Can a sugar pill replace your Zoloft?

Media outlets were abuzz early this year with the news that antidepressant drugs don't get rid of depression. The source was a Jan. 6, 2010, report in the prestigious Journal of the American Medical Association analyzing data from six clinical trials of antidepressants conducted between 1980 and 2009, involving 718 patients. The study concluded that when people with mild or even moderate depression took a pill, their symptoms improved. The kicker: The improvement was the same, on average, whether the pill was a real drug or a lookalike placebo. Only among people with more severe symptoms did the real drugs have an edge over the placebo.

Studies like those make great headlines. Far more complicated is how to apply the findings in real life. You can't diagnose or track depression with a blood test or an X-ray. Assessing improvement relies completely on the patient's subjective reporting of his or her state of mind. Plenty of studies have shown that the drugs work, though placebo effects are notoriously high in antidepressant trials, and the drugs' advantage over placebos is usually modest.

Then there's the fact that placebos can and do work in and of themselves. "It's always been known that people with milder depression are more likely to respond to placebo than people with more severe depression," says Michael First, M.D., a professor of clinical psychiatry at Columbia University. That may result from the power of suggestion, the reassurance of care from an empathetic therapist, or the confidence that comes from taking action to get better.

Depression also tends to improve over time naturally, so the placebo improvement might be, in part, just the illness running its course. "But practically speaking, how do you get around that?" First says. It's unethical for doctors to give patients placebos without telling them, unless they've agreed to be part of an approved clinical trial. But for a placebo to work, the doctor has to lie; research has documented that once people know they're getting the placebo, the effect goes away. As effective as the placebo response can be, "there's an inherent impossibility in real life to harness it," First says.

If you're not sure you need an antidepressant any more, talk with your doctor about reducing the dose or weaning off of it. Never stop an antidepressant suddenly or on your own; you may be risking severe withdrawal symptoms.