Predictive Value of Early Autism Detection Models Based on Electronic Health Record Data Collected Before Age 1 Year.

Researchers used computer models to predict autism using routine medical records from nearly 45,000 babies. The model could identify some children who would later be diagnosed with autism as early as 30 days old, with accuracy improving by their first birthday. While not perfect, this approach could help doctors spot autism earlier when combined with parent questionnaires, potentially leading to earlier support for families.

This retrospective study analyzed electronic health record (EHR) data from 45,080 children to develop predictive models for early autism detection. Using L2-regularized Cox proportional hazards models, researchers evaluated autism prediction accuracy from birth to 12 months. At 30 days, the model achieved 45.5% sensitivity and 23.0% positive predictive value at 90% specificity. By 12 months, performance improved to 59.8% sensitivity and 17.6% positive predictive value at 81.5% specificity.

The study demonstrates that routine healthcare data collected during infancy contains meaningful predictive signals for later autism diagnosis, suggesting potential for automated screening integration with existing caregiver surveys to enhance early detection accuracy.

Could enable passive autism screening integration into routine healthcare systems. May improve early detection when combined with caregiver surveys. Requires careful implementation to manage false positives and ensure equitable access across diverse populations.

Single health system data may limit generalizability. Retrospective design cannot establish causation. Unclear if models account for healthcare access disparities. Performance metrics suggest high false positive rates that could cause unnecessary anxiety for families.

Autism detection early in childhood is critical to ensure that autistic children and their families have access to early behavioral support. Early correlates of autism documented in electronic health records (EHRs) during routine care could allow passive, predictive model-based monitoring to improve the accuracy of early detection. To quantify the predictive value of early autism detection models based on EHR data collected before age 1 year. This retrospective diagnostic study used EHR data from children seen within the Duke University Health System before age 30 days between January 2006 and December 2020.

These data were used to train and evaluate L2-regularized Cox proportional hazards models predicting later autism diagnosis based on data collected from birth up to the time of prediction (ages 30-360 days). Statistical analyses were performed between August 1, 2020, and April 1, 2022. Prediction performance was quantified in terms of sensitivity, specificity, and positive predictive value (PPV) at clinically relevant model operating thresholds. Data from 45 080 children, including 924 (1.5%) meeting autism criteria, were included in this study.

Model-based autism detection at age 30 days achieved 45.5% sensitivity and 23.0% PPV at 90.0% specificity. Detection by age 360 days achieved 59.8% sensitivity and 17.6% PPV at 81.5% specificity and 38.8% sensitivity and 31.0% PPV at 94.3% specificity. In this diagnostic study of an autism screening test, EHR-based autism detection achieved clinically meaningful accuracy by age 30 days, improving by age 1 year. This automated approach could be integrated with caregiver surveys to improve the accuracy of early autism screening.