The USPSTF's reviews investigate screening in such a fragmented, tunnel-visioned manner. I'd say the American Academy of Pediatrics (AAP) has a much better understanding of the "big picture" than the USPSTF -- where primary care providers (PCPs) are supposed to be using various screening tools that complement each other in a planned and periodic manner. The "big picture" of developmental-behavioral surveillance and screening in children 0 to 6 years http://archive.brookespublishing.com/documents/Bricker-screening-algorit... includes thoughtfully interpreting screens and skillfully discussing results face-to-face with parents plus promoting developmentally enhancing activities and positive parenting. These "big picture" activities make the screening process more safe and effective, but was not captured in the USPSTF's statement.
The USPSTF autism screening recommendation statement lacks applicability to to a busy primary care setting. Pre-visit screening tools makes well-child visits more efficient and effective. They allow PCPs to spend less time gathering information and more time talking about next steps. The majority of PCPs perceive that parents really appreciate having a face-to-face conversation about the results of parent-report screening tools.
Even though the USPSTF is mainly concluding that further research is needed, administering an autism-specific screen/test only "as needed" (at the end of the well-child visit) simply isn't a practical approach in a busy office setting. "As needed" screening/testing only leads to children with emerging autism spectrum disorder (ASD) getting a late diagnosis and an unacceptable lengthy delay in treatment. As a PCP myself, my experience has been that using the Ages & Stages Questionnaire, 3rd edition (ASQ-3) and Modified Checklist for Autism in Toddlers-Revised with Follow up Interview (MCHAT-R/F) universally at 18 and 24 months leads to a much higher percentage of my patients swiftly receiving early intervention (EI) services and those with emerging ASD generally get their diagnosis prior to 36 months of age.
"As needed" screening/testing also doesn't work because PCPs do not accurately and reliably elicit parents' concerns about their child's developmental-behavioral skills in an evidence-based manner. The research favors the wise recommendations of the AAP -- which, at ages 18 and 24 months, is to universally administer a broadband developmental screening questionnaire in combination with a 2-stage autism screening tool like the MCHAT-R/F.
At this point, the AAP should clarify to PCPs not to use the old MCHAT in isolation at 18 and 24 months (without concurrent administration of a broadband developmental screening tool like the ASQ-3 or PEDS). If PCPs universally screen with the MCHAT in isolation, then the large majority (~90%) of positive/concerning screens end up being toddlers who are not ultimately diagnosed with autism. That's troublesome. However, these same children frequently end up being diagnosed with an unspecified developmental delay and do benefit from EI services. For a "medium risk" result on the new MCHAT-R, the "Follow up Interview" is the key to sorting out which toddlers really need a ~$3000 ASD-specific evaluation with a developmental-behavioral pediatrician and team of pediatric sub-specialists.
Finally, EI services for treating ASD has been proven in multiple studies to optimize long-term outcomes and to reduce ASD-related costs for families and systems of care. Nevertheless, all experts would agree with the USPSTF about the need for more high-quality research that investigates the treatment of ASD in children under 30 months of age.