Early Identification of Autism in Spain: Exploring Diverse Possibilities with the MCHAT

Background:  The AAP has recommended autism screening for all children at 18 and 24 months (AAP, 2007). The MCHAT is the most well-studied, validated and useful screening tool for autism at early ages (Robins, Fein, Barton & Green, 2001) but more empirical data is needed. It has been confirmed the difficulties of the MCHAT in distinguishing ASDs from other DDs (Ventola et al 2007, Canal et al 2011) at the screening level. High rate of false positives (FP) can create unnecessary stress for parents and lead to expensive and time consuming diagnostic evaluations. It also has been suggested that MCHAT is better at detecting autism in children aged 24 months than those aged 18 months (Pandey 2008) .Therefore best age to instigate screening neither added value of screening at two time points are still conclusive.

Objectives: This study seeks to investigate further the possibilities of the use of the MCHAT as an efficient yet effective screening instrument to detect autism in community and at-risk populations. The main focus is to test and discuss the different options to administer the MCHAT identifying the most operational items and best cut-offs  in assorted samples in order to determine if the measure needs to be adapted accordingly.

Methods:   Parents of 18 months and/or 24 months aged children were asked to fill in the Spanish version of the MCHAT (Canal, 2011) at the outpatient health services (compulsory vaccination and well-child check-up programs respectively) in several health areas in Spain (Low Risk -LR-). For validation purposes, parents of children referred to early intervention services were also asked to fill out the MCHAT questionnaires when the child was aged 16-36 months old (High Risk –HR-). Analysis and data collection are ongoing. MCHAT data has been gathered from 7835 children: (7835 LR) 5270 at 18 months and 4819 at 24 months, including a subset of 2254 children screened 2 times and from 104 HR-children. Additional information on ADOS, V/NV-IQ and clinical diagnosis from LR-children with MCHAT positive results and from all HR children (regardless screening results) has been likewise recorded. MCHAT Best7 (Robins, IMFAR 2010&2011), MCHAT-R (Robins, IMFAR2011) and new potential scoring algorithms of the MCHAT, based on item analysis were examined in the different samples for psychometric properties.

Results: The preliminary findings suggests that the MCHAT used in HR population does not have the same sensitivity and specificity as noted in the general population regardless the scoring or algorithm used. Sen. & Sp are also different at 18 vs 24 m.o. (.57&0.90 Vs 0.71&0.92 respectively) within the same re-screened children. Aspects of pointing, joint attention and social reference on the MCHAT (ITEMS 6,7,9,23) seemed to be indicative of children subsequently diagnosed with ASD more than children with other DDs.

Conclusions: Efforts must continue to reduce the false positive rate without significantly increasing the number of missed cases. The analysis of the variability of the samples characteristics will determine the reliability of the findings. Improvement in screening will benefit parents, children and very likely health and early intervention resources.