Trajectories of Autism Severity in Children from Toddlerhood to Childhood in Taiwan: The Roles of Maternal Education Level and Residence Area

Background: Literature had shown that autism spectrum disorder (ASD) is characterized by heterogeneous severity. However, most of recent studies were done in the West, seldom in the East. As the authors know, this is the first longitudinal study for 6 years from toddlerhood to childhood in Taiwan, to explore the developmental trajectories of autism symptom severity and its early predictors.

Objectives: Two study purposes: (1) to plot longitudinal trajectories of ASD severity from early childhood to elementary school period; (2) to find the early indicator to explain the different trajectory classes in the children with ASD.

Methods: Seventy children with ASD participated at time 1 (mean chronological age was 2.5 years old), and then followed at time 2 (mean chronological age was 4 years old) and time 3 (mean chronological age was 8.5 years old). Participants recruited from two local areas in Taiwan, 33 of them were from Taipei city, the capital area in Taiwan, and 37 of them were from Chia-Yi city and county, the rural area in Taiwan. The study assessed the autism symptom severity by ADOS and cognitive function by MSEL WPPSI-IV, or WISC-IV at three time points. Additionally, early indicators including maternal education level, child’s FSIQ, VIQ, NVIQ at Time 1, and age of first autism sign in ADI-R. Due to different modules were used in ADOS across three time points, ADOS total CSS (calibrated severity score) was used for analysis.

Results: In hierarchical cluster analysis, three classes model was found to represent the observed data. We found that 27% (n = 18) of participants were assigned to persistent high severe class, 61% (n = 43) of participants were assigned to worsening class and 13% (n = 9) of participants were classified as improving class. The proportions of the three classes among two areas revealed that the children living in Chia-Yi were more likely to be classified as worsening group (n = 30, 81%) and no one be classified as improving group. Using multinomial logistic regression analysis, we found only two variables, i.e., maternal education level (p < .01) and area (p < .000) can explain the subgrouping independently. However, when the former two variables were put into the multinomial logistic regression model simultaneously; the effect of maternal education level was disappeared and the area variable (p < .000) was the only significant indicator to explain the subgrouping.

Conclusions: In summary, this study showed that all of the participants with ASD were classified as persistent high, worsening and improving groups with three time points from toddlerhood to elementary school ages. However, the children with ASD living in the rural area such as Chia-Yi were more likely to be classified as worsening group. Additionally, only the residence area could explain the ASD subgrouping. The accessibility and quality of service system of early intervention in two areas are discussed.