Truth in the Most Optimistic Way FREE

While I was in residency at the Mayo Clinic, my oncology attending arrived for rounds one day with a sheepish grin on his face. It was June 3rd, you see, and for the 7th straight year a former patient of his had called him and angrily exclaimed, “I am still alive, you idiot!” before slamming down the telephone. Seven and a half years ago, the oncologist had given the man 6 months to live.

Since then, I have stopped giving patients specific predictions about their life expectancies. I recognize that patients need to know their prognoses to make treatment decisions and plan their affairs. However, I have found that relatively nonspecific prognoses are sufficient. I might say, “I cannot predict the future, but in my experience, patients with your illness typically live a matter of months, not years,” or “Many people in your condition will live for only a matter of weeks, but some live significantly longer. I do not know what your fate will be.” In these conversations, I discuss concrete treatment goals with patients. I do not hesitate to say when I think the goal should shift from cure to palliation. When things are grim, I suggest that it is time to visit with friends and family because “it is better to be safe than sorry.” I give enough prognostic information to help patients make decisions, but I avoid using numerical wording that suggests I have a prognostic crystal ball.

In this issue, Lamont and Christakis (1) report the results of a survey of physicians who had referred patients with cancer to local hospices. They asked each physician to estimate how long his or her patient would live. In a later question, they asked physicians what prognostic information they would communicate if patients insisted on receiving such information. They found that almost one quarter of physicians would not communicate a temporally specific prognosis, 37% would communicate the same prognosis that they had estimated, 28% would provide an optimistic prognosis (a longer survival than predicted), and a small number would provide a pessimistic prognosis.

I do not know where my practice style would fit in this classification scheme. On the basis of conversations I have had with other physicians since reading Lamont and Christakis's article, my communication style is relatively common. The physicians I spoke with, an admittedly unscientific sample, said that they were reluctant to provide specific predictions to patients. Their reluctance is based not on a desire to withhold information from patients but on uncertainty about their predicting abilities. When I mentioned that the physicians in Lamont and Christakis's study were a median of 70% confident in their prognoses, most of my colleagues replied that even if they were 70% confident (whatever that would mean), they would still be too uncertain to provide specific numerical prognoses to patients.

What prognostic information should physicians communicate to patients with terminal illnesses? When I discussed Lamont and Christakis's research with a taxi driver, he had a simple prescription for how he would want physicians to communicate with him if he were to develop a terminal illness: “Tell me the truth in the most optimistic way.”

But what is the truth here? The disparity that Lamont and Christakis found between physicians' formulated prognoses and communicated prognoses suggests that some of these physicians were being less than forthcoming by, for example, refusing to provide specific prognoses. Lamont and Christakis point out that such refusals were especially common among older physicians, who had trained in an era when patients were often left in the dark not only about their prognoses but also about their diagnoses (2). Perhaps these physicians represent the last vestige of paternalism.

Yet I am not so sure that physicians who refused to provide specific prognoses were withholding information from their patients. To encourage physicians to formulate prognoses, Lamont and Christakis asked the following: “For the next question, please again assume the patient has the most typical course and that the patient receives the type of care after referral to the hospice that you expect [their emphasis]. We recognize that it is very hard to make such predictions, but we would be grateful for your best estimate of how long you think that this patient has to live.”

This wording encourages physicians to take a stab at formulating a prognosis for a typical patient who receives a specific kind of care. Given these assumptions, physicians may have been able to formulate prognoses. But how confident were they that their specific patients would follow this course and receive this care? Physicians' reluctance to give specific predictions may result from their awareness that not all patients are typical.

What about the physicians who communicated optimistic prognoses to their patients? Were they being less than forthcoming? Some of these physicians may have stretched the truth beyond its optimistic limits, but such stretching was probably well intentioned. Patients want to hear that there is hope. And, just as important, physicians want to communicate hope to their patients; communicating hope can improve patients' prognoses. Mildly but unrealistically positive beliefs can improve outcomes in patients with chronic or terminal diseases (3). For example, a late-1980s study of patients with AIDS found that those who stated that they “refused to believe that this problem has happened” lived 9 months longer than those who indicated that they “tried to accept what might happen” (4). Moreover, unrealistically optimistic views have been shown to improve quality of life (5–7). Providing optimistic prognoses might even improve physicians' spirits. This may be especially important for oncologists, who regularly deal with dying patients and may therefore need to find ways to hope that patients will do better than expected.

Perhaps the most puzzling group of physicians is the small number who would have chosen to provide pessimistic prognoses. Were they less than forthcoming? Drawing on my clinical experience, I would guess that many physicians base their communication on the way patients react to their prognoses. When patients are “too pessimistic” about their illnesses, physicians emphasize hope in order to lift their spirits. When patients are “too optimistic,” physicians make sure patients understand the gravity of their situations. The small group of physicians in this study who said they would provide pessimistic prognoses probably did so because of how their specific patients had interpreted their illnesses so far.

Before judging the physicians in Lamont and Christakis's study, we should remember that all had recently referred their patients to local hospices. Thus, it is likely that the major goals of prognostic communication had already been achieved. Since the patients were receiving hospice care, they knew that they had terminal illnesses, had chosen to abandon any attempts at miraculous cures, and had elected a course of palliative care. Such decisions would have been unlikely if physicians had been systematically misleading them about prognosis.

Lamont and Christakis's study deserves widespread attention and discussion. It also deserves to be followed with further research that clarifies physicians' reasons for communicating or not communicating specific prognostic information to terminally ill patients. Prognostication will never be an exact science. The prognostic information we communicate to patients should be vague enough to include the truth—“usually weeks or months”—and specific enough to help people plan their lives and deaths. Numerically specific prognostic communication can be the enemy of hope. The truth we communicate to patients should help them prepare for the worst while allowing them to hope for the best.

Peter A. Ubel, MD

Ann Arbor Veterans Affairs Medical Center; University of Michigan Health System Ann Arbor, MI 48109-0429