Background: Most psychosocial risk factors appear to have general rather than specific patterns of association with common childhood and adolescence disorders. However, previous research has typically failed to 1) control for comorbidity among disorders, 2) include a wide range of risk factors, and 3) examine sex by developmental stage effects on risk factor-disorder associations. This study tests the specificity of putative psychosocial risk factors while addressing these criticisms.
Methods: Eight waves of data from the Great Smoky Mountains Study (N = 1,420) were used, covering children in the community age 9-16 years old. Youth and one parent were interviewed up to seven times using the Child and Adolescent Psychiatric Assessment, providing a total of 6,674 pairs of interviews. A wide range of putative neighborhood, school, peer, family, and child risk factors, and common and comorbid youth disorders were assessed.
Results: The majority of putative risk factors were specific to one disorder or one disorder domain. A unique or 'signature set' of putative risk factors was identified for each disorder. Several putative risk factors were associated with a disorder in preadolescent males, preadolescent females, adolescent males, or adolescent females only.
Conclusions: Our findings support the need to define risk factors and disorders narrowly, to control comorbidity and other risk factors, and to consider developmental patterns of specificity by sex.