Dr. Frances repeatedly chastises the writers of the DSM-5 for failing to ask Hippocrates’ question: Will this new diagnosis help patients or harm them? But the DSM-IV failed to ask the same question when, for example, it retained Multiple Personality Disorder. After the DSM-III included MPD in 1980, thousands of spurious cases emerged in the next two decades, and special psychiatric clinics arose to treat them. Yet faced with evidence of this disastrous epidemic, the DSM-IV did not delete the diagnosis. Instead, the manual renamed it Dissociative Identity Disorder. “MPD presented a dilemma for me,” Dr. Frances says. “We took scrupulous pains to present both sides of the controversy as fairly and effectively as possible—even though I believed one side was complete bunk.” How do you “fairly” argue for a diagnosis you think is complete bunk? Where’s the methodological rigor? Why did it take malpractice suits to close the psychiatric MPD clinics and not the presumed voice of scientific authority, the DSM? Dissociative Identity Disorder remains in the DSM-5.

Nonetheless, Dr. Frances does his best to save psychiatry. It is impossible to define “normal,” he explains, let alone “mental disorder.” But that doesn’t mean, he adds, that we can’t talk about the problems that cause human suffering. Mental disorders should be diagnosed only when a person’s symptoms are obvious, severe and haven’t gone away on their own. Watchful waiting should be the first step in treatment, and medication a last resort. A wise, and increasingly rare, approach.

Is there another way to classify mental problems? The National Institute of Mental Health has washed its hands of the DSM and is working on its own system for 2020. As NIMH director Thomas Insel pronounced: “Mental disorders are biological disorders involving brain circuits that implicate specific domains of cognition, emotion, or behavior.” A NIMH initiative plans to produce “precision medicine” to treat them.