M-CHAT Screening in Toddlers Referred to Early Intervention

Background:  

Early detection of autism spectrum disorder (ASD) is a priority to facilitate early intervention. The Childhood Autism Team CHeck (CATCH) Team is a cross-system specialized screening and assessment process in Chester County, PA for children in Early Intervention. Children are referred to the CATCH team if they screened positive on the Modified Checklist for Autism in Toddlers (M-CHAT) at 18-36 months and/or if the early intervention providers and parents express concerns about ASD. Referral to the CATCH team (including developmental pediatrics, school psychologist, and behavioral health professionals) results in a diagnostic evaluation.  

Objectives:  

The objectives of the current study were to (1) evaluate the utility of the M-CHAT to detect ASD in a community-based high-risk sample, (2) compare M-CHAT scores to final diagnosis, and (3) examine the preliminary psychometric properties of the M-CHAT among children referred to the CATCH team for ASD assessment.

Methods:  

Archival data (2005-2014) from CATCH included 246 toddlers who screened positive on the M-CHAT and 12 who screened negative but were referred by a provider due to concerns of ASD. From this sample, 197 were diagnosed with ASD (79.8% male; mean age=2.20 years, SD=.51, range: 1.10-2.90) and 61 were diagnosed with other developmental delays (69.6% male; mean age=2.32 years, SD=.45, range: 1.11-2.90). Diagnoses in the non-ASD group included language delay (n=40), global developmental delay (n=19), and other (n=2, separation anxiety, behavior concern). Evaluation tools included ADOS(2), CARS(2), medical history and parent interview to inform a DSM(-IV/5) checklist. Data were extracted from clinical evaluation records. M-CHAT scores included total and/or critical score (note: some children had only one score available); final diagnosis was a clinical best estimate considering all available data.

Results:  

Total M-CHAT score was not significantly different between ASD (mean=7.13, SD=3.8) and non-ASD (mean=6.50, SD=3.7), t(225)=-1.50, p=.294. However, critical score was significantly higher for the ASD group (mean=3.30, SD=1.6) compared to the non-ASD group (mean=2.32 SD=1.4, t(108)=-2.89, p=.005. The M-CHAT was positive (based on either total or critical score) for 190 (96.4%) of the ASD cases, indicating high sensitivity among those children referred for specialized ASD assessment by the CATCH team. Similar to other studies, the M-CHAT was not specific to ASD, but in the current sample, positive predictive value (PPV) was .77 indicating that positive screens were much more likely to be diagnosed with ASD than in low-risk samples (e.g., .06 PPV for the single-stage M-CHAT in Chlebowski et al., 2013).

Conclusions:  

ASD screening tools may perform differently in high-risk samples compared to low-risk samples.  Examining utility of ASD screeners in community-based settings facilitates the implementation of research tools in real-world settings. In a high-risk sample of toddlers, the M-CHAT critical score may be more useful in differentiating ASD from non-ASD delays, rather than the total score. PPV indicated high confidence that positive screens are likely to have ASD. Findings from the current study corroborate that standardized screening paired with referrals for diagnostic evaluation can facilitate identification of young children with ASD.