Behavior problems in children with diabetes: disentangling possible scoring confounds on the Child Behavior Checklist.

OBJECTIVE The Child Behavior Checklist (CBCL; T. M. Achenbach, 1991), when used to assess the behavior of children with diabetes, may contain confounds because some behavioral items can have a physiologic etiology, and may skew reports of behavioral disturbance. METHODS Two techniques were used to disentangle possible scoring confounds in the behavioral ratings of children with and without diabetes: (1) the Somatic Complaints scale was deleted, or (2) Diabetes Items, identified a priori with 89% agreement by nine medical personnel, were deleted. RESULTS As expected, with traditionally scored protocols, children with diabetes obtained higher Internalizing and Total Behavior Problem scores than controls. This group difference persisted whether the Somatic Complaints scale or the Diabetes Items were deleted. CONCLUSIONS Compared to controls, children with diabetes obtained mildly elevated scores on six of the eight CBCL scales, regardless of scoring method, suggesting that their mildly elevated behavioral profile is not confounded by physiologic symptomatology.