Summary: A new parent-report questionnaire can help detect signs of autism in children aged 18 to 30 months, though it does not identify every child who will later receive a diagnosis.
Source: University of Cambridge
New research led by the University of Cambridge shows that the Quantitative Checklist for Autism in Toddlers (Q-CHAT) can identify autism in some children aged 18–30 months, but it cannot detect all children who will later be diagnosed.
The findings are published in The BMJ Paediatrics Open.
Researchers at the Autism Research Centre in Cambridge carried out a prospective population screening study involving nearly 4,000 toddlers. Parents completed the Q-CHAT when their child was between 18 and 30 months old, and the children were reassessed at age four to evaluate diagnostic outcomes.
The Q-CHAT is an update of the original CHAT screening tool introduced by the same research team in the 1990s. It preserves core items from the original measure but expands them to assess early language development, repetitive and sensory behaviours, and broader social-communication traits.
The Q-CHAT contains 25 items with multiple response options for each, typically scored on a five-point frequency scale. This dimensional approach captures a range of behaviour rather than a simple yes/no answer, improving sensitivity to subtle differences in early development and enabling more nuanced screening across a range of early autistic traits.
In phase one of the study, 13,070 caregivers were invited to complete the Q-CHAT about their child; 3,770 returned the questionnaire. From those respondents, 121 children were invited for a detailed diagnostic assessment based on their scores and sampling strategy.
Phase two consisted of follow-up when the children reached four years of age. At that time the Childhood Autism Spectrum Test (CAST) and a checklist asking about referrals or diagnoses for developmental conditions were used. Children meeting thresholds on the CAST, those assessed in phase one, or those with parental concerns were invited for standardized diagnostic assessments using internationally recognized methods.
The study reports that the Q-CHAT had a sensitivity of 44% for predicting autism at age four, meaning it correctly identified 44% of the children who later met diagnostic criteria. Specificity was high at 98%, indicating most children without autism were correctly classified as not autistic. The positive predictive value (PPV)—the proportion of children who screened positive and were later confirmed autistic—was 28% after both phases of the study.
All 11 children who met diagnostic criteria for autism at phase one also scored at or above the Q-CHAT cut-point (39), demonstrating the tool’s ability to identify clear cases in toddlerhood. However, the Q-CHAT did not identify every child who was diagnosed at age four. The authors note that some children display less pronounced symptoms until later in childhood, which can delay detectable signs on toddler screens.
Previous work by the same team has shown that some autistic individuals are not diagnosed until adolescence or adulthood. Families and early supports can mask difficulties until social and academic demands increase, such as during transition to secondary education or adulthood, which can trigger recognition and referral.
Dr Carrie Allison, Director of Research Strategy at the Autism Research Centre and lead author of the study, said: “This research confirms that autism can be detected during the toddler years, while also highlighting that some children will only be identified later. Repeated screening and ongoing developmental surveillance may be more effective than relying on a single assessment at one time-point.”
Professor Tony Charman, Professor of Clinical Child Psychology at King’s College London and a study collaborator, added: “Screening in infancy can help fast-track children into early intervention services, which are associated with improved outcomes for many families. This study is notable because population-based toddler screening measures have rarely been evaluated with this level of methodological rigor.”
Professor Sir Simon Baron-Cohen, Director of the Autism Research Centre, commented: “Twenty-five years ago our team demonstrated that autism could be screened for and diagnosed as early as 18 months. The Q-CHAT represents an evolution of the original CHAT, providing a more sensitive instrument that helps identify young children who may benefit from diagnostic assessment and early support.”
About this autism research news
Source: University of Cambridge
Contact: Craig Brierley – University of Cambridge
Image: The image is in the public domain
Original Research: Open access. “Quantitative Checklist for Autism in Toddlers (Q-CHAT). A population screening study with follow-up: the case for multiple time-point screening for autism” by Allison, C., Matthews, F.E., Ruta, L., Pasco, G., Soufer, R., Brayne, C., Charman, T., & Baron-Cohen, S. BMJ Pediatrics Open
Abstract
Quantitative Checklist for Autism in Toddlers (Q-CHAT). A population screening study with follow-up: the case for multiple time-point screening for autism
Objective
This prospective population screening study evaluated the Q-CHAT for detecting autism in toddlers aged 18–30 months, with diagnostic follow-up at age four to assess accuracy over time.
Design
Observational, prospective study.
Setting
Luton, Bedfordshire, and Cambridgeshire, United Kingdom.
Participants
13,070 toddlers from the Child Health Surveillance Database (March 2008–April 2009) were eligible for follow-up at age four. Of these, 3,770 (29%) returned the Q-CHAT at 18–30 months and completed follow-up with the CAST at four years.
Interventions
A stratified sample across the Q-CHAT score distribution was invited for diagnostic assessment (phase 1). At age four, the follow-up used the CAST and a Checklist for Referral (CFR); children with CAST ≥15, those previously assessed, or those with developmental concerns on the CFR were invited for diagnostic assessment. Standardized diagnostic measures were applied at both time-points to evaluate Q-CHAT test accuracy.
Main outcome measures
Consensus diagnostic outcome at phase 1 and phase 2.
Results
In phase 1, 3,770 Q-CHATs were returned (29% response), and 121 children underwent diagnostic assessment; 11 met criteria for autism and all 11 had screened positive on the Q-CHAT. At a cut-point of 39 the PPV was 17% (95% CI 8% to 31%) in phase 1. At phase 2, 2,005 of 3,472 CASTs and CFRs were returned (58% response), and 159 children were assessed diagnostically (including 82 from phase 1). All children identified as autistic by the Q-CHAT in phase 1 continued to meet criteria at phase 2. The combined PPV after both phases was 28% (95% CI 15% to 46%).
Conclusions
The Q-CHAT can identify autism in some children aged 18–30 months, enabling earlier referral for diagnostic assessment. However, the relatively low positive predictive value means that several children who screen positive will not receive a later autism diagnosis. Follow-up screening and ongoing developmental surveillance at multiple time-points are recommended, as not all children who later meet diagnostic criteria are detectable in toddlerhood.