Indirect Comorbidity in Childhood and Adolescence

OBJECTIVE Comorbidity between psychiatric disorders is common, but pairwise associations between two disorders may be explained by the presence of other diagnoses that are associated with both disorders or "indirect" comorbidity. MATERIALS AND METHODS Comorbidities of common childhood psychiatric disorders were tested in three community samples of children ages 6-17 (8931 observations of 2965 subjects). Psychiatric disorder status in all three samples was assessed with the Child and Adolescent Psychiatric Assessment. Indirect comorbidity was defined as A-B associations that decreased from significance to non-significance after adjusting for other disorders. RESULTS All tested childhood psychiatric disorders were positively associated in bivariate analyses. After adjusting for comorbidities, many associations involving a behavioral disorder and an emotional disorder were attenuated suggesting indirect comorbidity. Generalized anxiety and depressive disorders displayed a very high level of overlap (adjusted OR = 37.9). All analyses were rerun with depressive disorders grouped with generalized anxiety disorder in a single "distress disorders" category. In these revised models, all associations between and emotional disorder and a behavior disorder met our criteria for indirect comorbidity except for the association of oppositional defiant disorder with distress disorders (OR = 11.3). Follow-up analyses suggested that the indirect associations were primarily accounted for by oppositional defiant disorder and the distress disorder category. There was little evidence of either sex differences or differences by developmental period. CONCLUSION After accounting for the overlap between depressive disorders with generalized anxiety disorder, direct comorbidity between emotional and behavioral disorders was uncommon. When there was evidence of indirect comorbidity, ODD, and distress disorders were the key intermediary diagnoses accounting for the apparent associations.