Why doctors are so bad at predicting how long their patients will live.

Why doctors are so bad at predicting how long their patients will live.

Why doctors are so bad at predicting how long their patients will live.

Health and medicine explained.
Aug. 18 2010 11:37 AM

The Worst Fortune Tellers

Why doctors are so bad at predicting how long their patients will live.

(Continued from Page 1)

Such technology can complicate predictions further: Take former Vice President  Dick Cheney, who last month had a left ventricular assist device implanted in his chest to augment the pumping of his weakened heart. By conventional measures, given the severity of his reported  congestive heart failure, the likelihood of his surviving past one year would be only 50 percent. But don't yet file his obituary: Doctors not involved with his treatment have given optimistic projections that carefully selected patients with LVADs can live more than an additional year with the device; one patient survived more than five.

So how do we become better at prognostication?


Christakis argues that studying and delivering prognoses to patients is part of the ethical obligation of doctors to their patients. "Furthermore," he writes, "physicians should legitimate discussions regarding prognosis not only with their patients but with each other." As such, doctors would recast the professional norm to include open and frank discussion of prognosis in medical care.

In so doing, we need to strive for honesty and avoid "hanging crepe," the idea of delivering a poor prognosis simply to combat our tendency to be overly optimistic and to keep our hands clean: If the patient dies, I predicted it and therefore appear accurate; if the patient outlives my prediction, everyone is pleasantly surprised and thus I'm not held accountable. Dahut reminded me of the oncologist's rule of thumb: If a patient was walking around three months ago and still can, he'll probably be around a little while longer; if he can't get out of bed, "the disease is likely to progress at the same rate." Framing it this way helps patients and families manage expectations, by not giving a limited time interval that can feel like either an immutable sentence or an obstacle to overcome.

Beyond that, there needs to be fundamental change in the way that doctors are educated in prognosis calculation and delivery. Prognosis should not be left to the realm of mordant comments in pathology lectures or sotto voce remarks to students and residents outside a patient's room during ward rounds. Instead, medical education at all levels should feature actuarial information on the major causes of death, with modification for the age of the patient and the various treatment options. Patients will have to more overtly demand prognosis to help spur these changes.

Of course, in an age of patient empowerment, you may well be the ones to lead the charge to make doctors more attuned to answering that fundamental question, which lies at the very heart of medical practice. We can only hope that future doctors will be better prognosticators than old Dr. Sikora.

Like Slate on Facebook. Follow us on Twitter.