Algorithm combining multiple digital phenotypes has high diagnostic accuracy, with area under the receiver operating characteristic curve of 0.90
By Elana Gotkine HealthDay Reporter
WEDNESDAY, Oct. 4, 2023 (HealthDay News) — An autism screening digital application (app) administered to young children can accurately detect autism, according to a study published online Oct. 2 in Nature Medicine.
Sam Perochon, from Duke University in Durham, North Carolina, and colleagues reported findings from a multiclinic prospective study assessing the accuracy of an autism screening digital app administered to 475 children aged 17 to 36 months during a pediatric well-child visit. Of the participants, 49 were diagnosed with autism and 98 were diagnosed with developmental delay without autism. Stimuli that elicited behavioral signs of autism were displayed by the app, which were then quantified using computer vision and machine learning.
The researchers found that an algorithm combining multiple digital phenotypes had high diagnostic accuracy, with an area under the receiver operating characteristic curve of 0.90, sensitivity and specificity of 87.8 and 80.8 percent, respectively, and negative and positive predictive values of 97.8 and 40.6 percent, respectively. Similar sensitivity performance was seen across subgroups defined by sex, race, and ethnicity.
“We conclude that quantitative, objective, and scalable digital phenotyping offers promise in increasing the accuracy of autism screening and reducing disparities in access to diagnosis and intervention, complementing existing autism screening questionnaires,” the authors write.
Several authors disclosed ties to the biopharmaceutical and medical technology industries; several authors developed technology related to the app that has been licensed by Apple.
Copyright © 2023 HealthDay. All rights reserved.