Palliative Care in Indigenous Populations: The Question That Does Harm

The advance directive is a cultural artifact. It has a federal statute behind it, which makes it easy to mistake for a neutral instrument.

The Patient Self-Determination Act, passed by Congress in 1990 and in force since the end of 1991, requires every hospital, skilled nursing facility, home health agency, and hospice that takes Medicare or Medicaid money to ask patients at admission whether they have an advance directive and to inform them of their right to execute one. The statute encodes a theory of a good death, and the theory has a sequence. The patient is told the prognosis. The patient forms a preference. The preference is written down. The writing governs.

Each step in that sequence is a claim about how a human being ought to meet the end of a life. Each is contestable. In at least one framework, well documented and still in force, the first step is not merely unwelcome. It is understood to cause harm.

In September 1995, the Journal of the American Medical Association published two papers in the same issue on culture and end-of-life decision-making. One of them is cited constantly. The other is cited far less, and it carries the harder finding.

Joseph Carrese and Lorna Rhodes, in "Western Bioethics on the Navajo Reservation: Benefit or Harm?", reported interviews conducted on the Navajo Nation with patients, traditional healers, and biomedical providers about how prognostic discussion is received. What they found was not a preference for less information. It was a different account of what speech does.

Navajo thought organizes around hózhó, usually rendered in English as harmony, order, or beauty. Maintaining hózhó is an active discipline of thought and language rather than a passive state. Gary Witherspoon's Language and Art in the Navajo Universe (1977) documents the underlying premise: language does not passively describe a world that exists independently of it. Language participates in the world's formation. To speak a thing is to move toward it.

The clinical consequence is direct. The question the statute requires, the one about whether you want us to try to restart your heart, is not the collection of a preference. It is the introduction of a possibility into a room where someone is working to stay alive. Carrese and Rhodes' informants described exactly that. The advance-directive conversation, conducted correctly and according to protocol, was experienced by some as an act that made the bad outcome more likely.

The physician who complies with the statute delivers a harm. The physician who declines to comply withholds a right the law was written to protect. Carrese and Rhodes put the alternatives in their title as a question. The honest answer is that both were true at once, and that the conflict was manufactured by the form rather than presented by the patient.

The second paper in that issue was Leslie Blackhall and colleagues' "Ethnicity and Attitudes Toward Patient Autonomy," a survey of eight hundred people aged sixty-five and older in Los Angeles across four groups: European American, African American, Korean American, and Mexican American. The differences were large. Korean American and Mexican American respondents were substantially less likely than the other two groups to believe that a patient should be told a terminal diagnosis, and less likely to believe that the patient should be the one to decide about life-sustaining treatment. Family-centered decision-making was not a departure from self-determination in those groups. It was how self-determination was exercised.

No American Indian or Alaska Native population was sampled.

That absence is the subject of this essay. The most-cited American study of culture and end-of-life decision-making did not include the population in which the conflict with the statute was sharpest, and the paper that did include it, published on the same day in the same journal, entered the ethics curriculum rather than the clinical one.

The gap is not only a matter of conversation. It is built into the benefit.

The Medicare Hospice Benefit, created in 1982, requires three conditions before care begins. The hospice medical director and the patient's attending physician must certify a prognosis of six months or less if the illness follows its expected course. The patient must elect the benefit, which for adults means giving up Medicare payment for treatment aimed at the terminal condition. And a certified hospice agency must be willing and able to serve the address.

Each condition is harder to satisfy in Indian Country. Certification requires the prognosis conversation, which is the disclosure problem again, now with a signature attached to it. Election requires a declaration that treatment has stopped, which in many families is heard as a statement about the person rather than about the medicine. Service area is geography: hospice is a home-visiting model, and it needs a house it can reach, on a road it can find, inside a radius a nurse can drive twice a week and still get home.

Children were released from the second condition. Section 2302 of the Affordable Care Act requires state Medicaid programs to let children receive hospice care concurrently with treatment directed at the illness. Adults were not released. The Medicare Care Choices Model, the federal demonstration that tested concurrent supportive care for adults, ran its course without changing benefit design.

American Indian and Alaska Native Medicare decedents enroll in hospice at lower rates than white decedents. The finding is usually reported as a difference in preference. Some of it is. The rest of it is that the benefit was built around an address, a family willing to say the word out loud, and a prognosis someone was prepared to speak.

The form does not perform especially well in the population it was designed for either. In 2021, R. Sean Morrison, Diane Meier, and Robert Arnold published a paper in JAMA titled "What's Wrong With Advance Care Planning?" arguing that three decades of research had not shown that completing advance directives reliably produces care consistent with what patients say they want, and that the field's investment in documents had crowded out investment in the conversations the documents were supposed to record. An instrument that underperforms where its assumptions hold is a poor candidate for export to places where they do not.

The Research That Followed the Practice

The standard account of clinical knowledge runs in one direction. Theory generates a hypothesis. A trial tests it. A guideline carries the result to the bedside. The palliative care knowledge base for Indigenous populations did not form that way. It formed backward.

Carrese and Rhodes did not begin with a theory of Navajo metaphysics and then look for a clinical application. They began with the observation that a legally required conversation was producing distress, and they went and asked people why. The question came out of the work.

This is the standpoint claim restated as research design: what can be asked depends on where the asker stands, and the distance between a question written in a university office and a question written by people who have watched the protocol fail is position rather than rigor. Issue 026 of this publication argued that community-based participatory research is what strong objectivity looks like once it is written into a protocol, and that individual consent cannot carry a risk the community is the one to bear; the Havasupai case is the standing demonstration of both. Palliative care research inherits the problem in its sharpest form, because dying is the most guarded subject a community has.

Canada wrote a governance answer. Marlene Brant Castellano, writing on the ethics of Aboriginal research in the Journal of Aboriginal Health in 2004, set out the terms, and the instrument that carries them is OCAP, for ownership, control, access, and possession, asserted through the First Nations Information Governance Centre. The community owns the knowledge produced about it and decides what is done with that knowledge. That is a claim of title rather than a permission, and it does not expire when the study closes, which is the one thing a consent form cannot do.

That is not a procedural nicety when the subject is dying. You cannot study dying in a place that will not tell you about its dead, and no protocol earns that on the strength of a signature collected at enrollment. Trust is the instrument, and instruments get built before they get used.

Who Decides Whether the Care Was Safe

Irihapeti Ramsden was a Māori nurse in Aotearoa New Zealand who spent two decades developing a concept she called kawa whakaruruhau, rendered in English as cultural safety, and completed a doctoral thesis on it at Victoria University of Wellington in 2002. Her move was small in statement and total in consequence.

The person who determines whether an episode of care was culturally safe is the person who received it.

Not the clinician. Not the clinician's profession. Not an accrediting body holding a checklist. Cultural competence is a property of the practitioner, assessed by the practitioner's own field, demonstrable by completing a course. Cultural safety is a property of the encounter, assessed by the person the encounter happened to. The two terms sound like variants of one idea. They put authority in opposite places.

Issue 019 of this publication set out Madeleine Leininger's transcultural nursing, and put to it the objection that her category of repatterning assumes a clinician who can identify a practice as harmful without ever saying who decides that, or from what standpoint. Ramsden's contribution is different in kind. Leininger supplied the method. Ramsden moved the judge.

That objection changes character at the end of life. Elsewhere it turns on a practice with an outcome attached, a delayed presentation or a refused treatment, and the clinician can at least point at the outcome. At the end of life the outcome is the thing in dispute, because what counts as a good death is itself a cultural product. The clinician who holds that dying patients should be told everything, should die at home, should have a documented surrogate, and should not receive treatment that will not extend life is holding four positions, all defensible, none neutral, and each easier to hold as a fact than as a position.

What follows is a composite and not a patient. An eighty-one-year-old man with metastatic cancer is admitted to a small rural hospital several hours from where he lives. He is alert and oriented. Before the team reaches the room, his daughter stops them in the hallway and says the family has decided he is not to be told how long.

The team's first instinct is that this violates the man's right to know and possibly federal law. The instinct is not wrong about the law. It is wrong about what was asked. Nobody said he should not be told he is dying. The family said not to name a number.

Those are separable acts, and the American form does not separate them. Prognostic disclosure and prognostic quantification arrived together in the template and have been treated as one thing since. Families draw the line where this one drew it more often than the template allows, and the line is defensible on evidence. Nicholas Christakis and Elizabeth Lamont reported in the BMJ in 2000 that physicians' survival predictions for terminally ill patients were accurate within a third of actual survival only about a fifth of the time, and that the errors ran heavily toward optimism. The number the family is being asked to receive carries an interval around it that rarely gets stated in the room.

The conversation that follows the hallway is where the clinical work sits. Who in the family is to be told first, and by whom. Whether there are people who must be present before anything is said. Whether being at home matters enough to reorganize the medicine around it, and what would have to be true for that to happen. Whether there are words that should not be spoken in the room, and what the man would want said instead of them.

Arthur Kleinman, Leon Eisenberg, and Byron Good set out the interview that produces these answers in "Culture, Illness, and Care," published in the Annals of Internal Medicine in 1978. What do you think caused this. Why do you think it started when it did. What do you fear most about it. The questions were written for a patient. They work on a family, and at the end of life the family is frequently the unit holding the answer. Nothing they retrieve is about the disease. All of it is clinical.

Where the Form Came From

The founding claim of palliative care is that the answers to those questions are medicine, and a nurse made it.

Cicely Saunders trained as a nurse before she trained as anything else. She entered the Nightingale School at St Thomas' Hospital during the Second World War, left bedside nursing after a back injury, worked as an almoner, which is what a hospital social worker was then called, and qualified in medicine in her late thirties. She opened St Christopher's Hospice in London in 1967. The concept she is remembered for is total pain: the claim that the suffering of a dying person has physical, psychological, social, and spiritual components, and that treating the physical component alone leaves the person in pain.

Total pain was an argument that dying is not only a physiological event, and that the parts of it which are not physiological are clinical rather than decorative. The form that American palliative care exported inverted the argument. It made the physical part the medicine and the social and spiritual parts a preference, recorded in a field, after the medical work was finished.

Saunders was formed in nursing first and medicine third. That order was the condition of the insight. An almoner's job was to establish what a household could absorb, which is not a question the ward round was asking. The discipline kept what she saw and lost the position she saw it from.

Ramsden's move and Saunders' training order point at the same thing. A form cannot be examined from the position that wrote it. What Navajo informants told Carrese and Rhodes was not new information about culture. It was information about the form. The form was the thing that had to change, and the only people positioned to say so were the people it was being used on.

About the author

Peter Schindler, MD, PhD is an Assistant Professor of Medicine and Associate Program Director of the Community Health Center Family Medicine Residency Program at the University of Nebraska Medical Center. He practices at Winnebago Comprehensive Health System, OneWorld Community Health Center, and Nebraska Medicine. He completed a Primary Care Research Fellowship at McGill University and holds a BSN from the University of Wisconsin-Oshkosh, an MS and PhD in nursing from Emory University's Laney Graduate School, a Diploma in Tropical Medicine from the Liverpool School of Tropical Medicine, and an MD from the Medical College of Wisconsin-Green Bay. He is board certified by the American Board of Family Medicine. The Interprofessional publishes new essays every week at the intersection of medicine, nursing, and the clinical knowledge that lives between them.

Disclaimer. The views in The Interprofessional are Peter Schindler's own and do not represent the official positions of the University of Nebraska Medical Center, Winnebago Comprehensive Health System, OneWorld Community Health Center, Nebraska Medicine, or any other affiliated institution.