Evaluation of the Aberrant Behavior Checklist for Developmental and Epileptic Encephalopathies

Abstract

OBJECTIVES\nTo determine the suitability of the Aberrant Behavior Checklist (ABC)-a common measure used in clinical trials for treatment of challenging behaviors of autism-as an outcome measure for pharmacological and behavioral interventions for young people with Developmental and Epileptic Encephalopathies (DEEs).\n\n\nMETHODS\nWe assessed score profiles on the ABC in a sample of 122 young people with DEEs, including Dravet and Lennox-Gastaut syndromes, and KCNQ2- SCN2A-, and KCNB1-associated disorders. Then we examined its internal structure using item cluster analysis. We used both unrestricted item cluster analysis to determine the number of item clusters that maximize reliability and restricted analyses in which we pre-specified models with 5-, 6-, and 7-clusters, to examine consistency with previous factor analytic studies. We also conducted validity analysis on the various scoring methods with age, sex, and autism spectrum screening measure scores.\n\n\nRESULTS\nUnrestricted item cluster analysis suggested that three clusters maximized reliability of ABC scores. These broadly represented other-directed behaviors (i.e., externalizing ), self-directed behaviors (i.e., internalizing ), and inappropriate speech. Restricted models separated item clusters for stereotypy from other self-directed problem behaviors, and self-injurious behaviors from the other externalizing behaviors. Validity analysis also supported these structures. Overall, all scores were low, and less than 20% of DEE participants had symptoms severe enough to qualify for most randomized trials of behavioral therapies.\n\n\nSIGNIFICANCE\nThese results are broadly consistent with the extant ABC scoring algorithms. They suggest a high internal consistency reliability, which may support the use of the ABC in future clinical trials in patients with DEEs who exhibit the behaviors assessed by the ABC. Alternatively, concerns about overall low scores raise cautions about using the ABC as a measure of behavior in unselected populations with DEE.