A study of tens of thousands of children suggests that autism can be diagnosed from routine medical information collected within the first month of life.
Nearly half of children subsequently diagnosed with autism were identifiable from information recorded in electronic health records (EHRs) within 30 days of life, according to results published in . JAMA network open.
The EHR includes data on low birth weight, prematurity, low Apgar scores and other perinatal complications, as well as self-reports such as postnatal hyperbilirubinemia, respiratory infections, sleep, crying and feeding problems. Can include data about autism-related conditions.
Accuracy of EHR diagnosis was comparable to caregiver surveys collected at 18–24 months of age and further improved when 360 days of data were included.
“The results of this diagnostic study, conducted in a large academic medical center, suggest that EHR-based autism prediction reaches clinically meaningful levels of accuracy as early as 30 days of age. reports Matthew Engelhard, Ph.D., of Duke University School of Medicine. and colleagues.
“We observed that almost half (45.5%) of children with autism could be identified in 30 days while maintaining high specificity (90.0%).”
Our findings suggest that combining EHR information with caregiver surveys may improve the accuracy of early autism screening and enable families to receive more timely behavioral support. increase.
Noting that early detection of autism is associated with improved outcomes, researchers found that between 2006 and 2020, within the Duke University health system, 45,080 infants seen before age 30 had A retrospective diagnostic study was performed using EHR data from children.
Subsequent autism spectrum disorders and other neurodevelopmental disorders were identified using claim codes.
The team used 60% of the data to train and fine-tune an EHR-based autism detection model, and deployed the remaining 40% as a test set to evaluate performance.
Overall, 18,032 children were randomly assigned to the test set, including 363 autistic patients and 3,721 control participants who met criteria for other neurodevelopmental conditions. A total of 3,615 control participants were followed up to her eight years of age and included when calculating key performance indicators.
The EHR-based autism detection model achieved a positive predictive value of 23.0% with a sensitivity of 45.5% and a specificity of 90.0% at 30 days of age.
This improved to a positive predictive value of 17.6% at 360 days with a sensitivity of 59.8% and a specificity of 81.5%, and a PPV of 38.8% with a specificity of 94.3% with a sensitivity of 38.8% and a PPV of 31.0%.
Researchers find EHR-based tool provides information on probable autism and complements commonly used Infant Autism Modified Checklist (M-CHAT) screening tool I am paying attention to that.
However, they add: “In contrast to existing screening tools such as M-CHAT, EHR-based detection of autism occurs at a much earlier age (>30 days) and is completely passive, so data collection is Not needed, except what is already done during routine care.”
The team concludes: “The results suggest that EHR-based monitoring should be integrated with his M-CHAT, other caregiver surveys, and other screening tools to improve the accuracy of early autism screening.” increase.”
