Comparing Remote Diagnosis of ASD to Gold Standard, in-Person Assessment

Background: A telemedicine approach to diagnosis of ASD, the Naturalistic Observation Diagnostic Assessment (NODA) may accelerate the diagnostic process by connecting families with diagnosticians via the internet. Parents download a mobile application and complete a developmental history questionnaire. The NODA app then guides parents to record four 10-minute videos of their child in their home (i.e., mealtime, playtime with others, playtime alone, and parent concerns) and uploads them to a HIPAA-compliant, web-based platform. A diagnostician logs in to the web portal, reviews each video to identify specific behavioral examples of atypical or typical development.  Examples are “tagged” from predefined behavior descriptors informed by clinical expertise. Each tag is linked to a specific DSM-5 criterion for ASD. The diagnostician reviews the DSM-5 checklist and determines whether each criterion is satisfied based on the tags, the developmental history, and their own clinical judgment.   

Objectives: The goal of this study was to determine the percentage agreement for DSM-5 diagnosis of ASD between the remote method (NODA) and an in-person assessment utilizing gold-standard diagnostic procedures across three participant groups: 30 children referred for an ASD evaluation (ASD), 10 typically developing children (TD), and 10 children with other diagnoses (OTH). 

Methods: To date, 44 children have completed both the in-person assessment and the NODA procedure.  Of these 44, 30 were in the ASD group (23 males), 9 in the TYP group (6 males) and 5 in the OTH group (4 males). Participants were between 21 and 86 months (M = 50.27, SD = 16.33); there were no significant between group differences in age ( p = .21). The in-person diagnostic assessment included the ADI-R, ADOS-2, Vineland, and Mullen. Families returned home and completed the NODA procedure. A blinded clinician completed the NODA diagnostic assessment. For borderline cases (e.g., only one DSM-5 criterion unsatisfied), a second review was conducted.    

Results:  For each child, the diagnostic decision (ASD, not ASD) between the in-person assessment and the NODA assessment was compared. The same diagnostic decision was reached in 86.42% of the cases. There was perfect agreement in the TD and OTH groups and 80% agreement in the ASD group (24/30).  In 20% of ASD cases, NODA did not endorse ASD when the in-person assessment did.  However, there were sufficient criteria established within NODA to identify developmental delays. 

Conclusions: The results from this study demonstrate that NODA results in a valid diagnosis of ASD in 80% of cases. Approximately 20% of cases may still require an in-person evaluation.  NODA represents a key opportunity to increase timely diagnosis for most children with ASD, a critical first step to accessing appropriate early intervention programs. Further, NODA may help alleviate the demand placed on professionals who perform diagnostic assessments as part of their practice and reduce wait lists, thereby allowing families who need an in-person assessment to be seen faster. Additional analyses will explore whether child characteristics (e.g., intellectual disability, language impairment) impact diagnostic agreement, and whether there are differences in expression of autism symptoms in a home vs. a clinical setting.