The tragedies I encountered throughout my posting were interwoven with Lillian's illness. She was not only depressed or manic but frequently psychotic: One morning I found her stranded, unable to figure out how to step across the hot lava that she saw flowing around her bed. My fears soared when I read that one out of every three cases like hers ends in suicide. So, late in my Vienna tenure, when our doctors had tried every treatment they knew and she wasn't getting better, I flew home to meet with the leading researcher on bipolar illness at the National Institute of Mental Health. Dr. Robert Post met with me in his home for two hours late one night.

I described Lillian's struggle just to get through a day. "Your daughter's system is being poisoned," he told me, his voice clipped with annoyance. "You shouldn't have accepted the debilitating side effects of her medications." He recommended that we start over with an experimental mood stabilizer that had helped 60 percent of his patients who were nonresponders to other drugs.

With a surge of optimism, I asked, "How many have you given it to?"

"Twelve."

I searched his eyes. "How many children?"

"None."

Dr. Post waited as I processed his words. "Do you think it's safe?" I asked nonsensically.

"Safe? You've tried everything else, and your daughter has a potentially fatal illness. It's time to throw the book out the window," he pushed.