Begley: Get Shrunk at Your Own Risk

No one bats an eye when a drug for a severe mental illness such as schizophrenia or depression causes serious side effects such as nausea, weight gain, blurred vision or a vanishing libido. But what few patients seeking psychotherapy know is that talking can be dangerous, too—and therapists have not exactly rushed to tell them so.

For treatments that come in a bottle, the Food and Drug Administration requires proof of safety and efficacy. For treatments that come from the lips of psychologists and psychiatrists, there's no such requirement. But while therapists fight over whether they should use only treatments for which there is rigorous scientific evidence for efficacy, they have largely ignored something more fundamental. "The profession hasn't shown much interest in the problem of treatments that can be harmful," says psychology professor Scott Lilienfeld of Emory University. "Of the few psychotherapies that have been tested for safety, too many cause harm to at least some patients."

The failure to heed Hippocrates reflects the assumption that psychotherapy is, at worst, innocuous. That naive trust should have been blown out of the water when "recovered memory" therapy actually created false memories, often of childhood sexual abuse, tearing families apart. But the "Handbook of Psychotherapy and Behavior Therapy," the clinicians' bible, devotes only 2.5 pages out of 821 to adverse effects, even though documented risks of therapies could fill a small book.

"Stress debriefing," for instance, is designed to prevent symptoms of posttraumatic stress disorder in those who have suffered or witnessed a trauma. In a three- to four-hour group session, a therapist pushes patients to discuss and "process" their feelings and to describe in detail what they experienced or witnessed. Many of those who undergo stress debriefing develop worse PTSD symptoms than those who deal with the trauma on their own, controlled studies show, probably because the intense reliving of the trauma impedes natural recovery. Burn victims who underwent stress debriefing, for instance, had worse PTSD 13 months later than victims who had no psychotherapy; people who went through it after being in a car crash had greater anxiety about travel three years later than those who did not.

Psychotherapy for dissociative-identity disorder (formerly called multiple-personality disorder) can pose even greater risks. Some therapists believe that the best treatment for these fractured souls is to bring out the hidden identities, called "alters," through hypnosis or helping alters leave messages for one another. Unfortunately, many alters cause "self-injurious behavior, suicide attempts, and verbal and physical aggression," notes Lilienfeld in a paper in the journal Perspectives on Psychological Science. In addition, the "let's meet the alters!" techniques can actually create alters in suggestible patients. "As more alters come out, it gets harder to get the patient back to having one identity," Lilienfeld says. The longer someone stays in therapy, the more alters show up, evidence that "many and perhaps most alters are products of inadvertent therapist suggestion." So much for "First, do no harm."

Few of us will need therapy for multiple-personality disorder. But everyone will experience grief—and counseling for normal bereavement may not always be benign. A 2000 study found that four in 10 people who lost a loved one would have been better off without grief counseling (based on a comparison with people who were randomly assigned to a no-therapy group). That was especially so for those who experienced normal grief. In that case, counseling sometimes prolonged and deepened grief, leaving more depression and anxiety than in those who worked through their loss on their own.

That 40 percent figure is likely inflated, argues psychologist Dale Larson of Santa Clara University. But he agrees with Lilienfeld's estimate that 10 to 20 percent of people who receive psychotherapy are harmed by it. Even the American Psychological Association acknowledges that too many clinicians practice "psychoquackery," as psychologist John Norcross of Scranton University puts it. If we had FDA-style regulation of psychotherapies—difficult though that would be to do, especially since the effects of psychotherapy depend on the therapist—"fringe therapies would not be on the market."

How fringe is "fringe"? In percentage terms, very. But the number of people undergoing potentially risky therapies reaches into the tens of thousands. Vioxx was yanked from the market for less. To be sure, even risky psychotherapies don't harm everyone, just as most people who took Vioxx will never have a heart attack. What is remarkable about psychotherapies, though, is that few patients have any idea that "just talking" can be dangerous to their mental health.