Antidepressants Useless? Far From It. Check the Study
John Hopkins Psychiatry Newsletter
by Marjorie Centofanti
July 10, 2010
Last winter, a study in JAMA became a media wonder in part because of its message but also because of its ease of fitting into a sound bite: Antidepressants are no better than a placebo unless your depression is severe. As a result, “patient response has been intense,” says psychiatrist Glenn Treisman. “Every week they tell me I read that these things don’t work.”
Treisman is one of many who questions that study—and others with a similar theme. They’re suspect in the way they interpret data, he says. “The problem hinges on ignoring something researchers learn in Epidemiology 101. And we do that,” he adds, “at our peril.”
A full professor in psychiatry and internal medicine, Treisman is best known internationally for his care of psychiatrically ill HIV-infected patients. His work with this population from early in the AIDS epidemic has translated into expertise in depression, addiction, personality disorders and chronic pain.
It’s his insight on the workings of psychopharmacology—he has a Ph.D. in the field—that we tap here:
Q. These studies are dangerous, you say.
A. Ten to 20 percent of people with major depression died by suicide before widespread antidepressant use. Depression doubles your risk for cardiovascular death after a heart attack, stroke, HIV. I’d say discouraging people from taking good medication for depression is dangerous.
Q. So what’s the problem with the studies?
A. Deep down, it’s failing to recognize that there’s bias in the way we select people for drug trials. In early studies of antidepressants, we were very exclusive. We’d use what’s called research diagnostic criteria (RDC) that made sure someone had major depression before entering a trial. It’s a bit like a surgeon being totally convinced someone has appendicitis before doing surgery. And in those settings, antidepressants are good: 75 percent of people get better.
But because major depression can be so deadly, and because we have no unequivocal test for it, we err on the side of overdiagnosis and treatment. We’ve become inclusive. The clinical criteria in our DSM [Diagnostic and Statistical Manual] for example, aim not to miss anyone who might have depression. But using inclusive DSM criteria for our trials, as we do today, means we include a lot of people with depressive symptoms who don’t have major depression. It’s just as, in the 1950s, some people had surgeries without having appendicitis when their surgeons weren’t sure.
Q. So?
A. What happens is that more people—the ones without the true depression the drug’s intended for—report feeling better. Between 1980 and 2000, for example, the placebo response in antidepressant trials jumped from 20 percent to 40 percent. And when you allow for a margin of error, it starts to look like there’s no difference between the drug and placebo. But that’s wrong. The antidepressants work for true depression, even if it’s mild.
Q. And your confidence comes from…
A. A critical review of new and older literature. My years as a psychiatrist. But if you doubt me, think of the evidence pharma companies have to present to the FDA. By the time firms get to the point of getting a new drug registered, they have tons of phase I and II trials and basic science showing these drugs work. They know they work. Registration trials are too costly for mights and maybes.
Q. You’re eager for a last word.
A. The take-home message is that doctors who prescribe need to be critical thinkers. Each patient is entitled to expert tailored treatment, not a recipe derived from the average response from the average clinical trial subject.
John Hopkins Psychiatry Newsletter
by Marjorie Centofanti
July 10, 2010
Last winter, a study in JAMA became a media wonder in part because of its message but also because of its ease of fitting into a sound bite: Antidepressants are no better than a placebo unless your depression is severe. As a result, “patient response has been intense,” says psychiatrist Glenn Treisman. “Every week they tell me I read that these things don’t work.”
Treisman is one of many who questions that study—and others with a similar theme. They’re suspect in the way they interpret data, he says. “The problem hinges on ignoring something researchers learn in Epidemiology 101. And we do that,” he adds, “at our peril.”
A full professor in psychiatry and internal medicine, Treisman is best known internationally for his care of psychiatrically ill HIV-infected patients. His work with this population from early in the AIDS epidemic has translated into expertise in depression, addiction, personality disorders and chronic pain.
It’s his insight on the workings of psychopharmacology—he has a Ph.D. in the field—that we tap here:
Q. These studies are dangerous, you say.
A. Ten to 20 percent of people with major depression died by suicide before widespread antidepressant use. Depression doubles your risk for cardiovascular death after a heart attack, stroke, HIV. I’d say discouraging people from taking good medication for depression is dangerous.
Q. So what’s the problem with the studies?
A. Deep down, it’s failing to recognize that there’s bias in the way we select people for drug trials. In early studies of antidepressants, we were very exclusive. We’d use what’s called research diagnostic criteria (RDC) that made sure someone had major depression before entering a trial. It’s a bit like a surgeon being totally convinced someone has appendicitis before doing surgery. And in those settings, antidepressants are good: 75 percent of people get better.
But because major depression can be so deadly, and because we have no unequivocal test for it, we err on the side of overdiagnosis and treatment. We’ve become inclusive. The clinical criteria in our DSM [Diagnostic and Statistical Manual] for example, aim not to miss anyone who might have depression. But using inclusive DSM criteria for our trials, as we do today, means we include a lot of people with depressive symptoms who don’t have major depression. It’s just as, in the 1950s, some people had surgeries without having appendicitis when their surgeons weren’t sure.
Q. So?
A. What happens is that more people—the ones without the true depression the drug’s intended for—report feeling better. Between 1980 and 2000, for example, the placebo response in antidepressant trials jumped from 20 percent to 40 percent. And when you allow for a margin of error, it starts to look like there’s no difference between the drug and placebo. But that’s wrong. The antidepressants work for true depression, even if it’s mild.
Q. And your confidence comes from…
A. A critical review of new and older literature. My years as a psychiatrist. But if you doubt me, think of the evidence pharma companies have to present to the FDA. By the time firms get to the point of getting a new drug registered, they have tons of phase I and II trials and basic science showing these drugs work. They know they work. Registration trials are too costly for mights and maybes.
Q. You’re eager for a last word.
A. The take-home message is that doctors who prescribe need to be critical thinkers. Each patient is entitled to expert tailored treatment, not a recipe derived from the average response from the average clinical trial subject.
