Optimizing a de novo artificial intelligence-based medical device under a predetermined change control plan: improved ability to detect or rule out asd in general pediatric settings

Canvas Dx is a software as a medical device (SAMD) diagnosis aid designed to support healthcare providers’ diagnosis of ASD in primary care settings. In a prospective, double-blinded validation trial (NCT04151290), Canvas Dx output was compared to the consensus specialist diagnosis. The device performed with high accuracy but abstained from producing an ASD positive or negative output in 68% of cases. Here we describe how the device's abstention rate was optimized under a predetermined change control plan.