Autism is prevalent (1 in 50 children) and impairing but early intervention has been shown to improve outcomes prompting recommendations for screening at both two year and 18 month health check-ups. Detection of autism early is crucial but accurate screening tools have been elusive. In Phase II, a practical method of improving accuracy of M-CHAT, the most widely used autism screen, was created by adapting the validated and recommended follow up interview, M-CHAT F/Ui (F/U), for efficient use during the visit. In the case of a positive screen, the primary care provider (PCP) can use CHADIS, a web-based questionnaire delivery, decision support and post-visit engagement system, to conduct F/U rather than requiring a visit with another professional. Phase II data showed F/U results by a PCP were equivalent to the autism center. Phase-II data on the M-CHAT plus F/U via CHADIS replicated findings of the F/U authors in showing both excellent positive predictive value (PPV) 0.96 for children >20 months (thus effective for 24 month olds) and also low PPV 0.54 for <20 months. Promisingly, Phase II exploratory analyses using a decision tree including supplementary data from a routinely used standard language screen (ASQ communication scale) and an item from a language measure (MCDI) plus the standard autism screen (M-CHAT plus F/U) reached PPV of 0.95 in the <20 month group. This screen completion could be done efficiently online by parents. Phase IIB plans a replication of this screening procedure which promises to be accurate for 18 month olds and comparison to alternatives using the community network of >400 Maryland doctors where >22,000 autism screens have been done using the CHADIS system. A more accurate screening test is of less value if it is not universally used. In Phase II, an approach was developed that reduces disparities in access to screening using a "talking" tablet kiosk that was preferred by parents to alternatives, but further workflow issues will be addressed in Phase IIB. To improve screen completion, we will program an automated reminder/completion confirmation text/email system in CHADIS with coupons as incentives for parents. In Phase II, CHADIS was adapted to capture both patient input and doctor decision-making and Maintenance of Certification (MOC) accreditation was awarded by the American Board of Pediatrics for this program and earned by 140 pediatricians. This monthly doctor quality improvement program will be extended to a daily continuous quality improvement process for the whole office team to further assure patient participation in screening while providing other clinical and financial analytics of value to the office. The Phase-IIB goal is to develop and test an innovative "screening system" not just a new test.
Autism is prevalent (1 in 50 children) and impairing but early intervention has been shown to improve outcomes prompting recommendations for screening at both two year and 18 month health check-ups. In Phase I, a standard screening procedure was adapted so that doctors could efficiently identify autistic children at two years of age but the screen did not prove accurate at 18 months. Phase II data showed that a novel method combining supplemental information and a new computerized scoring pathway promised accurate detection at the critical 18 month age. This Phase 2B proposal is aimed at replicating this unique approach to early autism identification in a community sample, comparing it with alternatives plus developing innovative strategies to assure that all children are screened by automating reminders and completion confirmation and by supporting continuous quality improvement in office implementation using automated reporting analytics.