Poster #15 - Phenotypic Structure of Common Early Adolescent Psychopathology Symptoms, With Prediction From Childhood Temperament

• In Event: Poster Session 04

Abstract

Introduction
Psychiatric disorders are a leading cause of disability in young people (Baranne & Falissard, 2018) and affect up to 1 in 5 adolescents in the United States (Costello et al., 2011). Thus, preventing mental disorders is a major public health goal. Early diagnosis and intervention are crucial for reducing disease, but problems with the nosology of mental disorders impede their diagnosis and treatment. Among the limitations of current psychiatric diagnoses are heterogeneity within disorders, phenotypic and genetic overlap across disorders, and high rates of comorbidity.

More accurate models of the structure of psychopathology will elucidate at what levels etiological variables relate to disorders, allowing for targeted prevention efforts. Studies on the phenotypic architecture of psychopathology show common variance between disorders that can be modeled using higher-order factors (Kotov, 2017). An especially promising predictor of psychopathology is temperament, or dispositions present from infancy that influence activity, reactivity, emotionality, and sociability and broad individual differences in the expression of affect (Goldsmith, 1987).

Using temperament as a predictor of psychopathology is especially useful during adolescence, a vulnerable period of peak onset of symptoms (Kessler et al., 2005) that remains understudied (Castellanos-Ryan et al., 2015). Thus, this study delineates the latent structure of mental disorders in 13-year-olds. Next, we examine whether temperament at age eight, controlling for early-emerging symptoms of psychopathology at age eight, is a useful predictor of later higher-order factors of psychopathology.

Method
Participants (N = 849) are from the Wisconsin Twin Panel (WTP), a birth-record based study of twins born in Wisconsin between 1989 and 2004 (Schmidt et al., 2019).

Temperament was assessed with the Child Behavior Questionnaire (CBQ; mother and father report) when participants were on average eight years old. Psychopathology was assessed with the MacArthur Health and Behavior Questionnaire (HBQ; mother and father report) and the Diagnostic Interview Schedule for Children (DISC-IV; mother report) to assess symptoms of psychopathology when participants were on average eight and 13 years old.

Results
Exploratory structural equation modeling (ESEM) was used to examine the structure of 8yo and 13yo psychopathology. The best-fitting model included two-factor solutions for both 8yo and 13yo psychopathology (internalizing and externalizing) and a three-factor solution for 8yo temperament (effortful control, surgency, negative affect). Model fit was acceptable (CFI = 0.906, TLI = 0.879, RMSA = .069, SRMR = .043). Other models tested included three-factor models of psychopathology (i.e., fear, distress, externalizing) and bifactor models (i.e., general factor, internalizing, externalizing). Age 8 temperament factors did not predict age 13 psychopathology factors over and above age 8 psychopathology.

Discussion
Results indicate that symptoms of psychopathology in this community sample are best captured by a two-factor solution of internalizing and externalizing factors at ages 8 and 13. Adolescence is characterized by biological and behavioral changes, including an increase in incidence of psychopathology. However, the same model of psychopathology symptoms fit in both middle childhood and adolescence. Further, middle childhood temperament did not have any utility as a predictor of adolescent psychopathology over and above early emerging symptoms of psychopathology.

Authors

• Emily Hilton, University of Wisconsin - Madison
  Presenting Author

• Robert Krueger

• Kathryn Lemery-Chalfant, Arizona State University

• H Hill Goldsmith, University of Wisconsin - Madison