The Bad News First

"NO surprises” is a basic rule in hospitals. Junior doctors are supposed to notify their superiors promptly about worrisome developments in a patient, and information is supposed to move smoothly up the chain of command. One of the gravest errors a doctor in training can make is to inform the attending physician well after the fact about a patient’s turn for the worse.

Unfortunately, this rule does not extend to seriously ill patients themselves. They and their families are frequently surprised by the sudden imminence — and the raging authority — of death.

Research has revealed doctors’ tendency to contribute to the problem by avoiding making prognoses. In one study of nearly 5,000 hospitalized adults who had roughly six months to live, only 15 percent were given clear prognoses. In a smaller study of 326 cancer patients in Chicago hospices, all of whom had about a month to live, only 37 percent of the doctors interviewed said they would share an accurate prognosis with their patients, and only if patients or their families pushed them to do so.

Even when doctors do prognosticate, the research shows, they typically overestimate the time a patient has left to live, often at least tripling it, perhaps because they feel overconfident. The pugilistic attitude most doctors adopt toward disease is understandable, even desirable, for much of the course of illness. But there comes a time when this attitude can lead to false optimism. Doctors who wrongly think that patients are going to live much longer wind up recommending needlessly painful and expensive treatments. This phenomenon is neatly captured by a gallows-humor joke told by hospice nurses: Why are coffins nailed shut? To keep doctors from administering more chemotherapy.

By not making or communicating prognoses, doctors can make the end of life more unpleasant. Patients are given no chance to draft wills, see distant loved ones, make peace with estranged relatives or even discuss with their families their wishes about how to live the end of their lives. And they are denied the chance to make decisions about what kind of medical care they want to receive.

Roughly half of Americans die with inadequately treated pain. Large minorities suffer symptoms like shortness of breath, nausea or depression. Four in five die in hospitals and nursing homes, rather than at home as most prefer. And more than half significantly burden family caregivers in the course of their final illness: the family loses its life savings, a caregiver has to quit work or a spouse falls seriously ill.

For reliable prognoses to become a routine part of medical care they must become a priority of medical research and education. Less than 5 percent of research focuses on prognosis. Textbook descriptions of diseases cover prognosis less than 25 percent of the time. And medical schools and residency programs almost completely neglect training in prognostication.

Greater investments in new statistical tools and databases that help physicians predict outcomes are also needed. With these, doctors could translate the clinical, biochemical and genetic information they collect on their patients into statistical predictions of life expectancy that could supplement their own clinical judgment.

Doctors often say they worry that predictions about survival may become self-fulfilling prophecies or cause patients to lose hope. But a realistic assessment of how long a patient has to live need not cause either the patient or doctor to become pessimistic. It should only refocus attention on the quality of the patient’s life. Sometimes living life to its fullest requires knowledge of its finitude.

Nicholas A. Christakis, a physician and a professor of sociology at Harvard, is the author of “Death Foretold: Prophecy and Prognosis in Medical Care.”