Lack of Correspondence Between Self- and Parent-Report on Structured Psychiatric Interviews of Adolescents with High-Functioning Autism Spectrum Disorders

Background:  The gold standard for diagnosing psychiatric disorders is to obtain information from multiple sources and achieve a consensus diagnosis. However, disagreement between sources is common and often a single source is used in research. Adolescent self-report results are frequently exclusively utilized over parent report in psychiatric research studies of typically-developing adolescents. This approach may not be best for even bright and verbal adolescents with ASD due to differences in emotional understanding, communication, and self-awareness that may impact their ability to self-report psychiatric symptoms. To date, studies of psychiatric comorbidity in ASD utilize a variety of methods for establishing diagnosis, and parent- and self-report on psychiatric interviews have not been directly compared to determine the degree of correspondence.

Objectives:  The study aimed to determine the degree of correspondence between self- and parent-report on a semi-structured psychiatric interview for adolescents with ASD.

Methods: Participants included 37 10 – 17 year old children with an ASD (confirmed with the ADOS and ADI-R) and without intellectual disability (mean FSIQ = 106). Current and lifetime comorbid psychiatric diagnoses were established via the Autism Comorbidity Interview, which is a modification of the Kiddie-Schedule for Affective Disorders and Schizophrenia. Lifetime and current diagnoses consistent with DSM-IV criteria were determined, as well as subsyndromal and subthreshold diagnoses which reflect milder variants.

Results:  Parent-report interviews resulted in substantially more current DSM-IV psychiatric diagnoses than self-report, with 42 diagnoses based on parent report compared to 8 diagnoses based on self-report. Diagnostic concordance was extremely low, with parent-child agreement for only 6 current DSM-IV diagnoses (14% of parent-reported diagnoses). When allowing for any level of diagnosis (e.g. collapsing subthreshold, subsyndromal, and DSM-IV) and not specifying a time frame (e.g. using lifetime scores), agreement improved, with parent-child agreement on 34% (33) of the 97 parent-reported diagnoses. For parent-reported major depression, results indicated 41% agreement (7/17) for any level lifetime diagnosis, which was the highest rate of agreement.

Conclusions: The results indicated very poor diagnostic agreement between parent- and self-report on a psychiatric interview. The findings were in stark contrast to research on typically-developing adolescents, revealing an opposite pattern of disagreement; namely, adolescent report among typically-developing populations results in higher rates of disorders than parent report, whereas the present findings revealed low to zero rates of disorders based on the self-report of adolescents with ASD despite high rates of psychiatric disorders based on parent-report.  Agreement for lifetime diagnoses at any level (e.g. subthreshold or higher) was better, suggesting that adolescents with ASD may be able to report on their psychiatric symptoms to a certain degree. Overall, the results imply caution should be exercised before dismissing concerns about a comorbid psychiatric disorder based on the adolescent’s self-report alone. Further, the best approach to research on psychiatric comorbidity in ASD may differ from typically-developing populations, in that parent report may be preferable to adolescent report data if only one source is utilized. However, obtaining information from multiple sources when determining a diagnosis is still preferable, particularly given the complex nature of differential diagnosis in ASD.