Patterns of autism symptoms: hidden structure in the ADOS and ADI-R instruments

Jérémy Lefort-Besnard,Danilo Bzdok,Guillaume Dumas,Bertrand Thirion,Leonhard Schilbach,Kai Vogeley,Gaël Varoquaux

doi:10.1038/s41398-020-00946-8

Abstract

We simultaneously revisited the Autism Diagnostic Interview-Revised (ADI-R) and Autism Diagnostic Observation Schedule (ADOS) with a comprehensive data-analytics strategy. Here, the combination of pattern-analysis algorithms and extensive data resources (n = 266 patients aged 7–49 years) allowed identifying coherent clinical constellations in and across ADI-R and ADOS assessments widespread in clinical practice. Our clustering approach revealed low- and high-severity patient groups, as well as a group scoring high only in the ADI-R domains, providing quantitative contours for the widely assumed autism subtypes. Sparse regression approaches uncovered the most clinically predictive questionnaire domains. The social and communication domains of the ADI-R showed convincing performance to predict the patients’ symptom severity. Finally, we explored the relative importance of each of the ADI-R and ADOS domains conditioning on age, sex, and fluid IQ in our sample. The collective results suggest that (i) identifying autism subtypes and severity for a given individual may be most manifested in the ADI-R social and communication domains, (ii) the ADI-R might be a more appropriate tool to accurately capture symptom severity, and (iii) the ADOS domains were more relevant than the ADI-R domains to capture sex differences.

Highlights

Autism-spectrum disorder is clinically defined by different types of symptoms: deficits in communication and social interaction, as well as restricted, repetitive patterns of behavior, interests, or activities
Two clinical tools have become established in this assessment procedure, and are considered to be the “gold standard” in symptom evaluation for autism, when combined with clinical judgment: (i) the Autism Diagnostic Interview-Revised (ADI-R)[6], a semistructured interview conducted with parents, which focuses on current presentation and lifelong developmental history, and (ii) the Autism Diagnostic Observation Schedule (ADOS)[7], a standardized semistructured diagnosis assessment conducted through one-to-one personal interaction and direct observation of an individual suspected to have autism using a range of activities and delivered by a trained examiner
To explore distinct subgroups related to the ADI-R and ADOS-assessment patterns among patients with autism, we assigned each patient to one dominant symptom constellation based on the two instruments

Summary

Introduction

Autism-spectrum disorder (hereafter “autism”) is clinically defined by different types of symptoms: deficits in communication and social interaction, as well as restricted, repetitive patterns of behavior, interests, or activities. Whereas the differentiation between clinical subgroups was skipped, the differentiation between degrees of severity has been newly included. Two clinical tools have become established in this assessment procedure, and are considered to be the “gold standard” in symptom evaluation for autism, when combined with clinical judgment: (i) the Autism Diagnostic Interview-Revised (ADI-R)[6], a semistructured interview conducted with parents, which focuses on current presentation and lifelong developmental history, and (ii) the Autism Diagnostic Observation Schedule (ADOS)[7], a standardized semistructured diagnosis assessment conducted through one-to-one personal interaction and direct observation of an individual suspected to have autism using a range of activities and delivered by a trained examiner.

Methods

Results

Discussion

Conclusion