinician and a patient.

The problem with advance directives is not that they have to be witnessed or written on special forms. If that was the problem, we’d expect to see much higher utilization rates in Idaho, where there are no witness or notary requirements, and somewhat higher utilization rates in Utah, where only one witness is required. The problem is that they reduce advance care planning to completing a form, to checking off boxes on a list. 

In Idaho, for example, individuals have the opportunity to say that if they are ever unable to communicate and “have an incurable or irreversible injury, disease, illness or condition,” and that a medical doctor, based on a physical examination, has concluded that the “injury, disease, illness or condition is terminal” and that “death is imminent” no matter what is done, and that the “application of artificial life-sustaining procedures” would only artificially prolong life, then they would (or would not) want medical treatment and procedures, nutrition and hydration, hydration but not nutrition…

The lawyers who design such forms believe they allow people to indicate with great precision just what they would want and under what circumstances they would want it. But in fact, as many others have observed, it is far from clear what exactly it means for a condition to be terminal. I would argue that dementia is a terminal disease—but with a typical time course of about five years from the time of diagnosis to death. Ah, you might say, but the advance directive forms include the qualifier that death must be imminent. But that's not good enough. How imminent? In a matter of hours? Days? Months? And if we could agree, based on a careful reading of the text of the directive (which I’m not so sure we can), that what is meant is the person has a  disease that in the normal course of things results in death within six months and that the person's disease has progressed to the point where death will occur within at most days; then is it really so useful to specify that in those very limited circumstances we wouldn’t want treatment that won’t make a difference anyway? Is that all the advance care planning is about—stopping futile treatment in the 72 hours of life? And what about “treatment” that is symptomatic, that is intended to ameliorate suffering, rather than to prolong life, though it might, as an unintended consequence, prolong life. Are such “medical treatment[s] and procedures” to be rejected?

Advance care planning, as we have come to understand it over the last several decades, is not about procedures or treatments—or checking boxes. It is about reflecting on what’s important, in the context of a realistic understanding of a person’s medical condition. It's about figuring out what medical treatments are most consistent with achieving whatever it is that the patient deems important in life. 

Making it easier to complete a form will not transform advance directives. Conceptually, advance directives are legal documents, whether or not they must be witnessed or notarized or completed on special paper. What people need is not a better document. It’s a different process, a process that is built on communication and that deals with the purpose of medical treatment, not the technical means of achieving those ends.

-->