Expert Clinician Certainty in Diagnosing Autism Spectrum Disorder in 16-30-Month-Olds: A Multi-site Trial Secondary Analysis

Abstract

Differential diagnosis of young children with suspected autism spectrum disorder (ASD) is challenging, and clinician uncertainty about a child's diagnosis may contribute to misdiagnosis and subsequent delays in access to early treatment. The current study was designed to replicate and expand a recent report in this Journal (McDonnell et al. in J Autism Dev Disord 49:1391-1401, https://doi.org/10.1080/15374416.2020.1823850 , 2019), in which only 60% of diagnoses were made with complete certainty by clinicians evaluating 478 toddlers and preschool children referred for possible ASD to specialized clinics. In this study, secondary analyses were performed on diagnostic, demographic and clinical data for 496 16-30-month-old children who were consecutive referrals to a 6-site clinical trial executed by specialized centers with experienced clinicians following best-practice procedures for the diagnosis of ASD. Overall, 70.2% of diagnoses were made with complete certainty. The most important factor associated with clinician uncertainty was mid-level autism-related symptomatology. Mid-level verbal age equivalents were also associated with clinician uncertainty, but measures of symptomatology were stronger predictors. None of the socio-demographic variables, including sex of the child, was significantly associated with clinician certainty. Close to one third of early diagnoses of ASD are made with a degree of uncertainty. The delineation of specific ranges on the ADOS-2 most likely to result in clinician uncertainty identified in this study may provide an opportunity to reduce random subjectivity in diagnostic decision-making via calibration of young-child diagnostic thresholds based on later-age longitudinal diagnostic outcome data, and via standardization of decision-making in regard to clinical scenarios frequently encountered by clinicians.

Keywords: Autism spectrum disorder; Diagnostic certainty; Differential diagnosis; Toddlers.

Fig. 1

Density Plots for ADOS-2 subscores—Social Affect (SA) and Restricted & Repetitive Behavior (RRB) and Mullen Verbal and Nonverbal Age Equivalents—by Expert Clinician Diagnosis (ECD) Certain vs Uncertain Cases^a (Density of 1.0 = 100% of the sample). Curves represent the number of Certain and of Uncertain Cases (in percentage) with given ADOS-2 and Mullen scores

Fig. 2

Natural cubic splines model estimates for the probabilities of diagnostic certainty over increasing values of ADOS-2—Social Affect (SA) and Restricted & Repetitive Behavior (RRB)– and Mullen sub-scores—Verbal (VAE) and Nonverbal (NVAE) Age Equivalents