Construct and Predictive Validity of Modified Checklist for Autism in Toddlers, Revised with Follow-up (M-CHAT-R/F) Taiwan Version for Toddlers in Taiwan

Background:  Identification of ASD at an earliest age allows for provision of early intervention to facilitate the development of affected children. The M-CHAT-R/F is one of the screening tools which refers to the widespread screening of a population at risk for ASD and can be administered to all children in primary care settings. The M-CHAT-R/F is designed as a two-stage screener that consists of parent-report items and clinician-initiated follow-up questions. Our research group has translated the M-CHAT-R/F to Mandarin Chinese as a Taiwan Version. Preliminary results showed that the M-CHAT-R/F Taiwan version had acceptable reliability and discriminative validity for developmental screening in Taiwanese toddlers. However, no previous factor analysis of the M-CHAT-R/F has yet been published in the literature and little is known about the predictive validity for child’s developmental outcome.

Objectives:  The objectives are to 1) explore the factor structure of the M-CHAT-R/F Taiwan version, and to 2) examine predictive validity of the M-CHAT-R/F total score on developmental scores of the Mullen Scales of Early Learning (MSEL) at child’s 30 and 36 months of age.

Methods:  Toddlers aged 16-30 months were prospectively enrolled in this study. Children with major sensory, motor or neurological impairment/disorder were excluded. Parents filled out the M-CHAT-R/F at the beginning of the study and received developmental assessments by using the MSEL when child approached 30 and 36 months of age. Exploratory Factor Analysis (EFA) was conducted to investigate the factor structure of the M-CHAT-R/F Taiwan version. Linear regression analysis was used to examine the association between the M-CHAT-R/F total scores and MSEL developmental scores.

Results:  One hundreds and seven toddlers were recruited in this study. An Initial analysis showed the Kaiser-Meyer-Olkin statistic was 0.87 (above the threshold of 0.5) and the Bartlett’s Test of Sphericity was significant (p < .0001), indicating that the M-CHAT-R/F data were suitable for EFA. Five factors were derived accounting for 55.6% of the variance (Figure 1): (1) social attention and play (9 items); (2) communication (3 items); (3) interest and imitation (3 items); (4) behavior (3 items); and (5) motor (2 items). All the factor loadings of each item were > 0.3 on the derived factors (ranging 0.45 – 0.84). The results of predictive validity showed the M-CHAT-R/F total scores at child’s early ages were significantly associated with the MSEL scores at child’s 30 and 36 months of ages (β= 112.9, R²=0.209, p<0.001). Children who were classified as high risk or median risk for ASD in the M-CHAT-R/F scores were found to have significantly lower MSEL developmental scores at 30 and 36 months of age (p<0.05).

Conclusions:  This study showed a five-factor structure for the M-CHAT-R/F Taiwan Version. The finding of factor structure reflects the DSM-5’s diagnostic criteria for ASD, providing some evidence towards a similar factor structure of autistic traits and related symptoms for toddlers in Taiwan. Furthermore, the associations between the M-CHAT-R/F and MSEL suggests an acceptable predictive validity that the M-CHAT-R/F is predictive of child’s developmental outcome on the MSEL scales.