SAN FRANCISCO – Low-risk toddlers who screened positive on the revised version of a widely used autism screening tool were 114 times more likely to be diagnosed with autism spectrum disorder, compared with those who screened negative in a validation study of 15,612 children.
The Modified Checklist for Autism in Toddlers, Revised with Follow-up (M-CHAT-R/F) reduced the number of toddlers needing follow-up and increased detection of autism spectrum disorder, compared with the original checklist with follow-up (M-CHAT-F). The screen-positive rate decreased significantly from 9% with the M-CHAT-R to 7% with the M-CHAT-R/F. The rate of detecting autism spectrum disorder significantly increased from 45 to 67 cases per 10,000 children screened, Diana L. Robins, Ph.D., of Georgia State University, Atlanta, and her associates reported (Pediatrics 2014;133:37-45).
In the two-stage screening tool, parents first answer 20 yes-or-no questions, which typically takes less than 5 minutes. In the study, 92% scored 0-2 at this stage, indicating low risk with no further evaluation needed unless other risk factors are present. One percent of toddlers were considered high risk based on scores of 8 or higher, and all of these were later diagnosed with some kind of development delay or concern, which justifies immediate referral for formal evaluation and possible early intervention, the investigators reported.
Six percent of toddlers scored 3-7, indicating medium risk for autism spectrum disorder, and their parents were invited to complete a second stage of screening by answering structured follow-up questions to obtain more information, a process that takes 5-10 minutes by phone with a nurse, physician assistant, or other health professional. Approximately one-third of toddlers whose parents complete the follow-up continue to show risk for autism spectrum disorder on subsequent evaluation and deserve referral, the investigators said.
Autism spectrum disorder was diagnosed in 48% of all toddlers who screened positive, making them 114 times more likely than screen-negative children to be diagnosed with the disorder. Among the rest of those who screened positive, 36% had other developmental delays, 11% had developmental concerns but no formal diagnosis, and 5% were deemed to be typically developing.
Subjects were screened in Georgia and Connecticut during the children’s 18- or 24-month well-child care visits with pediatricians. The pediatricians also were asked to check a box at the top of the form if they had concern about autism spectrum disorder based on their clinical judgment. Among the 64 toddlers flagged by physicians, 30 were diagnosed with autism spectrum disorder and 12 had developmental delays or concerns. Pediatricians were more likely to indicate concern about possible autism in toddlers of highly educated parents, and the overall sensitivity of physician concern alone for detecting autism spectrum disorder was 24%.
Combining M-CHAT-R/F results and pediatricians’ clinical judgment provided a very high detection rate for autism spectrum disorder, "indicating that standardized screening in conjunction with routine developmental surveillance optimizes early detection for autism spectrum disorder," Dr. Robins reported.
She and her associates revised the original M-CHAT-F by eliminating three items (peek-a-boo, playing with toys, and wandering without purpose), reorganizing the remaining 20 items, simplifying the language, giving examples for developmental context and clarity, simplifying the scoring, and making other changes.
Toddlers with autism spectrum disorder in the study were diagnosed on average just after their second birthday, 2 years earlier than the median age of diagnosis in national data.
The M-CHAT-R/F can be downloaded for free in multiple languages from www.mchatscreen.com.
The investigators are studying the validity of using the M-CHAT-R/F in high-risk toddlers and looking at various ways to find cases of autism missed by the screening tool.
The National Institutes of Health funded the study. Dr. Robins and two of her coinvestigators are co-owners of M-CHAT LLC and receive royalties from the licensing of M-CHAT in electronic products, although her coinvestigators allocate all their royalties to research and clinical training expenditures. Three other coinvestigators reported having no financial disclosures.
sboschert@frontlinemedcom.com On Twitter @sherryboschert