‘Paediatric bipolar disorder’ rates are lower than claimed – a reexamination of the epidemiological surveys used by a meta‐analysis

Abstract

'Paediatric bipolar disorder' (PBD) is a controversial diagnosis where often prepubertal children as well as adolescents, who may have a range of psychiatric disorders or symptoms, are diagnosed with a severe mental illness requiring lifelong medication. Clinically, it has often been applied in the United States but rarely in most other countries. A meta-analysis (Van Meter et al., Journal of Clinical Psychiatry, , 72, 1250) claimed that the prevalence of PBD was similar to adults at 1.8% with no difference between the United States and other countries. This conclusion has been highly cited. The heterogeneous nature of the original 12 epidemiological surveys warrants a qualitative analysis, rather than statistical meta-analysis as performed by Van Meter etal. (Journal of Clinical Psychiatry, 2011, 72, 1250). Thus, the meta-analysis and each of the 12 studies (six from the United States; six from other countries) were reexamined. Most of the 12 surveys predated the emergence of the PBD hypothesis. The 12 surveys were mainly of adolescents and at times young adults with few prepubertal children. Prevalence rates in the 12 studies suggest a lower rate of bipolar disorder, especially in non-US samples. For example, the Van Meter etal. (Journal of Clinical Psychiatry, 2011, 72, 1250) meta-analysis chose a rate of 2.8% by summation of adolescent and parent responses in a Dutch survey, however the rate fell to 0% if requiring concordance of adolescent and parent responses. Indeed, it could be argued that four of the non-US studies show 0% rates of PBD. Rates of PBD were generally substantially lower than 1.8%, particularly in non-US surveys, and if both parent and adolescent reports were required to meet the diagnostic threshold they fell to close to zero. The reanalysis suggests that bipolar disorder is rare before the expected age of onset in later adolescence.