"Comorbidity" in an epidemiological sample: a longitudinal perspective.

Parents' CBCL ratings of problem behaviors in 1117 4- to 11-year-olds from the general population were obtained across a 6-year interval. Data were analyzed categorically as well as quantitatively. Positive association, or overlap, was found between the majority of CBCL problem scales. Most CBCL scales were able to predict non-corresponding scales across the 6-year interval. Children who were scored in the deviant range on one particular syndrome but low on another syndrome had a better prognosis than children who were scored high on both syndromes. Studies concerning the etiology or treatment of certain conditions should control for co-occurring conditions because they may influence the results in unknown ways.