Signs of Autism Found on Chart Review of Screen-Negative Children with Autism

Background: Autism screening is widely employed to help pediatricians identify children at risk for autism, refer for prompt diagnostic evaluation, and start crucial early intervention. However, the mean age of diagnosis nationally is 52 months, approximately 2 years after autism screening. Screening instruments have estimated high sensitivity, but do miss some cases and can be falsely reassuring. Children who screen negative but go on to receive an autism diagnosis may have early signs of autism in routine medical and developmental history taken at well visits.

Objectives: To search medical records of children with autism who had false negative results on autism screening for behaviors or concerns associated with autism present prior to diagnosis in order to identify areas of development that should be added to the surveillance and screening process for autism.

Methods: Charts were reviewed within a large healthcare system for children who were 16-30 months old and screened for autism in 2013-2016. A cohort was identified who were screened with the Modified Checklist for Autism in Toddlers (M-CHAT) or the revised version (M-CHAT-R), and who had a later diagnosis of autism in their medical record based on the presence of an ICD code consistent with ASD. Charts of screen-negative children were first manually reviewed by autism experts to confirm autism diagnoses. Among those children with confirmed autism, provider notes were manually searched for documentation of early signs of autism and co-morbidities often associated with autism, and frequencies of these findings were calculated.

Results: 161 children with autism screened by the M-CHAT from 2013-2016 were identified. Of these, 99 (61%) screened positive on the M-CHAT and 62 (39%) children with confirmed autism screened negative. Among screen-negative cases, 66% had impairments noted in social communication, 60% had severe dysregulation of feeding or sleep or both, 48% had repetitive or restrictive play, 60% had abnormal sensory seeking or avoiding behaviors, 18% had a sibling with autism, 71% had a parental concern for autism, and in 61% the physician recorded a concern for autism. Overall, 79% of screen-negative children with autism had record of a parent or physician concern within the areas of social communication or repetitive/restricted play.

Conclusions: Most screen-negative children with autism had early markers of atypical development apparent in chart review prior to their autism diagnosis. A larger than expected proportion of children with autism had a negative screen, given previous studies of the M-CHAT showing high sensitivity. Notably, for most children studied, either the parent or the physician themselves reported a concern for autism. Despite signs of autism, a negative screen may be falsely reassuring against an autism diagnosis. Clinical signs of autism should prompt referral for further evaluation, and should not be overruled by a negative autism screen. Continued improvement of the screening process and screening instruments for autism are needed to improve early identification of children with autism.