Until a severe stroke sent him to a neurological intensive care unit in December 2014, Ernest Kohn was a particularly vibrant 90-year-old, still teaching a graduate economics class at Queens College.
So his family thought he might rebound. But when his son, Jerry, asked the rotating flotilla of neurologists what was likely to happen — Would his father survive? Go home? — no one really wanted to address his questions. “When you pushed them, they said, ‘We can’t say anything with surety,’” Kohn said.
“I kept saying, ‘I’m only asking for your opinion, not a guarantee. I’d really like to know what your 30 years of medical knowledge and experience tell you.’ Most of them would just clam up.”
We’ve known for years that doctors hesitate or even decline to discuss a poor prognosis with patients and their families. They fear that bad news will dash hopes; they don’t want to appear to be giving up. Often, their training hasn’t prepared them for sensitive conversations.
One researcher told me that oncologists believe that if they fail to offer yet more chemotherapy, even when that’s futile, patients will leave, seeking another doctor who will.
Yet the supposed cornerstone of contemporary medicine — patients making informed decisions — depends on their understanding their situation, their life expectancy, their probable quality of life, the pros and cons of any proposed treatment. (Or, when patients themselves are incapacitated, it depends on their surrogate decision-makers’ understanding.)
Experts have urged doctors to talk about the elephants in the room, especially at the end of life. But two recent studies show how achingly slow progress has been.
Even terminally ill patients still receive scant information, researchers have found, while family members acting for ICU patients commonly contend with confusion and misinformation. The studies also uncover some reasons for the disconnect. Doctors, it seems, shouldn’t get all the blame.
Dr. Holly Prigerson, director of the Center for Research on End-of-Life Care at Weill Cornell Medicine, and her colleagues interviewed 178 patients at cancer centers across the country. All had cancers that had progressed despite chemotherapy; their oncologists estimated their life expectancy at less than six months. “These patients were all dying, and everyone treating them was well aware of it,” Prigerson said.
Yet nearly 40 percent said they’d never discussed prognosis or life expectancy with their oncologists. Not surprisingly, when asked to answer four key questions about how well they understood their illness — including whether they grasped that their cancer was incurable and that they had months, not years, to live — only 5 percent answered correctly.
Researchers once attributed such results to communication failures: Doctors weren’t talking about prognosis or they weren’t using language that lay people understood.
“We made a lot of assumptions that with good information people would make good, clear, rational decisions,” said Dr. Douglas White, who directs a program on ethical decision making at the University of Pittsburgh.
Reality turns out to be considerably more complicated.
In White’s earlier research, asking surrogates with relatives in intensive care to respond to hypothetical prognoses, family members were markedly more likely to correctly interpret optimistic predictions (“a 90 percent chance of surviving”) than pessimistic ones (“a 5 percent chance”).
That’s called “optimism bias.” But the team’s latest study, in JAMA, shows that it is hardly the only muddying factor.
White and his colleagues surveyed 229 surrogates and 99 physicians involved in the care of critically ill patients, most older than 65, who’d already spent five days on ventilators in intensive care. The researchers asked physicians to estimate how likely each patient was to survive the hospitalization, from 0 to 100 percent; they asked family members what they thought their physician’s prognosis was.
More than half the time, the researchers found substantial “discordance” — at least a 20-percentage-point difference between the physician’s assessment and what the surrogates thought it would be.
Understanding what lies ahead can profoundly affect patients’ quality of life — and death. If they underestimate their life expectancy, they may forgo helpful treatment. If they overestimate it — the more common misperception — they may agree to tests and procedures that turn their final weeks and months into a medical treadmill.
Frank discussions don’t disrupt the bond between doctors and patients, Prigerson has shown. They do increase the likelihood that patients receive the end-of-life care they prefer, and leave survivors better able to cope with grief.