Prevalence of autism spectrum disorders among children (1-10 years of age): Findings of a midterm report from Northwest India.

Sir,

Raina et al.[1] found out in their interesting study that the prevalence of autism spectrum disorders (ASDs) among their studied pediatric population (1-10 years of age) was 0.9/1000. I presume that such prevalence ought to be cautiously interpreted. This is based on the presence of an important methodological limitation. Raina et al.[1] mentioned in the methodology that the participants were screened by trained field investigators utilizing the Hindi version of the Indian Scale for Assessment of Autism (ISAA). It is worthy to mention that ISAA has been considered to be a valid, accurate, and reliable tool for diagnosing ASDs and assessing the severity and disability among Indian children.[2] However, it has suboptimal validity in 3-9-year-old children. A recently published Indian study on the evaluation of ISAA in children aged 2-9 years at high risk of autism has shown that it had a sensitivity of 93.3; specificity of 97.4; positive and negative likelihood ratios of 85.7 and 98.7, respectively; and positive and negative predictive values of 35.5 and 0.08, respectively. Reliability was good and validity was suboptimal (r low, in 4/6 domains). The optimal threshold point demarcating “Autism” from “No autism” according to the Receiver Operating Characteristic curve was ISAA score of 70. The study concluded that ISAA could identify autism at a cutoff score of ≥70 and, thus certify disability of ≥40% in 3-9-year-old children.[3] Accordingly, I presume that there is a need to employ an assessment tool other than ISAA that could better determine ASDs prevalence. The ideal Indian diagnostic tool for ASDs requires accounting for variable literacy levels and heterogeneous culture and languages. It needs to be inexpensive, accurate, valid, reliable, easy to administer, and able to fulfill multiple purposes, including clinical (diagnosis, grading severity, planning intervention, and monitoring), research, and certification.[3] I presume that Autism Diagnostic Observation Schedule (ADOS)-Module 1 [ADOS (M1)] could be considered a better alternative ASD assessment tool as its cumulative diagnostic accuracy has been evaluated by a set of Indian researchers using the original diagnostic algorithm with meta-analysis and meta-regression. Tsheringla et al. addressed that the pooled sensitivity, specificity, diagnostic odds ratio (DOR), and summary receiver operating characteristic curves and its area under the curve (SROC-AUC) for the overall diagnostic accuracy of ADOS (M1) were: 0.91 (95% CI = 0.89-0.93), 0.73 (95% CI = 0.69-0.76), 44.20 (95% CI = 15.89-122.95), and 0.90, respectively. Moreover, the meta-regression analysis showed a nonsignificant relationship between ADOS (M1) and study quality as well as sample size. The researchers concluded that the ADOS (M1) with the original diagnostic algorithm had the overall diagnostic accuracy and pooled specificity suggesting moderate accuracy. The pooled sensitivity was high to be recommended as a screening test for ASDs.[4] I, therefore, presume that conducting a community based study utilizing ADOS (M1) could yield a better idea on the actual ASDs prevalence among Indian pediatric population.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.