Question 1: Does your child ever point with their index finger to ask for something?
Question 2: Is your child able to imitate you or your actions, for example if you pull a face?
Question 3: Does your child ever use pretend play, for example to talk on a phone or take care of a doll?
Question 4: Does your child look at something across a room when you point to it?
Question 5: Does your child understand what people say?
Question 6: Does your child ever bring an object to you to show you something?
The paper from Yoko Kamio and colleagues  (open-access) suggests that these 6 questions taken from the 23-item M-CHAT Japanese version (JV) might have the methodological strength to screen for possible autism in toddlers -- at least in Japan. M-CHAT by the way, is one of the instruments of choice when it comes to screening for possible autism and has seen some developments in recent times (see here).
Based on data derived from "two prospective community cohorts in Japan, Fukuoka (cohort 1) and Tokyo (cohort 2)" cumulatively including some 2500 children "who received health check-ups when aged 18 months", researchers analysed data using a model of discriminant function analysis based on groupings of those who were eventually diagnosed with an autism spectrum disorder (ASD) compared with those who weren't. They concluded that their study "identified a highly discriminative 6-item set from the 23-item M-CHAT-JV and demonstrated its reliability and validity with cohort data from 2 geographically different regions in Japan." The results, I might add, were not 100% reliable in terms of the 6-item screening method used, but this is real life and, as far as I am aware, we don't have a perfectly reliable autism screen at the moment.
These are interesting results as a function of the important autism science on the best way to 'red flag' autism in its very earliest days (see here). Indeed, the focus on social-communicative functions (including pointing) follows a trend in the peer-reviewed research literature in this area, as something to focus on when it comes to early screening for autism. I say this bearing in mind that within the very heterogeneous label of autism, there are cases of regression into autism at a time later than 18 months.
"Considering the tight time constraints in primary care settings, a brief screening tool might be helpful in facilitating the integration of autism-specific screening within routine general developmental screening." These are noble sentiments from the authors and kinda accords with some increasing moves in autism practice to make things more streamlined in these resource-austere times that we live in (see here). Obviously we await further research in this area on whether the Kamio findings cross cultures and geographies with other infant cohorts or not. If they do however, combined with the rise and rise of telemedicine for example, the days of the [often] long and expensive autism screening and diagnosis process might be numbered. Oh, and screening might just start to be interactive too  (see here for more information on the RITA-T).
Music: iLL BLU - Lonely People ft. James Morrison.
 Kamio Y. et al. Brief Report: Best Discriminators for Identifying Children with Autism Spectrum Disorder at an 18-Month Health Check-Up in Japan. Journal of Autism and Developmental Disorders. 2015. July 19.
 Choueiri R. & Wagner S. A New Interactive Screening Test for Autism Spectrum Disorders in Toddlers. J Pediatr. 2015 Aug;167(2):460-466.
Kamio, Y., Haraguchi, H., Stickley, A., Ogino, K., Ishitobi, M., & Takahashi, H. (2015). Brief Report: Best Discriminators for Identifying Children with Autism Spectrum Disorder at an 18-Month Health Check-Up in Japan Journal of Autism and Developmental Disorders DOI: 10.1007/s10803-015-2527-1