Category: Child / Adolescent - Anxiety
Rates of adolescent internalizing problems are a significant public health concern. Little focus has been placed on specific biological and environmental factors associated with the shift from subthreshold to clinical levels; thus, there is a critical need to characterize adolescents with subthreshold impairment who are at risk for escalating to clinical levels. The objective of this study was to characterize risk profiles for adolescents who developed clinical levels of internalizing problems by age 16, using data from brain, genetic, personality, cognitive, history, psychopathology, and demographic measures.
Data were from the IMAGEN study, which includes biological and environmental data for approximately 2,000 14-year olds, with follow-up assessment at 16. Adolescents had complete data on clinical diagnoses at ages 14 and 16, and did not meet clinical criteria at age 14. 93 participants met criteria for Anxiety and Depressive Disorders at age 16, and 1,394 who did not were used as controls. Ten-fold cross-validated regression analyses using elastic net regularization were used to identify risk profiles associated with an escalation in internalizing symptomatology. Logistic and linear regressions were conducted, with clinical diagnoses as the outcome for the logistic regression and symptom-severity scores as the outcome for the linear regression performed on controls.
Eight variables from demographic, personality, psychopathology, functional, and structural domains were associated with having a clinical diagnosis two years later (mean area under the ROC curve 0.70, SD=0.02, p < .001). Twenty variables from life events, personality, psychopathology, behavioral, functional/structural, and genetic domains were associated with an escalation in symptomatology in the non-clinical group (R2=.23, SD=.005).
Overall, factors from demographic, personality, psychopathology, structural, and functional domains characterized adolescents at risk for increased levels of impairment. One subset of variables emerged as related to clinically-diagnostic levels of impairment, and appears to represent a set of factors that identifies more extreme risk. One subset of variables appeared to characterize a dimension of risk, such that they emerged when examining symptomatology both as a categorical and continuous measure. Specifically, puberty, neuroticism, and anxiety levels predicted clinical diagnoses and symptom-severity in the control sample; however, distinct brain regions were related to clinical impairment versus non-clinical impairment. The other subset of variables related to the classification of symptom variation from a dimensional perspective, and appeared to only classify less extreme symptomatology variation within the sample. We not only identified risk profiles for adolescents at the highest risk for escalating to clinical levels of internalizing disorders but also characterized risk within the overall population. Findings enrich our theoretical understanding of internalizing disorders and the factors that are associated with varying degrees of impairment, while also having clinical potential to aid in identifying adolescents most at risk.
Kelsey Hudson– Doctoral Candidate, University of Vermont, Burlington, Vermont
Philip Spechler– University of Vermont
Lee Jollans– Trinity College Dublin
Bader Chaarani– University of Vermont
Scott Mackey– University of Vermont
Nicholas Allgaier– University of Vermont
Catherine Orr– University of Vermont
Matthew Albaugh– University of Vermont
Alexandra Potter– University of Vermont
Robert Althoff– University of Vermont
Richard Watts– University of Vermont
Robert Whelan– Trinity College Dublin
Hugh Garavan– University of Vermont