Structure of Psychopathology in Preschool-Aged Children in an Epidemiological and High-Risk Sample

• In Event: Persisting Effects of Foster Care as an Alternative to Institutional Care for Institutionalized Young Children

Abstract

Objective. Research on bifactor models of psychopathology in early childhood has been limited to community samples with limited longitudinal follow-up, restricting generalizability to understanding the structure of psychopathology across levels of risk and across the lifespan. We examined general and specific forms of psychopathology within two independent samples of preschool-aged Romanian children, including longitudinal outcomes within a high-risk sample of children exposed to early psychosocial deprivation.

Method. Study 1 consisted of 350 Romanian children recruited from pediatric clinics for an epidemiological study on young childhood psychopathology. Study 2 consisted of 158 young Romanian children from the Bucharest Early Intervention Project exposed to severe psychosocial deprivation and community comparison children. In both studies, young children’s psychopathology symptoms were assessed through the same caregiver-reported structured psychiatric interview (Study 1 M age 39.7 months, SD=10.9; Study 2 M age 55.6 months, SD=1.9). In Study 2, we examined prospective associations of general and specific psychopathology in young childhood with outcomes at ages 8 and 12 years, including a multi-domain composite of competent functioning (peer relations, family relations, academic performance, mental health impairment, physical health, substance use, and risk-taking behavior), full-scale IQ, and psychopathology. Prospective analyses controlled for young childhood total symptom count to isolate the effect of general and specific forms of psychopathology above a traditional symptom count.

Results. A bifactor model of psychopathology was supported in both studies, however, the structure of specific factors differed somewhat (Table 1). Study 1 had specific factors for externalizing symptoms, internalizing symptoms, and disturbed relatedness (signs of reactive attachment disorder [RAD] and disinhibited social engagement disorder [DSED]) whereas Study 2 had specific factors for externalizing symptoms, fear (symptoms of separation anxiety, specific phobia, and posttraumatic stress disorder), and disturbed relatedness. In Study 2, young children with higher levels on the general psychopathology factor had lower subsequent competent functioning and full-scale IQ and higher subsequent psychopathology symptoms at ages 8 and 12 years over and above total symptom count (Table 2). Further, children with higher levels on disturbed relatedness had lower subsequent competent functioning and higher psychopathology while covarying for total symptom count. Additionally, higher disturbed relatedness scores predicted subsequent internalizing symptoms over and above total symptom count. The externalizing and fear factors were not associated with measured outcomes.

Conclusion. We replicated the bifactor structure of psychopathology in early childhood in a community and high-risk sample, and extended understanding of outcomes of general and specific forms of psychopathology in middle childhood and adolescence in a high-risk sample. This is the first bifactor modeling study of psychopathology to include symptoms of RAD and DSED in the modeling, and to demonstrate their contribution to the general factor and to a disturbed relatedness factor. Both the general and disturbed relatedness factors were powerful early indicators of risk for poor developmental outcomes that yield information not obtained through traditional symptom counts. An important question is how bifactor models of psychopathology can be useful in clinical practice, as challenges remain regarding how to measure and implement general and specific domains in clinical practice.

Authors

• Katherine Leigh Guyon-Harris, University of Pittsburgh
  Presenting Author

• Kathryn Humphreys, Vanderbilt University
  Non-Presenting Author

• Mark Wade, University of Toronto
  Non-Presenting Author

• Mary Margaret Gleason, Children's Hospital of The King's Daughters
  Non-Presenting Author

• Florin Tibu, Universitatea de stat din Moldova
  Non-Presenting Author

• Charles A. Nelson, Harvard Medical School
  Non-Presenting Author

• Nathan A Fox, University of Maryland - College Park
  Non-Presenting Author

• Charles H Zeanah, Tulane University
  Non-Presenting Author