David Shaffer, M.D., is the chief of the Division of Child and Adolescent Psychiatry at Columbia University Medical Center. His work has centered on understanding the cause of depressive illness in youth and and how this factors into the diagnosis of unipolar and bipolar illness in childhood and adolescence. He is the co-author of The Many Faces of Depression in Children and Adolescents. This is the edited transcript of an interview conducted on July 19, 2007.
- Interview Highlights
- How a psychiatrist decides between normal and extreme behavior in a child
- Why he thinks bipolar has become a "fashionable" diagnosis for kids
- The twists and turns in the story of youth suicide and the use of antidepressants
- What's been learned about early onset of depression in kids that informs thinking about early onset of bipolar
How was it that the field of child psychiatry has lagged behind the rest of medicine?
I can't really give you a terribly good explanation except that child psychiatry seemed to be dominated by people who were interested in psychoanalytic concepts or in object relations development concepts. ... Adult psychiatry was based on observations that were made in insane asylums -- that's where all the clinical descriptions derive from -- and children, on the whole, were not admitted to those. That may be one of the historical reasons why you didn't get these very good descriptions of the different child disorders until the 1940s.
So that was sort of the age of darkness. Where are we now in this field?
Well, there's a lot that we don't know. What is different is that we're more aware of what we don't know than the kind of certainty that dominated a lot of professions up until the last 15 or 20 years, when general medicine was dominated by professors and experts who knew the answers. The answers were often intuitive or based on very partial clinical experience. I think that the ability to admit to what we don't know is much more prevalent now, and it's a much healthier state to be in....
The whole area of bipolar, which is what you're asking about, is filled with dispute and uncertainty and argument. So I don't think that there are too many people out there saying that we're sure this is the way it is, at least not among the leaders in the field.
[And what about one of those leaders in the field, Dr. Joseph Biederman?]
I think that he was the first to recognize that there are lots of things that he might have put forward at one point which he no longer subscribes to. There are a whole lot of uncertainties about what are the key symptoms and what really happens to kids that we now call juvenile bipolar when they grow up. I think that he's fairly open-minded about it at this point.
He was very influential because he put forward a hypothesis which attracted a lot of people. He's done that right through his career. What he did was he took the written criteria for attention deficit [hyperactivity] disorder and the written criteria for bipolar disorder and he said, "Hey, a lot of these kids that you call ADHD actually fit the criteria for bipolar disorder."
Initially when he said that, it was because the criteria for mood disorder in children includes something called irritability; that's to say that the kids lose their temper or get upset or get annoyed at quite trivial provocations. He said that maybe all these kids that we call ADHD are just a childhood form of an adult disorder that we call bipolar.
So he put forward that hypothesis, and what allowed that to happen was the faults in the DSM [Diagnostic and Statistical Manual of Mental Disorders] system, which on the one hand have got very specific names for a whole lot of symptoms but on the other hand don't have very rich descriptions of what those symptoms are, so they can be interpreted in lots of different ways. A symptom like irritability occurs in something like 26 different diagnoses, so it probably means very different things in different contexts and to different people.
He put forward that idea, and it was jumped on. Historically that's happened before. There are some disorders which are very rare in children, like schizophrenia and depression -- this is before puberty -- and people have never quite come to terms with that. They've always thought, well, maybe we're just missing it, or maybe because they're children, they're showing the disorder in a different kind of way.
That led to autism initially being called childhood schizophrenia, because the person who first gave it that name said, "These are schizophrenia kids, but, because their brains are not fully developed, they're not showing the same symptoms." With depression, the same arguments were applied.
Much of the challenge in diagnosing mental illness is the issue of whether irritability is sometimes normal behavior in kids and adults. And inattention is also normal in a lot of kids. How do child psychiatrists decide what's normal and what's extreme?
A lot of symptoms that you see in disturbed children could be quite normal if you just looked at them by themselves. How do we decide? Well, you hit on a very important point. There are very few of what are called pathognomonic symptoms. In adulthood there are certain symptoms, and if you see them, you can be pretty sure that there's a diagnosis underneath them. For example, hallucinations or delusional thinking and certain kinds of very repetitive thoughts and so on usually mean a particular kind of diagnosis in adults. In children they don't, because they've got an imperfectly developed thinking and feeling system. Almost every symptom that you can name may occur in normal kids.
So the way that one ends up with, what is a disturbed child or, what's a child with a mental disorder, is if a symptom is interfering with their life -- if, as a whole, the child is not really making the progress either developing socially or developing at school or developing in their relationships or developing in their understanding of what's going on; or if they have a lot of those symptoms; or if the symptoms have lasted much too long. It's more those aspects of a symptom rather than their actually descriptive nature. I think people misunderstand that. That's usually what we take into account.
Has bipolar replaced the ADHD diagnosis as the diagnosis of the moment? ADHD is this very gray area, and there was an explosion in this diagnosis in the 1990s. And it seems that the controversy on what it is has sort of been resolved. ...
Well, I don't think that ADHD has been fully resolved. ADHD is a collection of symptoms, and the reason it's not fully resolved is that nobody really believes that it's due to any one single cause. It's a disorder like anemia that could be produced through many different roots. One of the reasons why people believe that is because it so often occurs along with other conditions. So that makes you wonder a little bit about what its very nature is.
The other thing is ADHD did go through a period when it was very controversial. Back in the '80s it became a center of a lot of social, political debate. There were people who felt that kids were being labeled with a name, ADHD, so that they could be given medication when what they were suffering from was bad schoolteachers, and they were going to inadequately funded schools or inadequately trained teachers, and they were responding just like any normal kid would to a situation where the teacher wasn't giving over their message in a way that they understood.
So the pills became an area of political controversy. I think what changed with ADHD -- first, what happened in the '90s was that with the advent of the new antidepressants and direct marketing of the antidepressants, a lot of the concern about psychoactive medication was reduced. It became a much more respectable thing to do. People would talk about it. They would compare notes. Kids would compare notes and so on. So there was the general reduction in the stigma that was attached to giving medication, not in all groups, but generally. I think that the prescription of Ritalin and the stimulants generally benefited from that.
The other thing is that there was a very large study that was done by NIMH [National Institute of Mental Health] under very rigorous conditions that showed overwhelmingly that if you had ADHD and you didn't get Ritalin, you were not going to get better no matter what kind of psychotherapy or how intense the psychotherapy was. Those things helped move it out of the area of controversy.
Why is bipolar popular? I think bipolar is fashionable. It may also be popular because most of the psychiatric disorders of childhood are chronic: You don't just have one and then you take some pills, and then you get better, like an episode of depression. Most kids who have psychiatric problems, they tend to last or to resurface repeatedly. That's demoralizing, and people are always looking for alternative explanations.
And one of the great things about bipolar disorder [in adults] was that among the psychoses, it was a treatable one. You could keep it under control if you took your lithium and so on. So it had a flavor of being not such a bad diagnosis, and I think that that was attractive.
I think that a lot of kids with the very severe kinds of behavior disturbance that you see in what is called juvenile bipolar, those parents are very sorely tried. They have a rough time, and the kids have a rough time. So looking for a diagnosis that seems to offer a lot of hope is very attractive.
It is fashionable because for many years it was thought to not exist in children, and then, all of a sudden, it became fashionable because it was common: Out of nowhere bipolar disorder suddenly was being diagnosed left, right and center.
At specialized or university centers, a lot of the cases they were seeing were kids who had been diagnosed as bipolar by a local psychiatrist or pediatrician. The parents wanted a second opinion to see if that was actually true.
The reason why I call it fashionable rather than real -- fashion I guess implies that it may not be real -- is because there are certain bits of evidence that are sorely lacking in that if you do follow-up studies of kids, if you follow back studies of adults who have bipolar illness -- clear, conventional bipolar illness -- you don't find an awful lot of ADHD in their childhood histories. Equally, if you follow up kids with ADHD, you don't find bipolar as they grow up. So that's a big problem.
The other thing is that the family history and the heritability of juvenile bipolar is not nearly as strong. It's very different to the heritability that you get in adult bipolar. So there were good scientific reasons to worry about the validity of the diagnosis.
And there are very good practical reasons to worry about it, because the medications that are used are by and large untested in children. Most were developed as anticonvulsants. The amount of research on whether they work in children who have had these diagnoses is not in any way adequate or satisfactory.
We know that the medications have the potential for being quite worrying. Something like Depakote, for example, causes ovarian dysfunction in adolescent girls, and some of the other medications used can cause endocrine problems. Some of the medications that are used, like Topamate, which is attractive to people because you lose weight on it, also can cause severe impairment of memory and other kinds of thinking. So what you've got out there is a whole lot of kids diagnosed with a condition that hasn't really attained respectability yet, and they're receiving medications that have not been fully tested in children.
That's the reason why I call it fashionable. It's not like here is something that we definitely know about.
Sort of a great experiment?
Well, it's not an experiment, unfortunately, because the way it's used is not experimental. We don't really know what would have happened if those very same children would have received a different treatment.
It's a worry to many, many psychiatrists, pediatricians and parents, because it seems to be given so loosely. It has a great scientific value because it's actually stimulating research in that area, so it should lead to some clarification eventually.
Compared to other countries, is the U.S. too eager to solve problems using medications?
You know, there's a lot of wonderful biology that goes on in British labs, universities and pharmaceutical companies and so on. But it's hard to get that accepted in practice there, where the average physician and psychiatrist really still feels strongly that everything is due to the environment.
America, I think, is much more open to new ideas, and it's open to try things. And I don't think it's a bad thing. I think that it means that periodically, maybe we overdo it.
Tell me a bit about the DSM-IV and what impact that has on the field.
Well, the big DSM that had an impact on the field was DSM-III, because the major element in it accurately said, "We don't know the cause of most of these conditions." All the previous classifications of psychiatric disorder had assumed that we did, so they would normally have a sum causal element built into the description. And DSM-III said, "We don't know what the cause is." We fight about causes, we disagree and argue about them, so let's just describe what we can see and leave it that.
It was an enormous development. It really allowed people to escape from their boxes and to go ahead. It was a momentous event. What they had done in Europe or the rest of the world with the international classification of diseases was they kind of did that, but in a half-baked way.
What happened with DSM was that they actually laid out all the different symptoms, and described the different symptoms more or less fully, and they said how many you have to have and so on and so forth. This was leading the way toward a kind of precision which was unknown before.What the ICD [International Statistical Classification of Diseases and Related Health Problems] had done was simply to give a very broad and general description.
Some people seem to advocate that it's not helping, the DSM-IV, because there's so much overlapping in the criteria for different diagnoses and you cannot sort of discreetly say much about anything, the overlap is so confusing.
What DSM did was, previously, if you had somebody who had, say, anxiety and depression, or schizophrenia and depression, they said: "Oh, they have schizophrenia, and they have a secondary depression. They've become depressed because of some consequence of the other illness or whatever."
People would use this notion of primary disorder and then a secondary disorder. Again, it was because DSM said: "Well, we don't know the causes of all these things. Who knows? They may be caused by different things." They said, "You can have as many diagnoses as you've got, and we're not trying to get into the business of guessing what's primary or secondary."
So DSM allowed comorbidity to be discovered because it didn't exist before then. It had all been rolled into some notion of what's primary, and people differed in what's primary. Some people said it was what comes first. Other people said it's what's impacting your life. Some people said it's what is worst. So that was one of the big benefits of DSM. But one of the consequences has been that it adds uncertainty to what the significance of the comorbidity is.
I think the DSM was right, because if you look at anxiety and depression in adolescents, a lot of kids who are anxious have periods of depression and vice versa. People have thought, well, ... if you're socially anxious, that means you don't have friends; you don't have support systems. It's going to make you depressed. So they assume that anxiety was the primary disorder and depression was secondary. But there's some wonderful studies which have been done recently which show that that's not how it works; that if you have anxiety and depression, they alternate in how they appear, so it's not that one comes first. It's more complicated than we thought. I don't think that it's a weakness that we discovered that there's a lot of comorbidity.
And DSM-V, where do you think that will head?
Originally, people hoped that by the time DSM-V was ready we would know the genetic or physiological origins of most of the disorders and that we would be able to have a classification system that was more like you get in other diseases where you might get vitamin deficiency, which can be manifest in lots of different ways.
In practice, it hasn't worked out like that for psychiatry, so my guess is it will end up clearing up some of the uncertainties and problems that were present in the last one. I don't think that we're going to see anything very dramatically innovative, and maybe that's good.
One of the most interesting things in depression that has emerged in the last decade has been a discovery of the difference between early-onset depression and depression which comes on later in adolescence. I think that that's still not widely recognized.
Depression before puberty is really rare. Depression in adolescence is common, maybe even more common than adults. What people have discovered by following up previously depressed kids in adolescence is that the depression that occurs in very young children is quite different. It is associated with a lot of psychiatric problems in the family but not necessarily mood problems, whereas in adolescence, as in adulthood, it's usually associated with a family history of mood problems.
Almost any kind of psychiatric illness, and presumably some kind of impact that that has on the family, is found in the very small number of kids who get depressed when they're young. What happens to them is also quite different to adolescents who get depressed in that when they grow up, they don't necessarily have depression; they have a variety of psychiatric problems.
It's a very special group, and I think there's a message there for the bipolar people that age of onset may signal something quite different. They're assuming that age of onset for bipolar disorder is simply telling you that it's going to look different, but it's going to behave the same later on. If we take the example of depression, that's not what's happened.
So I don't think my ideas of depression have changed much. There's always the question about whether all of the ways in which depressed people think, is that primary, or is it secondary? Does the state of depression make you self-critical? Does it make you think badly of yourself? Does it make you pessimistic? Or do you become depressed because you're pessimistic and you put yourself down and so on?
More and more, it's clear that the first is right; that one of the consequences of being depressed is that you start to think in those very negative ways. That's a very important distinction, because there's an enormous number of people who are doing research and practicing on the assumption [that] how you think determines whether or not you're going to get depressed rather than the other way around.
The other interesting thing about depression is that findings from the recent STAR*D Study [Sequenced Treatment Alternatives to Relieve Depression, a study funded by NIMH], showed that it's very important for a child psychiatrist that if you treat a mother's depression with pills or with psychotherapy -- just with pills will do -- that a very high proportion of their disturbed children will get better without you doing anything with the children.
Medicate the parents.
Yes. ... If you can do that, you're getting at a root cause. If you don't treat the parents and they remain depressed, and either so depressed that they can't really watch their kids or give them the time or the attention that they would like to, or if they go on crying in front of their children or go on talking about wishing they were dead in front of their children, or go on drinking because they're depressed and they're not being treated, then all the things that you try to do with the kids are not going to be as effective as they might be.
Something unusual happened with suicide study in the '60s. The rates had been fairly constant and hadn't changed very much. But starting in the early '60s, the rate in males started to shoot up. Between the early '60s and the mid-80s, that period of some 25 years, the rate in males had tripled.
People hadn't realized that it was only affecting males. They were aware that youth suicide was much more common, and they weren't really quite sure why.
It gradually emerged as the studies started to be done that the probable reason for that increase was due to much greater use of drugs and alcohol.
Alcohol is intimately related to suicide. If you look at older adolescents who commit suicide, two-thirds of them are abusing alcohol. Alcohol is related to suicide throughout the life cycle. It's a particularly suicidogenic substance. We don't fully understand why.
Then starting in 1994, the rates in males started to come down. It came down a full third between 1994 and 2003. Historically, if you look at the teenage suicide rates in historic terms, there had never been the same increase, and there had only been two previous occasions when the rate had come down. The first one was during Prohibition, so before World War I and ending up after World War I, and the second was during World War II. So there was a lot of excitement about why; something good must have been happening. The rate of suicide had gone down by a third, far greater than anybody could have conceivably had hoped for.
One looked for all sorts of explanations. If the increase had been due to drugs and alcohol, maybe kids weren't using so many drugs or drinking so much alcohol. But all of the surveys of that really didn't show much reduction. If anything, [there was] a slight increase.
It wasn't that more kids were getting treatment, because, in fact, the rates of psychotherapy at any rate had fallen considerably. It could be that more kids were going into the emergency room and being better treated than they had been. That's always possible. But it's unlikely, because we might have seen that before.
What most of us concluded was that what was happening was that finally there was an antidepressant medication that was safe to give young people, and it was being very widely used, and that one of the good effects of that was that it was reducing the suicide rate.
And it wasn't just in the United States. You've got the same phenomenon occurring at the same time in nearly all of Western Europe and in Canada, New Zealand, [Australia], other countries where ... you've got the medications and good statistics. You saw exactly the same thing happening. That seemed to be what was driving it, and that was the general assumption.
The other thing was that there had been a few autopsy studies in which people who had committed suicide were examined. Their blood fluids were looked at. It was first done in Sweden in adults, and it was noted that most of these people had no trace of an antidepressant in their system. So the investigator who did this said, "Well, maybe this is more evidence that suicide occurs when you don't treat depression adequately."
Then, of course, the whole thing was turned on its head when the British Medicines Commission and the FDA held hearings, and a whole lot of data were looked at, and it was concluded that SSRIs [selective serotonin reuptake inhibitors] -- the antidepressants that I still think led to the decline in suicide -- had actually seemed to be associated with an increase in talking about [suicide], thinking about suicide, and also in suicide attempts. But I've got a lot of problems with that conclusion because I think it was based on evidence that is not perfect.
So in 2003, suddenly the sort of savior drug is being questioned in England. Paxil is the major culprit for all sorts of bad things, apparently. You're watching this from your office up at Columbia University. What are you thinking is going on in England? And how could you explain that period?
Similar concerns about antidepressants being related to people talking and thinking about suicide had been voiced shortly after they were introduced in the 1990s. That had been followed by an enormous volume of restudy of tens of thousands of subjects, and it was concluded that the suicide rate was the same or higher in placebo than in people who had received antidepressants. But those hadn't included any child studies.
What happened with the black box then?
The FDA, to be absolutely fair, took their time in deciding what to do, and they had another set of hearings, and they decided on a labeling change, which included a black box.
The black box, by and large, is terrific. It tells you all the things that should happen when you go on a new medication: Stay in touch with your doctor; know how to reach them, and they should reach you; they should ask about this, that and the other; you should get a list of possible side effects and so on. Nobody would quarrel with that black box. But from what we can tell, what happened was prescriptions for antidepressants fell by between 15 and 25 percent, starting in 2004.
Now, suicide figures take about two years to come out because they're collected from medical examiners' office[s] throughout the country. So the date when the 2004 figures were going to come out was a day that those of us who were interested were waiting with great suspense, you know? We'd had nine years of uninterrupted decline. Would that be continued, or would that be interrupted?
And then the [numbers] came out the week before Thanksgiving, and lo and behold, they'd increased: Nearly 350 more kids committed suicide in 2004 than did in 2003. And although you could say that this points to the fact that SSRIs probably are protective, it was hard to get much attention to that finding. In general, I think newspapers have been more interested in talking about the dangers of this than talking about the terrible dangers of not getting treatment.
Have there been other spikes? Or doesn't the suicide rate sometimes blip up and trends down?
In 2001 or 2002 there had been a very small, brief uptick and then had come down. And you're right, it may be that we just have to wait to see whether, first of all, the trend to prescribe fewer antidepressants continues, and if while it does, whether the suicide rate adapts to a new level or whether it's just a single blip and it goes down again.
Do you think this is a tragedy, though? We sort of put a black box on the one medication that might really be safe and that may really be helpful.
People were concerned that that might happen, and that's why the FDA in contrast to some European countries did not put limits on prescribing the antidepressants. It's a tragedy for those families if this is true. But then, life is full of taking two steps forward, one step backward, two steps forward. It's the way knowledge progresses.
Do you think that this has been a backlash, that the black box came out because there was this concern that too many psychiatrists are being pushed by the pharmaceutical industry and that we don't know enough?
I think it had occurred in the context of some misbehavior by one corporation that was not completely frank. If they had been frank when they knew, that could have stimulated research, which could have actually answered the question, instead of which it was hidden. So that's a tragedy.
Many good things occurred as a result. First, not only were people more prepared to take antidepressants, which have made a major change to their lives, but it more than anything destigmatized mental illness, because up until these antidepressants became available, everybody went around thinking that if you were depressed, you were a loser; you were depressed because you couldn't cope.
Insofar as the people who were trying to get you to take medication put forward a more neutral model of "This is in your brain; it's in your genes; don't blame yourself," I think that will always be helpful, and I hope it persists. My feeling is that probably more and more evidence will emerge to clarify the situation.