Occurrence and accuracy of a register-based diagnosis of pediatric bipolar disorder: A nationwide cohort study

Objective: To investigate the accuracy of a diagnosis of pediatric bipolar disorder in the Danish National Register compared to the patient charts. Second, we reported on the occurrence of a diagnosis of pediatric bipolar disorder during the study period. Methods: All persons appearing in the Danish nationwide registers between 1995 and 2014 with an incident ICD-10 diagnosis of single hypomanic/manic episode or a diagnosis of bipolar disorder (summarized as bipolar disorder [BD]) before turning 18 years were identified (n = 521) and a random sample (n = 131) hereof was selected for chart review. Each chart was reviewed by two independent Schedules for Clinical Assessment in Neuropsychiatry (SCAN) certified raters to assess whether symptoms documented in the chart were consistent with a formal diagnosis of BD according to the ICD-10 criteria or not. Results: The formal diagnostic criteria for BD according to the ICD-10 were fulfilled in 48 charts (45.3%, 95% CI: [36.1%, 54.8%]) out of 106 reviewable charts, age at index = 16.4 +/- 1.6 (range = 9.1-18.3) years. Cohen's Kappa ranged from 94.4% to 100%. The estimate of a lifetime prevalence up till the current age for bipolar disorder for those of aged 5-18 years, was 0.019% in 2014. Conclusion: Less than half of the register-based pediatric BD diagnoses were confirmed by chart review, which was lower than expected. The occurrence of a register diagnosis of pediatric BD was relatively low.