ICD-10 and alternative diagnostic criteria for childhood autism among children with intellectual disability.

Abstract

The diagnosis of Childhood Autism (CA) among children with Intellectual Disability (ID) based on clinical criteria validated for populations with average intelligence compromises it's diagnostic accuracy in this special population. This study documents the diagnostic accuracy of ICD-10 and alternative criteria for CA in ID population. Consensus clinical diagnosis of autism made by a multi-disciplinary team as the reference data were extracted from the case notes and psychological tests details by two trainee psychologists, and ICD-10 based clinical diagnosis, demographics, ID, CA related data documented independently of the psychiatrist. The appropriate statistical analyses were completed. Repetitive behaviors formed the most common symptom cluster. The high internal consistency (κ = 0.75) among the three groups of ICD-10 symptoms indicate their usefulness in the diagnosis of CA among children with ID, but significantly more children with ID failed to meet the ICD-10 criteria for CA. The fourth alternative criteria had the highest concordance with the reference standard (100%) and the first alternative criteria had the highest agreement with the reference standard (κ = 0.88) in identifying CA among ID. The diagnostic accuracy of all the alternative criteria was significantly better than that of ICD-10 with the first alternative criteria having the best diagnostic accuracy (Sn = 98%; Sp = 100%; PPV = 1; NPV =0.83) closest to the reference standard. The symptom clusters in ICD-10 for CA should be retained, however reorganising the diagnostic criteria is required for diagnosing CA accurately among children with ID. The first alternative criteria can significantly improve the case identifying ability and diagnostic accuracy altering there by the epidemiological data on Childhood Autism among children with Intellectual Disability.