Comparing the Indian Autism Screening Questionnaire (IASQ) and the Indian Scale for Assessment of Autism (ISAA) with the Childhood Autism Rating Scale-Second Edition (CARS2) in Indian settings

The Indian Autism Screening Questionnaire (IASQ), derived from the Indian Scale for Assessment of Autism ISAA (the mandated tool for autism in India), is an autism screening instrument for use in the general population by minimally trained workers. While ISAA has 40 items with four anchor points, the IASQ is a 10-item questionnaire with yes/no answers. It was initially validated using the ISAA. During its development the ISAA was itself compared to the Childhood Autism Rating Scale version 1 (ISAA Manual). In the present study, we evaluated both the ISAA and the IASQ in relation to the Childhood Autism Rating Scale version 2 (CARS-2). Methods: Participants were recruited from three settings: a referral clinic for neurodevelopmental conditions run by the Department of Paediatrics of a tertiary care teaching hospital (NDC OPD), the outpatient department of an institute for disability and rehabilitation (NIEPID), and from the community (CGOC). Persons between ages 3-18 were recruited following consent or assent (parent and child/adolescent). The IASQ was administered by a minimally trained administrator. It was followed by ISAA and the CARS-2 (in alternating order, by different evaluators blind to each other) (CARS2 SV (Standard Version) and CARS2 HF (High Functioning) as applicable). Sensitivity, specificity and area under the Receiver Operator Characteristics (ROC) curve were calculated for IASQ and CARS2, as well as for ISAA and CARS2. Concordance between CARS2 and ISAA was calculated using kappa coefficient. Results: A total of 285 participants (NIEPD n = 124; NDC OPD, n = 4; CGOC n = 157) (a total of 70 with autism and 215 controls) participated. IASQ and CARS2 were administered on 285 participants, while IASQ and ISAA were administered on 264 participants. When IASQ was compared to CARS2, sensitivity was 97%, specificity 81%, PPV 63%, NPV 99% at cut off 1 while these values were 97%, 92%, 79% and 99% respectively at cut off 2. There was high concordance between CARS2 and ISAA (Kappa 0.907, p<0.0001). Conclusions: IASQ has satisfactory sensitivity, specificity and concordance when compared with CARS2; it can be used for screening children with autism in community. The ISAA also showed a high concordance with CARS2, as it had with the older version of CARS.