Stability of Early Diagnoses and Symptom Presentation In Toddlers Referred for Autism Evaluation

Background:  The impact of early intervention on toddlers with ASD and emergence of developmentally appropriate, evidence-based methods (Dawson et al., 2010; Wetherby et al., 2006) have led to increased priority of early identification. Some research has shown diagnoses of very young children  to be largely stable (Chawarska et al., 2009; Kleinman et al., 2008; Lord et al., 2005) while other research indicates that symptoms change across the toddler and preschool years (Landa et al., 2006; Gamliel et al., 2009). However, these studies have primarily included children at high-risk and overrepresented low-functioning children. 

Objectives:  In order to address the need for research utilizing representative samples, the present study examined the short-term stability of diagnosis made by 24 months for toddlers referred for possible ASD in a prospectively identified sample. Secondly, this study examined child characteristics associated with change or stability of diagnosis.

Methods:  A preliminary study of 37 toddlers selected from the FIRST WORDS® Project, a screening of a general population sample, was conducted. Additional analyses will include 77 children with ASD or nonspectrum diagnoses. Children participated in two autism diagnostic evaluations before 42 months (preliminary sample: initial evaluation, M=19.35 months, SD=2.08; follow-up, M=34.71, SD=4.44). Best-estimate diagnosis was made by experienced clinicians using the Autism Diagnostic Observation Schedule (ADOS; Lord et al., 1999; Lord, Luyster, Gotham, & Guthrie, in press), standardized measures of developmental level and adaptive behavior, parent checklists, and home observation. A consensus clinical diagnosis was made by the one of three experienced diagnosticians using all information available, including initial best estimate diagnosis (made by the evaluating clinician and a licensed SLP), ADOS algorithm total, diagnostic report, and ADOS and home observation tapes. Diagnostic certainty was rated on a scale measuring certainty of current and predicted diagnostic symptom presentation.

Results:  The preliminary study of 37 toddlers indicated 100% stability for ASD diagnoses made at Time 1 (T1), as all 37 children were diagnosed with ASD at follow-up. Change in diagnostic symptoms was measured using ADOS classifications. The ADOS provides three classifications which reflect the number and intensity of symptoms. Classifications were stable for 68% of children, while classification worsened for 19%, and improved for 13%. Results indicated that age at T1, time between evaluations, and developmental skills as measured by the Mullen Scales of Early Learning (Mullen, 1995) at T1 did not differentiate between the stable, worsening, and improving groups. These groups were differentiated by clinician certainty at T1 in that clinicians were more certain about the stable group than 

Conclusions:  Findings support the increasing practice of diagnosing ASD in toddlers, as very young children were assigned diagnoses that remained stable from the second to the fourth year of life. Further, for some children, symptom presentation changed substantially indicating that marked improvement or worsening may occur within diagnostic stability. There is a clear need for increased understanding of factors associated with this change, in order to tailor intervention to a child’s likely developmental trajectory.