Predicting autism traits from baby wellness records: A machine learning approach
Ayelet Ben-Sasson1, Joshua Guedalia1, Keren Ilan1
1University of Haifa, Israel.
Insights
Early autism spectrum condition (ASC) identification is crucial for intervention. A new model uses baby wellness visit records from the first two years to predict ASC likelihood, aiding early detection in boys and girls.
Area of Science:
- Developmental Pediatrics
- Computational Health
- Public Health
Background:
- Early identification of autism spectrum conditions (ASC) is vital for effective intervention.
- Routine health data offers potential for detecting early ASC indicators.
- Technological advancements facilitate novel approaches to developmental screening.
Purpose of the Study:
- To develop and test a predictive model for autism spectrum condition (ASC) likelihood using routinely collected baby wellness visit records.
- To assess the model's performance in identifying ASC in children within the first two years of life.
- To explore sex-specific patterns in early ASC detection using health records.
Main Methods:
- A predictive model was developed using electronic health records from baby wellness visits for children up to 2 years old.
- The study included a large cohort: 591,989 non-autistic children and 12,846 children diagnosed with ASC.
- Model performance was evaluated for identifying ASC, with analyses considering sex-specific differences.
Main Results:
- The model identified two-thirds of children with ASC (63% of boys, 66% of girls).
- Key predictive features included language, fine motor, and social milestones (12-24 months), maternal age, and growth patterns.
- Parental concerns about development or hearing were significant predictors, alongside variations in birth and growth parameters between models.
Conclusions:
- Routinely collected health data during the first two years of life can be leveraged to support early autism spectrum condition (ASC) detection.
- Developed models show promise in identifying early signs of ASC in both boys and girls.
- These findings support the integration of predictive analytics into standard pediatric care for timely ASC identification and intervention.
Lay Abstract:
Timely identification of autism spectrum conditions is a necessity to enable children to receive the most benefit from early interventions. Emerging technological advancements provide avenues for detecting subtle, early indicators of autism from routinely collected health information. This study tested a model that provides a likelihood score for autism diagnosis from baby wellness visit records collected during the first 2 years of life. It included records of 591,989 non-autistic children and 12,846 children with autism. The model identified two-thirds of the autism spectrum condition group (boys 63% and girls 66%). Sex-specific models had several predictive features in common. These included language development, fine motor skills, and social milestones from visits at 12-24 months, mother's age, and lower initial growth but higher last growth measurements. Parental concerns about development or hearing impairment were other predictors. The models differed in other growth measurements and birth parameters. These models can support the detection of early signs of autism in girls and boys by using information routinely recorded during the first 2 years of life.
Related Concept Videos
01:52Regression Toward the Mean
01:19Autism Spectrum Disorder
These core symptoms manifest differently among individuals, ranging from mild to severe. The disorder's complexity extends beyond its clinical presentation, encompassing a diverse range of biological, cognitive, and sociocultural influences.
01:18Steps in Outbreak Investigation