Semantically Rich Approach to Automating Regulations of Medical Devices

Abstract

Advanced medical devices increasingly use sophisticated AI/ML models to enable real-time analytics for monitoring patients. In the US, these AI models, which often form the underlying device software, are regulated by the Center for Devices & Radiological Health (CDRH) at the Food & Drug Administration (FDA) to ensure the safety & efficacy of the medical device. These regulations for medical devices are currently available as large textual documents, called Code of Federal Regulations (CFR) Title 21, that cross-reference other documents & so require substantial human effort to parse & comprehend. Hence, the device manufacturers incur significant costs during the regulatory process to adhere to all the rules & policies laid down by the FDA. We have developed a novel, semantically rich approach to extract the knowledge from the rules & policies for Medical devices & translate it into a machine-processable format that can be reasoned over. This framework was developed using AI/Knowledge Management approaches & Semantic Web technologies like OWL/RDF & SPARQL. This paper presents the detailed Ontology/Knowledge graph we developed for medical device regulations & the Use case results that validate our design. Regulators & manufacturers alike can use our framework to significantly reduce the human effort required during the device regulatory process.