The Predictive Value of the M-CHAT for ASD Screening Among Preterm Infants

Background: The prevalence of Autism Spectrum Disorder (ASD) is 1% in the general population, whereas among preterm born children it is estimated as 1%-8%. Therefore, special attention has been given to early identification of social-communication deficiencies and ASD-symptoms in this high-risk population.

The Modified Checklist for Autism in Toddlers (M-CHAT) is a widely used screening instrument for ASD. The rate of positive screening for risk for ASD among preterm born children is 20%-41%, and is significantly higher compared to that of full-term infants.

Abnormal scores on the M-CHAT correlate with lower socio-economic status, male gender, low birth weight and gestational age, as well as with sensory, motor, social-communication, cognitive and emotional impairments. Given the association among the M-CHAT and the aforementioned characteristics, the predictive value of the M-CHAT should be carefully assessed utilizing gold standard measures.

Objectives:  To examine the sensitivity and specificity of the M-CHAT administered at 18 months, concurrently at 18 months and predictively at 36 months among preterm born children.

Methods: One hundred and ten preterm infants (Demographic characteristics are presented in Table 1) participated in this longitudinal study. The Mullen Scales of Early Learning, the M-CHAT and the Autism Diagnostic Observation Schedule (ADOS) were administered at 18 and 36 months.

Among 110 infants in our sample, 3 infants were excluded from the analyses because of severe neurological impairments. Eighty-nine families completed the assessment at 18 months, and 81 completed the assessment at 36 months.

Results: Applying the recommended criteria for the M-CHAT at 18 months, 29 infants (32.6%) screened positive for risk for ASD whereas only 8 infants (9.0%) were identified at risk for ASD using the ADOS. At 36 months, only 2 (2.5%) infants passed the cutoff for ASD diagnosis using the ADOS.

The sensitivity of the M-CHAT conducted at the 18 months in comparison to the ADOS administered concurrently was 62.5%, and the specificity was 70.4%. Assessing the predictive value of the M-CHAT administered at 18 months in comparison to the ADOS administered at 36 months, the sensitivity was 100.0%, and the specificity was 69.6%. The sensitivity of the ADOS administered at the 18 months in comparison to the ADOS administered at 36 months was 100.0%, and the specificity was 93.7% (Table 2).

One-Way ANOVAs indicated that infants who screened positive for ASD by the M-CHAT at 18 months were more likely to be male, from lower income families, and had lower receptive and expressive language abilities at 18 months compared to infants who screened negative. At 36 months, infants who screened positive for ASD at 18 months, exhibited significantly lower fine-motor skills and receptive language abilities and a higher (worse) ADOS total score.

Conclusions: In line with previous reports, positive screening for risk for ASD by the M-CHAT was associated with characteristics not specific to ASD. Given the importance of early intervention, the M-CHAT may be useful in identifying infants who can benefit from intervention yet clinicians should be careful in suggesting risk for ASD, especially among high-risk groups such as preterm infants.