Medical Conditions During the First Year of Life Provide a New Way to Detect Autism Earlier

It is well known that autism is associated with a wide range of medical conditions, including sleep disturbances, epilepsy, and gastrointestinal (GI) problems, among others. Until recently, we didn’t know how early these medical problems start. In earlier work, Duke Center for Autism investigator Matthew Engelhard, MD, PhD, and his team provided answers to this question. Based on a retrospective analysis of the medical records of 29,931 patients seen at Duke, Engelhard and a team of center investigators found that, during the first year of life, infants later diagnosed with autism were three times more likely to be treated by a gastroenterologist, neurologist, or ophthalmologist, compared to those without autism. They also found evidence that infants later diagnosed with autism versus those diagnosed with ADHD showed different patterns of early medical problems.

Current studies are taking these new findings a step further. The team is exploring whether information from an infant’s medical record can reliably predict whether the infant will later be diagnosed with autism. To answer this question, the team examined the infant medical records of over 45,000 patients seen at Duke Primary Care clinics. Of these, approximately 900 were later diagnosed with autism. Using machine learning, an algorithm based on information in the medical record was able to predict which infants would later be diagnosed with autism by one year of age.

“The long-term goal is for this algorithm to automatically monitor infants receiving care and alert pediatricians when an infant has a higher likelihood of a diagnosis of autism. This information can be combined with specific recommendations for the pediatrician, such as how to refer the infant for further evaluation and services,” said Engelhard. “It would also draw the pediatrician’s attention to medical conditions that may warrant further treatment.”

By addressing medical conditions, such as sleep and GI problems, very early in life, it is hoped that children will be better able to benefit from behavioral interventions and have improved quality of life. This innovative work has led to new funding from the National Institutes of Health (NIH). Engelhard recently received an NIH K01 Mentored Research Scientist Career Development Award, which will provide 5 years of support and protected time for a mentored career development experience. His goal is to use analysis of the medical records to develop tools for prediction of autism, ADHD, and other neurodevelopmental conditions, such as language delay, and to explore how to incorporate the prediction algorithms into regular clinical care at Duke.