Good machine learning practice for medical device development

AI Summary

The U.S. Food and Drug Administration (FDA), Health Canada, and the UK Medicines and Healthcare products Regulatory Agency (MHRA) have jointly established ten guiding principles for Good Machine Learning Practice (GMLP) to oversee medical devices incorporating artificial intelligence and machine learning (AI/ML). This international regulatory framework focuses on ensuring the safety, efficacy, and quality of digital health technologies throughout their lifecycle. Key principles highlight the importance of utilizing multi-disciplinary expertise during development and adhering to rigorous software engineering and cybersecurity standards. A central requirement is that clinical data sets must accurately represent the target patient population to prevent bias and ensure equity. Developers must also maintain strict independence between training and testing data sets to ensure objective performance evaluation. The guidance emphasizes that AI models should be designed with the intended clinical use in mind, prioritizing the performance of the human-AI team rather than just the algorithm. Required actions for manufacturers include maintaining transparency by providing users with essential information about the device's functionality, performance, and limitations. Furthermore, companies must implement ongoing monitoring of deployed models to manage risks related to performance changes or model drift over time. This collaborative approach provides a foundational roadmap for developers to align with global expectations for high-quality medical AI development.