Structure of Hypomanic Symptoms in Adolescents With Bipolar Disorders: A Network Approach.

ABSTRACT

Background: Bipolar disorders (BD) are severe mental illnesses that are often misdiagnosed or under-diagnosed. The self-report 33-item Hypomania Checklist (HCL-33) and the 33-item Hypomania Checklist - external assessment (HCL-33-EA) are well-validated scales for BD symptom detection. This study compared the network structure, central symptoms, and network stability of hypomanic symptoms measured by the HCL-33 vs. the HCL-33-EA. Methods: This cross-sectional study was conducted from January to December 2019. Adolescents (aged between 12 and 18 years) with BD were recruited from the outpatient department of Child Psychiatry, First Affiliated Hospital of Zhengzhou University. All participants were asked to complete the HCL-33, and their caregivers completed the HCL-33-EA. Network analyses were conducted. Results: A total of 215 adolescents with BD and their family caregivers were recruited. Node HCL17 ("talk more," node strength = 4.044) was the most central symptom in the HCL-33 network, followed by node HCL2 ("more energetic," node strength = 3.822), and HCL18 ("think faster," node strength = 3.801). For the HCL-33-EA network model, node HCL27 ("more optimistic," node strength = 3.867) was the most central node, followed by node HCL18 ("think faster," node strength = 3.077), and HCL17 ("talk more," node strength = 2.998). In the network comparison test, there was no significant difference at the levels of network structure (M = 0.946, P = 0.931), global strength (S 5.174, P = 0.274), or each specific edge (all P's > 0.05 after Holm-Bonferroni corrections) between HCL-33 and HCL-33-EA items. Network stabilities for both models were acceptable. Conclusion: The nodes "talk more" and "think faster" acted as central symptoms in BD symptom network models based on the HCL-33 and HCL-33-EA. Although the most prominent central symptom differed between the two models ("talk more" in HCL-33 vs. "more optimistic" in HCL-33-EA model), networks based on each measure were highly similar and underscored similarities in BD symptom relations perceived by adolescents and their caregivers. This research provides foundations for future studies with larger sample sizes toward improving the accuracy and robustness of observed network structures.