Something led my mind this morning to ponder why no-one has yet developed a screening instrument that mimics the way that clinicians use questioning to differentiate between ASD and "ASD-like symptoms resulting from something else". I think I was chewing over my evaluation interview again and thinking that this would surely not be too difficult(*), and would potentially reduce the load on diagnostic clinics caused by "false positives" in the queue, allowing signposting of these to more appropriate help earlier, enabling clinics to concentrate on unclear and complex cases with lots of co-morbid conditions that might be obscuring the diagnosis, and reducing waiting times overall.
I was also trying to reassure myself by saying "Surely if I scored 40 or so in the AQ50 I can relax a bit and expect a positive diagnosis, can't I?". Unfortunately it seems that this might not be the case.
Anyway, a quick google led me to this paper, which indicates that others have had this very thought "Finally, a novel scale more predictive of ASD diagnosis might be developed through a study of how clinicians discriminate ASD from ‘ASD-like’ symptoms."
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4988267/
Anyone into statistics will probably love this paper too!
(*) especially now that we have computers able to implement a huge battery of questions with many instances of "if you answered yes to Q12, go straight to Q243" without the person taking the test even knowing about it. You could even imagine some kind of Artificial Intelligence involved.
