Is Bipolar Disorder being over-diagnosed in children?
Many people erroneously argue that it is, so let's take a look at the evidence. First, I am talking about children aged five years into adolescence who actually meet the full DSM-IV criteria for Bipolar Disorder. I am not discussing children who only have irritability or aggression without any other manic criteria.
In 1994/1995 the frequency of the Bipolar Diagnosis was 25 cases per 100,000. In 2002/03 the frequecny increaed over forty-fold to 1003 cases per 100,000. (See: National Trends in the Outpatient Diagnosis and Treatment of Bipolar Disorder in Youth, Carmen Moreno, MD; Gonzalo Laje, MD; Carlos Blanco, MD, PhD; Huiping Jiang, PhD; Andrew B. Schmidt, CSW; Mark Olfson, MD, MPH, Arch Gen Psychiatry. 2007;64(9):1032-1039.)
Put another way, the diagnosis of Bipolar Disorder rose from 0.01% in 1994 to 0.44% in 2002. That is certainly a 40X increase, but is it really over-diagnosis?
The standard method to estimate the actual prevalence of psychiatric conditions is to conduct an epidemiological study, where trained researchers study large representative samples of the general population (not preselected clinical samples) with standardized, validated instruments (like the SCID or MINI or CIDI) to assess and diagnose the general population using DSM-IV criteria (not just irritability or aggression in children as bipolar disorder, for instance). Using those methods in over 5000 persons, the most recent analysis (see: Lifetime Prevalence and Age-of-Onset Distributions of DSM-IV Disorders in the National Comorbidity Survey Replication
Ronald C. Kessler, PhD; Patricia Berglund, MBA; Olga Demler, MA, MS; Robert Jin, MA; Kathleen R. Merikangas, PhD; Ellen E. Walters, MS, Arch Gen Psychiatry. 2005;62:593-602.) reports a one year adult prevalence of bipolar disorder (types I and II) of 3.9%. In an analysis of age of onset in the same citation above, the study found that bipolar disorder began before adulthood in a substantial minority of persons: in 25% of persons, it began by age 17, in 10% it began by age 13.
Let's do the math now and see what results. Excluding the adolescents up to age 17, and using the lower range of the NCS data:
If 10% of bipolar disorder begins by age 13, and 3.9% of the total population is diagnosable with the condition using DSM-IV criteria in a community-based epidemiological study of actual prevalence, then how many children should that clinical study have diagnosed with bipolar disorder? 0.10 multiplied by 0.39 is 0.039, or 0.39%. If we round to 0.4%, then we have the actual prevalence of bipolar disorder in children. The vaunted 40-fold increase from near zero led to the diagnosis being made in 0.4% of children based on percentage of office visits.
0.39% is about the same as 0.44%, so the purported evidence for over-diagnosis seems underwhelming, to say the least.