News & Politics » Feature

Prognosis: Death

Why won't doctors level with terminally ill patients? Nicholas Christakis revives the lost art of prognosis.



By Jeff Sharlet

It's a rainy Sunday afternoon, and Nicholas A. Christakis has just received sad news. Frances Holbrook*--"a wonderful old lady," he says--died that morning before he could get to her. Not that he could have saved her, or would have tried: Christakis is a doctor, but his patients rarely survive. His practice is in hospice care.

When a classmate asked his oldest son, then five, what his daddy did, the boy replied, "He teaches people how to die." In fact Christakis teaches sociology and medicine at the University of Chicago. But it's true that he thinks some deaths are better than others, and he believes doctors can help people achieve them. One day a week he does so as a hospice physician, but he hopes to go beyond house calls with a new book, published this month by the University of Chicago Press, Death Foretold: Prophecy and Prognosis in Medical Care--the first full general discussion of prognosis ever written. Traditionally the third leg of medicine's triad, after diagnosis and treatment, prognosis has been neglected for decades. But Arthur Kleinman, a Harvard professor of medicine and anthropology, declares Death Foretold "a clarion call for physicians to return to the core of our profession" and believes it may well change the way medicine is taught and practiced.

Christakis is no Kevorkian. For him, dying well has to do with knowledge, not needles. But the knowledge that interests him as a sociologist and drives him as a doctor is not the molecular science of most medical schools but something more difficult to measure: the number of days a patient has to live. A handsome, happily married 37-year-old man, he bears a scar about an inch and a half long on the right side of his neck, the result of an operation that fortunately turned out to be about nothing. When he talks about dying--which he does a great deal--Christakis tends to run his finger back and forth along the faint red line. It gives him a pensive air at odds with the concentrated energy that carried him straight from Harvard Medical School and a PhD in sociology from the University of Pennsylvania to dual appointments at the U. of C.

"It's very trendy to talk about 'vulnerable populations,'" he says, anger lacing his normally measured tone. "But to my eye, it's hard to imagine a more vulnerable population than the terminally ill. And yet in our rush not to abandon them therapeutically, we often abandon them prognostically."

Based on dozens of studies conducted by Christakis and other sociologists and physicians, Death Foretold moves from fact to fact about the medical neglect of prognosis like a doctor delivering bad news in well-reasoned tones--almost soothing as he tells you you've got cancer. But the book is more than a description of a condition. It's a manifesto for a form of prognosis that's equal parts prediction--an assessment of likely outcomes based on statistical averages--and prophecy, an intuition of what lies ahead.

Many doctors will view the book as an indictment of current medical practice. Most physicians, Christakis charges, "are so unwilling to see death coming that they don't see death coming." That may be because, like many Americans, they subscribe to a "never surrender" attitude more fitting to a western than a hospital room. "Our model for death is the Dylan Thomas poem 'Rage, rage against the dying of the light,'" notes David B. Morris, author of Illness and Culture in the Postmodern Age. "Which we've taken and turned into 'Go down with your guns blazing.'" Or your IV dripping, your radiation machine blasting--anything to prolong life, no matter how horrible it's become.

Christakis's studies show that most doctors refuse to prognosticate at all, and those who do tend to overestimate their patients' remaining days by two to five times. As a result, patients don't know what's coming and make no preparations. "Americans die deaths they deplore in locations they despise," he says, 80 percent of them in hospitals or nursing homes, as many as 70 percent of them in unnecessary pain. "We care so pathetically poorly for them. Frankly, it's a disgrace."

Frances Holbrook died the way Christakis thinks people should die. At 93, she passed away slowly but with certainty. "It didn't take a rocket scientist to see it coming," he says. Although he missed her last day, he was there not long before, making sure that she felt no pain and that she was as ready as anyone can be for death. "She was at home, where she wanted to be," he says. "She'd made her arrangements. And she wasn't hurting."

The same can't be said for Mary O'Reilly, a south-side resident Christakis decides to check on the same afternoon that he learns of Holbrook's death. Standing in his dining room in a Hyde Park brownstone, he calls Horizon Hospice to ask if O'Reilly needs a visit. She's 68. She has emphysema. There's nothing left for her doctors to do, so she's gone home to die, with Horizon's help.

Christakis last visited O'Reilly about three months earlier (hospice nurses provide the majority of home care). She's been sick for seven years, but in September she fell and fractured her ribs. The injury sent her into a downward spiral. Still, when Christakis saw her last he thought she had several hard but worthwhile months left. Her six children all lived nearby and arranged their schedules to give her constant company. When Christakis talked to her then, she said that only two things bothered her: the burden she was placing on her kids and the fact that she'd smoked for so long. Then she smiled and said she'd love to have a cigarette, even so.

But from the nurses at Horizon Christakis learns that O'Reilly's condition has crumbled in just a matter of days. Thursday she was playing pinochle with her brother, and today she's nearly speechless, her frail frame emaciated from four days of not eating. Emphysema is a hard death. Most patients suffocate, and although you can dose them up with morphine, when their lungs fail to fill with air there's not much you can do for the terror that makes their eyes grow huge. Last October Christakis kept his prognosis vague: 2 to 18 months, he wrote in his notes. Now the time has come to refine his estimate so that O'Reilly's family can prepare.

When Christakis arrives at O'Reilly's home, her daughter Constance answers the door with a hardened face and the darting glance of someone who's desperately trying to stay busy. A bricklayer by trade, she's shouldered the heaviest portion of her mother's care: washing sheets, mixing medicines, cleaning the floors until they shine. Her father, also a bricklayer, built this house. A decade earlier he died of cancer in the basement room in which Constance's mother now lies, propped up in the same hospital bed.

Mrs. O'Reilly's son Justin stands beside her. A lawyer, he's dressed like Constance in caretaking gear, sweatpants and a sweatshirt. The two look like instructors on an aerobics video. Behind them, at one remove, stand Mrs. O'Reilly's brother and his wife, who flutter away from the bed like butterflies when Christakis enters. He goes down on one knee and takes Mrs. O'Reilly's right hand while Justin squeezes the left. "Hello, Mrs. O'Reilly," Christakis says, his voice quiet and low. "It's so good to see you. I'm Dr. Christakis. Remember when I was here before?"

Mrs. O'Reilly glances at the stranger who's seized her hand. Her skin is pale yellow and her hair is the white of snow under twilight. Her eyes, enormous and black, fly away from him to the ceiling. "Do you know who I am?" Christakis asks again. Mrs. O'Reilly's eyes alight on his. Her lips stretch tight into an open-mouthed smile of perfect teeth--dentures. And they're bothering her. She releases her son's hand and tries to point. Christakis asks her if she'd like them removed. She can't answer.

"I think she recognizes you, though," says Justin.

"I think she does," Christakis agrees.

Justin explains that just a few days ago her vision was as clear as if she were standing on a mountaintop. But then she fell into a valley. It was a trip she'd taken in the past, but she'd always come out in a day or two. This has been the longest time down he can recall. Four days. "And we don't seem to be leaving the valley," he says. "That's the history of how we came to now."

His concern is not with now but with what's next. He and Constance have been giving their mother a new painkiller they got from the hospice nurse as well as antibiotics for the various infections that afflict her. The latter, he says, are keeping her alive; the former, he worries, may be hastening her toward death.

"Sometimes her hands reach up," Constance interjects. She wants to know if that's a sign of the end.

"What do you do then?" Christakis asks.

"I give her her rosary beads," Constance says. "And she's busy."

"Well then," says Christakis, "I don't think anxiety is the problem."

Nor are her painkillers, he decides after he examines her. He can't be sure, but he thinks that Mrs. O'Reilly has simply turned a corner in her slow progression toward death. The remaining distance is the shortest part of the trip. But--how long? Justin wants to know. The family has gathered in the living room, arranged on couches and chairs in a circle around Christakis. "I think," he says, "she's comfortably deteriorating." The family knows what he means. But their deliberately blank looks tell him they want him to be more explicit. "This is a downswing without an upswing," he says.

Then Constance changes the subject: Are chocolate shakes good for her mother? Christakis smiles. Yes, he says. Anything she enjoys is good for her. And are the antibiotics good for her? Justin wants to know. Will they make her live longer? Yes, Christakis answers--to the latter question.

"Have you seen many cases of emphysema?" Justin says.

"I have," says Christakis. "It's not easy. But--she could also die of a lung infection."

Justin follows his point. "How long," he says, "how long without the antibiotics?"

"I think a month. A month or less."

"That's, that's good to know," says Justin. "A month or so."

"I think," says Christakis, holding Justin's eye, "less."

If she improves, Christakis adds, she might live longer. If they treat her infections, he spells it out, she could live to die of emphysema. Without the antibiotics, the infection might spread. From her lungs to her blood. A condition called sepsis. "She would get very sleepy," he says. "A little delirious. Her lips would be dry. And she would die."

"But if it was emphysema...," says Constance.

"She'll suffocate. But I don't see the look of fear in her eyes right now."

"Oh," says Constance, "it's been there."

"The decision...," says Justin.

"Is whether to continue the antibiotics," answers Christakis.

"We don't want, you know...," says Justin.

"I know," says Christakis.

"For her to die...," says Constance.

Then Christakis says: "Hungry for air."

Mary O'Reilly won't die gasping. Based on Christakis's assessment, Justin, Constance, and their siblings take their mother off antibiotics. A few days later, she passes away in her sleep.

"Look," says Christakis one afternoon, sitting in a conference room beside his office and chain-sipping Diet Coke. "I'm not talking about truth dumping, or terminal candor, or ramming the horrible truth down the patient's throat. I'm talking about providing information that many patients want and ask for."

In some cases, he admits, patients simply don't want to know. But too often his hospice work takes him into homes like Mary O'Reilly's, where the patient hasn't been fully informed of what may lie before her. Worse, he's visited hospice patients who haven't even been told that a terminal prognosis is a requisite for the home care they're receiving. In such cases, he insists, doctors have not only a professional responsibility but a moral duty to be more forthright.

The relationship between a patient and a doctor, he says, resembles the relationship between a supplicant and a seer--more so than most scientific-minded doctors would like to admit. Both depend on a ritual of communication. You go to a doctor. Someone who's studied secrets you can't imagine. Who's been initiated into an almost priestly order. You grant this person authority to tell you things about yourself. This doctor wears a white coat, a symbol of his knowledge; maybe he's even got a stethoscope around his neck, the better to know your heart with. You allow him to examine you. A prophet might look into your eyes or count the lines on your palm. Your doctor takes a picture of your bones, peeks beneath your skin. You give him knowledge of yourself, and he gives knowledge back to you. He speaks a language you can't understand, but he translates. He tells you what's happened to you, what sort of disease you've been stricken with: the past. And he discusses the treatments he'll attempt to save you with: the present. And, maybe, just maybe, he'll try to predict the future.

Or not. One study Christakis conducted suggests that fewer than 20 percent of doctors offer prognostic information. Most confess they know little or nothing about prognostication--and prefer not to share what they do know with their patients for fear of robbing them of hope. But doctors need to broaden their definition of hope, says Christakis. "They need to understand the hope for a good death."

At some buried level, though, he thinks most of them already do. Many of the doctors he spoke with in his surveys on prognosis brought up their religious beliefs. The way they insisted on connecting theology and healing no matter what their point of view, Christakis writes, reflects the physician's naturally prophetic role--"a role that seeks to find grace in suffering, order in disorder, and meaning in disease." Not surprisingly, they often try to shrug off such a difficult task. Christakis's interviews with his colleagues turned up time and again an unwritten code of conduct: don't foresee, and if you do, don't foretell. And if you absolutely have to, keep your prognosis upbeat and vague.

Some doctors attributed their reticence to overwhelming uncertainty. "If physicians were that good at prognosis, then we'd spend more time at the track," one internist told him. Others spoke of the power they felt when they approached their patients with an arsenal of therapy and blinders of optimism. One doctor told Christakis that the practice of medicine "is like wine, and prognosis is like the dregs."

In fact, many doctors and scholars see prognosis as oftentimes morally wrong. "One of the healing tools is trust," says Joan Cassell, a Washington University anthropologist who's written several ethnographies of surgeons. "Why would you ever want to take it away?" She adds, "Prognosis for an individual patient is almost impossible, if you're really going to be honest about it."

Even doctors who make prognosis a prominent part of their practice express doubts. Michael D. Lockshin, a New York rheumatologist who deals with terminal illness and has written a book called Guarded Prognosis: A Doctor and His Patients Talk About Chronic Disease and How to Cope With It, is quick to agree with Christakis's point that contemporary medical education leaves doctors in the dark when it comes to seeing into the future. Many have a hard time just seeing a whole person. "We're taught the molecular-biological understanding of disease," he says. As technology advances, he adds, that particular focus may grow even more narrow--and the practitioner more arrogant. "There's an assumption that if we know the genes, we know the disease," he says. That leads some doctors to a "fix it" mentality that allows them to ignore the human being. Lockshin says he'll be delighted if physicians take Christakis's call for change seriously, but he worries that some doctors may respond with another quick fix: a careless approach to prognosis based on statistics. That would only compound the problem: "Anytime you use an aggregate number, an index of something, you're hiding information."

Howard Spiro, a medical scholar at Yale, is less skeptical than Lockshin about the ability of physicians to foresee accurately but more doubtful about the benefit of passing on bad news. In his book The Power of Hope: A Doctor's Perspective, he admits that his opinion that doctors should sometimes keep the truth to themselves falls "under the rubric of what used to be called paternalism, and is now called parentalism," adding that his ideas "are not thought to be correct." But he insists that his reluctance to name the number of days remaining to a patient is a form of humility. "Everybody knows no living man has the power to predict exactly who's going to die and when," he says--though in his own practice he's usually been right about when a patient's days were coming to a close. But even when the future runs along a predictable track, he believes that "hope makes logical sense."

Spiro believes doctors should "keep hope alive" even when they know their patients will die. For an earlier book he edited, When Doctors Get Sick, he examined the ways in which his own colleagues reacted to life-threatening illnesses. "They didn't want terminal candor, they wanted mercy," he recalls. And if medical professionals don't want honest appraisals, wouldn't it be brutal to force such information on laypeople even less prepared to understand?

Spiro cites many studies suggesting that the will to live may occasionally lead to miracles, perhaps a few extra weeks or an unexpected recovery. He's not sure he buys that line, but he thinks it's worth considering. Moreover, he argues, a good death may sometimes be one in which the patient doesn't realize what's happening, in which he or she goes down believing in life continuing to the very end. That may mean missing the opportunity to die at home, but Spiro thinks doctors should consider the benefits of dying in a hospital: clean sheets, plenty of professionals, a burden lifted from the family. Of course, he adds, it's ultimately the patient's decision, but in preparing the patient he'll usually err on the side of optimism. "The art of medicine is in phrasing," he says. "A good prognosis is always better, within the limits of what one can tell a patient without lying."

The art may well be in the way you tell the story, agrees Harvard med school professor Kleinman, whose book The Illness Narratives: Suffering, Healing, and the Human Condition offers case studies of chronic pain sufferers. But it's a poor storyteller, he says, who denies the end of his own tale. As Christakis points out, doctors have drifted away from prognosis as their ability to treat disease has grown. The quest for magic-bullet cures has become an obsession. And in Kleinman's view, the power to simply enter a room, lift a finger, and make the lame walk or the blind see has simultaneously thrilled and terrified doctors. "You see an enormous amount of belief in self-fulfilling prophecy among physicians," he says. More doctors than would care to admit it, he charges, believe that their words are as powerful as their medicines. They fear that their predictions of death may become death sentences. "Maybe a little on psychosomatic grounds. Mostly on a belief system of rationality tied to, for lack of a better word, magic."

Not that Kleinman thinks doctors should forgo their hunches. Divide the world neatly into two categories, the known and the ignored, and you'll cordon yourself off from your patient. "Doctoring is about being a part of the illness experience," he says, "of seeking to understand not only what is happening to the patient but also of seeking to understand along with the patient."

Some scholars might call that view postmodern relativism or, even worse, New Age nonsense. But Christakis is quick to point out that Kleinman's ideas--on which he drew heavily for Death Foretold--are both new and old. Two hundred years ago, says Christakis, doctors couldn't really treat disease but didn't know that--and as a result would bleed a person to death as a great way to get rid of a nasty cold. One hundred years ago doctors still couldn't do much to treat disease, but medical knowledge had advanced to the point that they were aware of that fact. So they spent a lot of time thinking about such nebulous factors as the character of the patient--time well spent, Christakis says. Even with more limited resources those doctors might have been defter at prognosis, which in many cases was all they could offer.

But around the turn of the century, Christakis argues, doctors got the idea that diseases rather than people had characters. Soon, instead of seeing a patient, they saw a disease with some skin and bones around it. As impersonal as that approach was, it led to amazing steps forward in medicine. It also led doctors away from prognosis, away from considering the future, and toward the hard data of the present. Or, in Christakis's words, "a clinical view that looks through rather than upon the individual case."

In that one respect Christakis argues for a return to the past. Of course he advocates pushing technology as far as it can go, and as a sociologist he has great faith in the power of surveys and numbers to provide probable outcomes. Among his suggestions for improving prognosis are putting more money into statistical data banks and computer programs that can double-check doctors' intuitive predictions. But when it comes to foreseeing the future of a person rather than a disease, he's convinced that physicians must look to the art of medicine as well as its science; the constitution of their patients as well as the information on their computers; and the vague, even frightening terrain ahead as well as the knowable details of disease.

"We have to search for the meanings of diseases as well as their manifestations," he says. "Otherwise, healing is as random as suffering."

Christakis will concede that randomness may be the order of things, but he remains certain that prognosis can provide shape to the past as well as the future. In 1969, when Christakis was six years old, his mother was diagnosed with Hodgkin's disease. "She was, you know, going to die," he says, recalling that the family was told that his mother would almost certainly not live beyond three weeks. But she did. Then she lived another 3 weeks, and then 6, and then 12, until weeks gave way to years. His mother lived nearly two decades longer than she was expected to.

Christakis admires the courage of the doctor who gave the original prognosis, but its inaccuracy frightened him. "I grew up both craving and detesting prognostic precision. Wanting desperately to know what the future held, but also not to know."

In his mind, his mother's good fortune is in some ways the exception that proves the rule. "I know how hard prognosis is," he says, admitting that he's delighted when he's wrong. "But I also know we can do better." His mother's case notwithstanding, he believes that the systematic pattern of optimism in prognostication proves that there are systematic means to give more accurate predictions. The place to start is medical school. "I believe this is a perfectible science," he says. And a perfectible art.

"Prognosis is what gives meaning and emotion to the diagnosis," he says. Most patients don't care all that much about the pathologic details--whether they have metaplasia or neoplasia or anaplasia. They want to know "What does this mean for me?" They want their doctors to be optimistic but not at the cost of meaning. Nobody, he says, wants a doctor who's a nihilist.

Art accompanying story in printed newspaper (not available in this archive): photographs/Lloyd DeGrane.

Add a comment