In 1972, Jerry Canterbury went in for back surgery, suffered a postoperative fall from his hospital bed, and ended up paralyzed. His surgeon, Dr. William Spence, hadn't mentioned that paralysis was a risk. Canterbury sued, and the D.C. Circuit Court of Appeals ruled that physicians have a duty to disclose whatever a reasonable patient would want to know before consenting to a procedure. The decision helped launch a revolution. Over the next three decades, American medicine would transform its foundational ethic from "the doctor decides" to "the patient decides," enshrining autonomy, informed consent, and patient choice as the bedrock principles of clinical care.
That revolution was right. It corrected real abuses: decades of physicians withholding diagnoses, performing procedures without meaningful consent, and treating patients as passive recipients of medical benevolence. The Belmont Report in 1979 made "respect for persons" a foundational principle. Informed consent law expanded. Bioethicists wrote entire careers' worth of scholarship on why the old paternalism had to go.
But somewhere between "the doctor should not decide for you" and "you must decide for yourself," the project went sideways. And a growing body of scholarship is arguing that we need to talk about it.
You might expect patients to want full decisional sovereignty. The empirical literature suggests otherwise, and the pattern is very consistent across studies.
A 2012 study of hospitalized patients at the University of Chicago found that 97% wanted their doctors to offer choices and consider their opinions. So far, so autonomy. But 67% of those same patients preferred to leave the final medical decision to the doctor. Read that again: two-thirds of patients, in a modern American hospital, wanted their physician to make the call. They wanted to be heard, not enthroned.
The paradox is that shared decision making (SDM), as commonly practiced, often degrades into exactly the thing it was designed to prevent. Many clinicians interpreted "shared decision-making" to mean "never recommend," fearing that any expression of professional opinion would make them paternalistic. The result was a distinctive clinical posture: scrupulously neutral, informationally generous, and existentially useless. Present the options, describe the risks, list the benefits, and then stare expectantly at the person in the hospital gown, as if they just materialized on earth five minutes ago with no preferences, no fears, no need for professional guidance.
The argument: in serious illness and end-of-life care, the autonomy framework often becomes a mechanism for offloading impossible decisions onto patients and families. A surrogate who is told "your mother can go on the ventilator or we can pursue comfort measures; it's your choice" isn't being respected. They're being burdened with a life-and-death decision they have no framework for making, and they may carry guilt about that decision for years.
The anti-paternalist revolution happened because physicians really did silence, mislead, and overrule patients. Patients were routinely not told they had cancer. Women were sterilized without consent. Research subjects were experimented on without knowledge. Any argument for restoring physician authority has to contend with the fact that physician authority was, within living memory, regularly abused. That history doesn't disappear because we've gotten better at ethics training.
But pure menu autonomy is often a fiction, and sometimes a cruel one. A frightened, exhausted, cognitively overloaded patient staring at a list of treatment options they cannot evaluate is not exercising self-governance in any philosophically serious sense. They are exercising the right to be confused and alone. Most patients don't want that. The informed consent data suggests they aren't getting real autonomy anyway. And the clinicians who refuse to recommend aren't being respectful; they're being absent.
None of this requires going back to the bad old days. Medicine is a relationship, not a vending machine. The patient puts in their values; the doctor puts in their knowledge; and what comes out, ideally, is a decision neither could have reached alone.
The pendulum swung away from paternalism for excellent reasons. But it swung too far. The profession built an elaborate ethical infrastructure around the idea that doctors should present and patients should choose, and in doing so it created a system where the most common patient encounter with “autonomy” is bewilderment. The interesting question now is how to build a clinical culture where physicians are neither dictators nor bystanders, where recommendations are expected and transparent and revisable, and where “autonomy” means something richer than being left alone with a terrible choice.
I think doctors should recommend more, hedge less, and trust that a patient who disagrees will say so. That’s not paternalism. But it’s closer to paternalism than the current orthodoxy is comfortable with, and I think the current orthodoxy is wrong.