Onset Patterns: Correspondence Between Home Video and Parent Report

Background:  

The onset of behavioral signs of autism is thought to occur in two patterns, an early onset and a regressive course. Recently, significant concerns about this two-category onset classification have been raised.

Objectives:  

To compare two methods of defining onset, parent report and coded home video, and examine the number of categories that best describe onset.

Methods:  

Participants were children with Autistic Disorder (n=52) and typical development (n=23). Home movies from 6 through 24 months were collected (n=3199 segments). Four social-communication behaviors were coded: looks at people, smiles at people, language (e.g., simple babble, complex babble, words, and phrases), and joint attention. These were combined into a composite score reflecting social-communication frequency per minute. Individual trajectories of social-communication from 6 to 24 months were grouped into interpretable classes using latent class modeling.

 

Results:  

Bayesian Information Criteria were used to select the number of trajectories that best fit the data. There was strong support from coded home video for 3 onset trajectories. The first "early onset" trajectory (n = 20) displayed low rates of social-communication behavior at all ages. The second "regression" trajectory (n = 20) displayed high levels of social-communication behavior early in life and significantly declined over time. The third "plateau" trajectory (n = 12) was similar to the typical children early in life but did not progress as expected. There was no support for a mixed (early signs + regression) trajectory.

Conclusions:  

This work suggests that traditional two-group models for classifying onset are insufficient to describe all the patterns in which symptoms emerge. It also raises significant concerns about the methods used to classify onset for research purposes.

 The ADI-R was used to create parent-report onset types. Given the lack of support from the trajectory analyses for a mixed onset group, participants with this parent-reported pattern were collapsed with those whose parents reported regression alone. This resulted in three categories of parent-reported onset: PR_Early Onset (n = 18), PR_Regression (n = 24), and PR_Plateau (n = 10).

 There was poor correspondence between parent report and home video classifications (kappa = .11, p = .30). Only 9 of 20 participants whose home video displayed clear evidence of a major decline in social-communication behavior were reported to have had a regression by parents. Only 8 of 20 participants with evidence of early delays in social-communication on video were reported to demonstrate an early onset pattern by parents. Of the 10 whose parents described a plateau, only 3 had home video consistent with this pattern.

 Onset was also classified using parent report, based on the ADI-R. Four onset categories were defined: Early onset – symptoms present in the first year of life and no skill loss; Regression – no symptoms in the first year of life, followed by loss of skills; Plateau – no symptoms in the first year and no skill loss; Mixed – both symptoms in the first year and later skill loss. The correspondence between these categories and those derived from coded home movies was then examined.