14 Problems With Geller Study Of Child Bipolar Disorder

After reading the full paper a few times, I am concerned about this paper and its conclusions and how it may well be spun by the press for many reasons. The study was funded by an NIMH grant. The lead researcher was Barbara Geller, a child psychiatrist at Washington University.

I suppose studies like this are open to a range of interpretations and quibbles. Here are mine. BTW, I am not asserting that something isn't wrong with these teens, just pointing out that I have real questions about just how aggressive we've gotten in American culture in softening bipolar disorder away from its classic manic-depression roots.

1. The media is already spinning this as a study showing that child bipolar disorder (the formal term is pediatric bipolar disorder per the FDA) persists into adulthood. Bloomberg does so here. Geller has contributed to the problem by entitling her paper "Child Bipolar 1 Disorder and arguing that her study "strongly support[es] continuity between child and adult BP-1." Her study does something different since it doesn't measure outcomes once these kids hit 18 year of age with adult measures.

2. The study isn't really about kids per se--and certainly not the toddlers and 6-year-olds who are at the center of all the controversy around the diagnosis--but it's mostly about teens. Children had a first episode of what Geller calls mania at an average of 11.1 years of age, and the 8-year follow-up study ends at about 20 years old. Some of the kids entered--meaning they had their first "manic" episodes--at about 8.5 years old, but globally this study is about adolescents which sort of argues for calling all of this "teen bipolar disorder." Or teenage wasteland. Or something else. But child bipolar disorder is a stretch and pediatric bipolar disorder is a bit broad.

3. Teens who receive a higher level of "maternal warmth" fared better than teens who didn't, dramatically better in fact. It's unclear to me what Geller means by maternal warmth in a formal sense, but it's safe to assume that having a mother who gives a damn about their kid is--duh--important.

4. There is no mention in the study of paternal warmth. So is Geller arguing by implication that dads don't matter or are all these teens' fathers missing in action?

5. The study is male-heavy--67 percent male, 33 percent female. It is also a small study with 108 out of 115 initial enrollees finishing out their eight years. I hope people in the media pay attention to how small this study was and reign in their conclusions appropriately.

6. The conclusion that the evidence is "strongly supporting continuity between child and adult BP-1" is based upon her finding that 44.4 percent of study subjects who reached 18 years of age--or or 24 of 54 subjects--experienced some symptoms of mania. Twenty-four subjects only represents 22.2 percent of the 108 subjects who finished the study. That's a pretty small subset upon which to make broad claims. What's more, over half of the participants haven't hit 18 yet. Thin reeds, as the saying goes.

7. Study participants are 85.2 percent white. Geller notes that another limitation of the study is that it mostly drew from kids and teens of higher socioeconomic status. She then steers readers to the paper's "Methods" section where there is no discussion of what higher socioeconomic status beyond noting that study subjects had health insurance. That's sort of incomplete and sloppy.

8. There is no discussion whatsoever of what medications or psychotherapy or both these teens are receiving although Geller makes it clear they are receiving care in the community. Not sure what that means. It would've been nice to at least have this addressed in the paper's discussion section.

9. The term BP-1 mania as used by Geller isn't the same thing as adult BP-1 mania. It's a kiddie mania that instead of requiring 3 of 7 symptoms (sometimes four) to be present for the better part of seven days (outside a hospitalization) only requires that one of two "cardinal" symptoms be present--namely, either elation or grandiosity. Part of this watering down of BP-1 mania is to limit overlap with some of the more ADHD-ish symptoms of classic mania and to keep the "irritability" type of bipolar kids we hear so much about from the Harvard child psychiatrists either out of the study or their symptoms out of being counted as mania; ie, distractibility, increased goal-oriented activity, talkativeness and so on. Also out are flights of ideas and the classic "excessive involvement in pleasurable activities that have a high potential for painful consequences." So no spending sprees or sexual indiscretions.

Separately, I have to wonder why they are using a mania proxy designed for adolescents and measuring it in young adults and calling that proof of the transition to adult mania (meaning real full-blown DSM BP-1 mania). Wouldn't you use adult measures of mania (both the DSM and the Young Mania Rating Scale) as your end point to prove the transition? In this study, Geller hasn't and I can find no discussion of this point in the paper. The editor and peer reviewers didn't query her on this point? She didn't feel it necessary to explain going with the Children's Global Assessment Scale or the Global Assessment Scale for adult-aged subjects to the exclusion of the Y-MRS (at least on the adult end)? Weird.

The GAS isn't designed to capture mania per se. Psychological troubles and social functioning, yes.

10. It's not clear to me exactly what kinds of episode durations counted as manic episodes, but the episodes durations of this watered down mania that Geller leaned on were very long, hundreds of days long. Which makes me a bit suspicious. In some cases, the teens were having episodes of four hours in length several times a day. In some cases not.

11. Sixty-one percent of the kids were reported as experiencing psychosis upon entering the study. That's a troubling rate, but it's not clearly to me exactly what psychosis was going on and how it was delineated. Geller uses the WASH-U-KSADS scale and I cannot find a relevant copy of it.

12. The same day this study appeared a study based upon a survey of 199 child psychiatrists in Australia and New Zealand was published in the journal Child and Adolescent Mental Health.

"Opinion varied as to whether PBD was over-diagnosed (25%), appropriately diagnosed (42%), or under-diagnosed (28%) in Australia and New Zealand, 5% were unsure. In contrast there was a consensus of views that PBD was over-diagnosed in the USA (90%), whilst less felt it appropriately diagnosed (3%), or under-diagnosed (1%) and 6% were unsure.

"Conclusions: The majority view was consistent with classical descriptions of bipolar disorder."

13. Let's assume for a minute that I am wrong and Geller is right and that what we've got here is a type of bipolar disorder than persists into young adulthood. If that's the case, then the outcomes certainly don't speak very well of the prevention paradigm--in which the earlier you diagnose and treat mental disorders, the more successful you are at limiting its problems--because 44.4 percent of these young adults, diagnosed with BP-1 mania as adolescents or children, are still symptomatic with some form of what Geller calls mania and another 29.6 percent have some form of unipolar depression. So whatever treatment these teens are getting don't seem to be working especially well. Forty of the 54 now-adults are symptomatic for something, if I read the study correctly.

14. I find it difficult to believe that Geller has made her case of "strongly supporting continuity between child and adult BP-1." What she has done is made the case of continued psychological issues for these former teens at the cusp of adulthood but hasn't proved much else.

But then I am a skeptic. Or an Aussie or Kiwi child psychiatrist.

Posted by Philip Dawdy at October 7, 2008 12:03 AM
del.icio.us Digg it reddit