RESUMO
With the market development and demand change, the use of adaptive algorithms in medical devices has become a possible trend. However, some uncertainties in the adaptive algorithm itself will bring challenges to the existing current supervisory work model. This article focuses on the ademptions of US agencies on artificial intelligence device supervision, and discusses the problems that existing pilot policies may encounter when facing devices with adaptive algorithms. In this way, we will provide relevant suggestions and look forward to discussing with scholars.
Assuntos
Algoritmos , Inteligência ArtificialRESUMO
This article analyzes common problems and offers some suggestions on full life-cycle supervision and administration process of in vitro diagnostics, including premarket technical evaluations, pre-market QMS on-site inspections, postmarket adverse event surveillance, unannounced inspections and re-evaluation, from regulators' perspectives. It aims to provide technical reference for revisions on regulations, and help government administrators to figure out major points in daily regulatory practice.