Taking a new approach to psychiatric research

In April, Dr. Thomas Insel, director of the National Institute of Mental Health, announced that the nation's largest funder of psychiatric research would no longer rely on the Diagnostic and Statistical Manual to categorize mental illnesses. The DSM, published by the American Psychiatric Association, had been the bible of psychiatric diagnosis for 60 years, and is still used by clinicians and insurers. Insel rejected the DSM's use in research in favor of the Research Domain Criteria (RDoC), a system based not only on patients' symptoms, as the DSM is, but also, when possible, on their brain anatomy and chemistry, and genes.

Insel's announcement was nothing short of cataclysmic, stirring a heated international discussion about the validity of psychiatric diagnoses and the relative contributions of life experience vs. brain pathology in mental illness.

Q. Can you talk about the decision to adopt RDoC and about how you think psychiatry will change as we learn more about the biology of mental illness?

A. The reason I went into psychiatry is that I was interested in patients' stories. In some ways the most boring patient is still a lot more interesting than our most interesting science. By that I mean, it's the narrative that's most compelling and most complex. And the science that we have to try to understand that narrative is incredibly oversimplified.

The RDoC approach is not going to be any better than anything else at explaining holistically what somebody experiences. It might be that if you had a patient's genome sequenced and you had his or her brain scanned, it wouldn't change anything that you'd do. We're really doing this partly to find that out.

Q. Do you think the DSM, which diagnoses patients based on their symptoms, is too subjective?

A. It's too narrow. It's too unidimensional. If cardiologists couldn't use an EKG or an echocardiogram, couldn't use anything except a patient's description of his or her chest pain, I wouldn't call that so much rigid as limited.

Q. Does the way psychiatrists currently diagnose mental illness limit patients' care?

A. I don't know that it limits care. It limits our understanding. The argument I'm making is that giving someone a psychiatric label tells you 2 percent of what you need to know.

Within a label like "schizophrenia" it's possible that there are 10 different disorders. My major complaint with psychiatry today is that it hasn't expanded to take in lots of kinds of information.

Q. Couldn't it be argued that, though limited, the DSM does offer a common language in which clinicians and researchers can communicate about mental illness?

A. I actually agree with that. I trained in the late '70s, and it was very chaotic. We would get reports about the treatment of schizophrenia in the United States and the treatment of manic depressive illness in England and they were talking about people with basically the same symptoms. It was impossible to read the literature and know what it meant because there was no common definition for any of the disorders. So to the extent that the DSM is a dictionary, it's extremely helpful.

What it can't do, though, is provide a guide to research. Researchers need to deconstruct these categories so that we can understand all the different disorders that probably underlie autism or schizophrenia or depression. And if you're convinced that everything you discover has to match those DSM criteria, you'll never be able to make progress.

Q. But is it good for clinicians and researchers in the same field to be speaking different languages: DSM and RDoC?

A. I don't think that's what we're doing. The DSM is the best we can do with what we know now. But we need to come up with something quite different for research. And to be fair, we haven't really had the tools to do that until very recently — functional MRI and so forth — and I'm not sure that we have the tools to do it completely even now. But we ought to make that the priority: to do for people with mental disorders what we do for people with other brain disorders.

Basically what happens in psychiatry now is that providers are taught to memorize all the DSM criteria, and then they go out like they're bird-watchers and they find what the diagnosis is. The patient gets a label, and then a set of treatments. I just think we can do better than that.

Q. Hasn't there always been disagreement about how or even whether various forms of mental suffering should be labeled as a disease?

A. There's always been some tension in psychiatry about being medical. Freud himself was apoplectic about the idea that psychoanalysis was going to be restricted to physicians [in the United States]. He was not a big proponent of diagnosis or any kind of medical model. And maybe I stand in the middle, which is that if we're going to make progress here — and I don't think we have made progress to the extent that we should have — we need to bring in all of the tools that we have in modern science and modern medicine to try to help people with serious mental illness.

Q. Fifty years from now, can you imagine that clinicians will have information about patients' brains and genes that will allow them to prescribe drugs targeting mental health problems more specifically than they do today?

A. I don't know if it will be drugs. It may be some other kind of intervention. I'm really interested these days in cognitive interventions.

Q. Do you think that psychotherapy will still be used 50 years from now?

A. It might be. But it might look very different. It might be that what we call "psychotherapy" is a mobile app that you download and is crafted for your specific cognitive domain. It would be really fun to think about this.

But what I'd really love to think about is that, in the future, a patient wouldn't become sick. We don't wait for people to have heart attacks to treat them. We treat their high cholesterol, look at their family histories, do everything we can to make sure they don't have heart attacks. That would be my vision for psychiatry: preemptive mental health care.