False Positives to the M-CHAT In An Italian Population

Background:  

Early screening at a general population level lowers the age of first diagnosis for autism, but caution must be taken related to possible over referrals and unneeded worries for the families due to the false positives rate. Nonetheless, research has shown that the majority of false positives do not show fully typical development. Little is known though about specific developmental course that may differentiate these subjects prior to screening. A deeper understanding of conditions at birth and younger ages associated to requiring a second level of screening to exclude autism is also needed to improve the psychometric properties of screening tools. This is especially important for Italy where currently no data is available from screening programs.

Objectives:  

The objective of this investigation was to study the first systematic screening program in Italy by means of the FYI and the M-CHAT, currently translated and culturally adapted. The specific purpose of this study was to compare features of false positives and negatives to the M-CHAT with respect to conditions at birth (order of birth, weight at birth) and social-communication level at 12 and 18 months in an Italian population.

Methods:  

The study was conducted in Tuscany, Italy on a preliminary ongoing sample of 165 subjects during the pediatric surveillance visits. During the 12^th month visit the FYI and a ‘response to name’ task were administered; the M-CHAT was administered during the 18^th month visit. Subjects failing the M-CHAT were given a follow-up phone call to receive the M-CHAT Follow-Up Interview and the Socialization and Communication Scales of the Vineland Scales II. Study associations were completed through using multiple logistic regression.

Results:  

Toddlers with low weight at birth were significantly more likely to be false positive (OR= 1.52; CI 97%: 1.13 -2.04) while firstborns were found less likely to be false positive (OR= 0.34; IC 97%: 0.26-0.44). Higher scores on the Social Affective Engagement and the Repetitive Behavior Scales of the FYI were also associated with an increased risk of being false positive (OR = 1.36; CI 97%: 1.31-1.41 and OR = 1.14; CI 97%: 1.12-1.16 respectively). Moreover, the chance of passing the M-CHAT only at the Follow-Up level was more than three fold for subjects failing a “response to name” task at 12 months (OR = 4.95; CI 97%: 3.70-6.61). Comparison analysis of the false positives Vineland scores with population norms showed no differences with typical development for the majority of the subjects.

Conclusions:  

The current study suggests possible different developmental pathways for those subjects requiring a second level screening with the M-CHAT to exclude risk for autism. The significant associations of conditions of vulnerability at birth and developmental atipicalities at 12 months in presence of typical communication and socialization competence at 18 months is suggestive that false positives to the M-CHAT represent variations of typical development in this Italian sample. Further research is needed to study early features of false positives to implement multiple level screening processes in Italy and define a culturally sensitive screening protocol for the Italian context.