A novel computational method that reliably predicts eventual autism spectrum disorder (ASD) diagnosis in young children has been developed by researchers at the University of Chicago. It eliminates the need for further bloodwork by only using diagnostic codes from previous doctor visits. Reportedly, it also reduces the instances of false-positive ASD diagnosis obtained by traditional screening techniques by half.
The study titled, ‘Reduced false positives in autism screening via digital biomarkers inferred from deep comorbidity patterns’ was published on Science Advances. It was undertaken by the University of Chicago’s developmental pediatricians Dr Michael E Msall and Dr Peter J Smith, in collaboration with the Zero Knowledge Discovery (ZeD) Lab led by Professor of Medicine Ishanu Chattopadhyay.
While a diagnosis of ASD can be obtained as early as two years of age, false positives produced by traditional screening tests can delay confirmed diagnosis. Since early intervention is of the essence and the number of trained professionals is limited, narrowing down the pool of patients by streamlining the diagnosis process can improve patient care by manifolds.
The researchers were able to reliably predict the diagnosis by using the sequences of ICD9 and ICD10 (International Classification of Diseases) codes generated during previous doctor visits (with the patient’s consent). This is achieved by a groundbreaking new algorithm that determines an autism comorbid risk score (ACoR) to successfully estimate the risk of an eventual ASD diagnosis.
Speaking of the newly developed diagnostic method, Dr Smith told Newswise, “A lot of what we have done is take the data and processes available in better-connected systems and apply them to less well-supported healthcare communities. This type of technology could overcome some of these structural barriers.”