"I've sat in many staff meetings where we've decided that patients should start lithium," she told me. "Occasionally, one of the doctors will say, 'Yeah, but it can cause weight gain.' 'So what's a little weight gain?' another doctor will answer. 'The patient can exercise.' "

"We just ignore it as a side effect. But then, when I was diagnosed with bipolar disorder, and started lithium, I gained 45 pounds. And my skin got bad. I used to be thin and attractive. Now look at me!" Tears welled in her eyes.

Her loose-fitting clothes fell in rumpled folds. Her cheeks puffed out and her skin looked rough from acne. Her calves bulged below her skirt.

She was a psychiatrist. Two years before, just after finishing residency, she received a diagnosis of bipolar disorder, or manic depression.

Her symptoms, before medication, included irritability, too much energy and insomnia. Lithium helped, but like many drugs for mental disorders it can cause weight gain.

Many of the chemicals in the brain that bring on psychiatric problems also regulate appetite. These drugs can thus increase hunger and thirst. They also alter metabolism: users feel tired and exercise becomes more difficult.

"Now, I have trouble getting out of bed," she continued. "I drive home and just want to nap. I used to have a great body and jogged five miles four times a week. But the worst part is, I can't get a date. Doctors don't think that's a big issue, but it is.

"Most doctors at these staff meetings are men. What do they know about what weight gain means to a woman? Before I got fat, I never even thought about it myself."

Part of the difficulty is that weight gain is not, strictly speaking, a medical disorder. But more important, many doctors weigh risks and potential benefits of treatments in ways different from their patients without realizing that wide contrasts exist.

Risks, after all, are relative: What one person considers too dangerous, another might not. The way risks are presented and framed shapes our perceptions of them.

A patient once told me: "The night before my open heart operation, my surgeon told me I had a 5 percent chance I may die. I couldn't sleep all night." This patient, if instead told that he had a 95 percent chance of thriving after the surgery, would have slept much better.

In research, too, investigators are supposed to warn participants of possible dangers. Yet at times, they minimize such hazards and promote only the benefits.

According to research, humans do not always think rationally about risks, but instead rely on stereotypes, overestimate the likelihood of bad outcomes, underestimate the possibility of good results and think they see patterns where none exist.

But this psychiatrist was telling me something else, too: that doctors and patients view risks and benefits in drastically different manners.

I wondered what other problems doctors and patients saw in contrasting lights. I thought of patients who had complained of going bald from chemotherapy; having to lie on a stretcher in a chilly hospital corridor for hours waiting for tests; having to wait an hour or more to see a doctor.

With a side effect, too, it is one thing to say the odds of its occurrence are 30 percent. But the importance or unimportance of that symptom may range widely between people in ways that doctors do not take into account. Doctors often have trouble dealing with the inadvertent side effects of their own treatments.

Whereas this woman's psychiatrist assumed the benefits of reducing signs of mania far outpaced the risks of weight gain, the woman calculated the ratio differently.

She knitted her black eyebrows together. "I also can't talk about these issues to colleagues," she said, sighing. "I can just imagine what they would say if they knew that I was bipolar. But I wish there was some way they could wake up and see."

Slowly, she shook her head to herself. "Colleagues tell me, 'Watch what you're eating, lay off the doughnuts, join a gym.'" She reflected on these discrepancies and laughed. "If they only knew."

Dr. Robert Klitzman is a co-director of the Center for Bioethics at Columbia University.