Is Bipolar Disorder Overdiagnosed in Children and Adolescents: A Virtual Debate

Kiki Chang, MD; Gabrielle Carlson, MD; Stephen M. Strakowski, MD


September 10, 2010

In This Article

The DSM: a Blessing and a Curse

Dr. Chang: All points well taken. The DSM, as we all know, is a blessing and a curse. The common language at least gives us a template to work with in the pediatric realm, but the difficulty of porting adult-type symptoms into children leaves much room for interpretation, and possible misinterpretation. Having clinicians and researchers sensitive to developmental issues helps distinguish "normal" childhood behavior and mood from "abnormal," but in no way solves the problem.[10]Gaye's point regarding schizophrenia is informative -- similar dimensional approaches could be taken to target bipolar disorders in children. However, as Steve mentions, due to multiple etiologies leading to similar (but still separate) final phenotypes ("equifinality," to quote Dante Cicchetti[11]) it is challenging to identify these various moving targets and to classify them correctly, as Judy Rapoport and the multidimensionally impaired project found.[1]

Identifying a few good spectrum-type of disorders outside of Cade's disease in kids, and following them longitudinally, is a good place to start. And yes, I agree we are behind on that initiative. For example, although I am against the inclusion of temper dysregulation disorder with dysphoria (TDD-D) in DSM-V, at least it is an attempt to start this spectrum process. However, more thought needs to go into this initiative, with several phenotypes represented. For example, BD not otherwise specified (NOS) in kids is making good progress I believe, given the Course and Outcome of Bipolar Illness in Youth (COBY)[12] data. These researchers have found that strictly defined BD NOS -- consisting mainly of children who have irritability or euphoria together with 2-3 manic symptoms that last at least 4 hours, recur, and cause impairment -- is a phenotype that often eventually leads to the BD I or II phenotypes. Other phenotypes could include unipolar depression, unipolar depression with brief manic symptoms, extreme irritability in different contexts (a la TDD-D), and anxiety (diffuse) with irritability, etc.

Also, we should realize that there may be "new" emerging disorders in children that simply were not that common 50 years ago. People I respect who have been in the field longer than I, clearly report increasing incidence of troubled kids -- younger, more pervasive developmental disorder (PDD)-type problems and more mood difficulties. Gaye, I wonder what you think of the possibility of this overall phenomenon of increasing psychiatric disorders in youth, or is it like global warming (controversial I know) in that some say we are just hitting a dip in a normal cycle, and that overall the big picture is no different if you wait another 20-50 years?

Dr. Carlson: I will accept responsibility for being the person who sees diagnosis a bit differently than some of you who have been trained more recently. I realize that is for a couple of reasons.

First, as I said earlier, I did my training before DSM-III, when one made a diagnosis by the resemblance of a condition to a template. We did semistructured interviews. However, there was the pattern recognition approach, as done with ICD 10. DSM tried to deconstruct the pattern into criteria but didn't quite do it, as we are finding out now in the definition of episodes that have an onset but not an offset. It's kind of like how paint by numbers tried to deconstruct a masterpiece into its component parts but never really got the masterpiece, just something that resembled it. So, I'm afraid that we have changed bipolar disorder into a condition of positive responses to structured interviews that ask about bipolar disorder. This is true for adults and kids.[13] Interestingly, a recent study suggests that the ICD approach to bipolar disorder has better diagnostic consistency than DSM IV so perhaps there is some merit in this approach.

Second, I always get a clinical history, and an extensive one, before I do a K-SADS. To me, the K-SADS is a mop-up operation that does a confirmatory psychiatric review of systems. I think many folks do the opposite. They do a K-SADS first and then, if they feel the need, a clinical interview. It might be worth doing a study to see whether that in fact makes a difference.

Third, when you don't really have a specialty clinic and take all comers (which is what our clinic does), and folks who see me bring me their most diagnostically complex cases, I have to do more than say "Bipolar -- in or out" of a study. I have to figure out what is wrong and why, and I have to go beyond whether something meets criteria. Many times the poor kid has the major problem of not meeting criteria for anything, or meeting criteria for something that really isn't his major problem (like ADHD). One of the nice things we did at Washington University's Renard Hospital many years ago (with the birth of systematic interviews) was to be honest and say "Undiagnosed, most like depression, or mania, or schizophrenia, or whatever." That I can live with.

I can verify that the forced choices of structured interviews give an interviewer something different than a open-ended clinical interview in which decisions are not forced into a specific algorithm. There are times when I do a K-SADS when I simply cannot answer a question honestly. I can push the parent or child into saying something just to move on, but I feel unclean when I do it. I'm sure I'm not alone in this. But it isn't anything we ever talk about.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: