Follow-up ASD Screening Identifies Children Missed at Initial Screening Timepoint

Background:  The M-CHAT-R/F is a valid screening tool for ASD at 18-30 months.  However, no screening instrument can identify all potential ASD cases. Children who screen negative at 18-24 months old but later screen positive at 30-48 months old with the M-CHAT-R/F are “potential miss” cases.  A longitudinal evaluation of the M-CHAT-R/F provides an opportunity to identify and study these children. 

Objectives: This study present preliminary data from a 2-year follow-up screening with the M-CHAT-R/F and examines items that were failed at both or either time points by “potential miss” cases. 

Methods: Participants were drawn from a sample of children who screened negative on the M-CHAT-R/F (n=10,504) and the M-CHAT (n=838) when screened at 18-24 months of age. Approximately 2 years after the initial screening, children were rescreened with the M-CHAT-R/F; it was mailed to families with a stamped return envelope. A subset of rescreened children screened positive on the M-CHAT-R/F and were invited to participate in an ASD evaluation. 

Results: At rescreening, 29 children were identified as “potential misses” and invited for diagnostic evaluation. Eleven of 17 evaluated children were given developmental diagnoses. Six children received an ASD diagnosis and were identified as “true misses.” Seven children received non-spectrum diagnoses, and four received no diagnosis. The average age of “potential misses” at initial screening was 21.6 months (SD=3.2), the average age at rescreening was 45.6 months (SD=5.6), and the average age at evaluation was 48.3 months (SD=3.2). “True misses” and “potential misses” did not differ on the number of items failed at initial screening (M=0.8, SD=0.9) or at rescreening (M=3.6, SD=1.7). At initial screening, failure on Item 13 (walks) and Item 17 (seeks parent attention) were each moderately associated with receiving an ASD diagnosis at 48 months (χ² (1, N=10)=1.7, p=.19, φ_c=.41), though sample size precluded significance. At rescreening, failure of Item 1 (follows point) (χ² (1, N=16)=1.6, p=.21, φ_c=.31), Item 3 (pretend play) (χ² (1, N=16)=1.7, p=.18, φ_c=.33), Item 5 (unusual finger movements) (χ² (1, N=16)=1.3, p=.25, φ_c=.29), and Item 15 (imitates) (χ² (1, N=16)=1.7, p=.18, φ_c=.33) were each moderately associated with a diagnosis of ASD at 48 months at a non-significant level.

Conclusions: This study presents preliminary data on a sample of screening cases that screened negative as toddlers but later screened positive on the M-CHAT-R/F, including some verified ASD “true misses.” 17 children were evaluated, 76% of whom received developmental diagnoses at 48 months; therefore, rescreening did not misidentify many typically developing children with potential concerns. At initial screening, failed items associated with ASD diagnosis were not core ASD features. This is potentially due to inaccuracies in parent reporting, or because core features manifested after the age of first screening. Interestingly, failed items at rescreening associated with ASD diagnosis represented core ASD features. These results emphasize the importance of ASD screening at multiple timepoints, as a single screening may miss cases. Further analyses will compare clinical characteristics and item-level responses in a larger sample of “potential miss” and “true miss” cases identified through additional screening measures.