On Wednesday, May 10, 2023, the Food and Drug Administration (FDA) announced the publication of a new discussion paper titled “Using Artificial Intelligence and Machine Learning in Drug and Biologics Development.” This discussion paper is intended to facilitate discussion. Discuss with stakeholders the use of artificial intelligence and machine learning (AI/ML) in drug development, including the development of medical devices intended for pharmaceutical use. This discussion paper covers his three main topics: ML; considerations for using AI/ML. And then the next steps and stakeholder engagement.
Here are 5 key points:
1.The FDA recognizes that AI/ML applications exist at each stage of drug development. AI/ML may have applications at every stage of drug development, from drug discovery to drug manufacturing. AI/ML is being applied to data from real-world data (RWD) and digital health technology (DHT) to support drug development. The first section of the discussion paper summarizes the different ways AI/ML can be used in drug discovery, clinical and non-clinical research, post-marketing surveillance, and advanced drug manufacturing.
2.FDA has unique experience with AI/ML for drug development. In recent years, the FDA has seen an increase in drug and biologics submissions referencing AI/ML. In response, FDA has taken a number of actions, including the establishment of the CDER AI Steering Committee, the Innovative Scientific and Technological Approaches for New Drugs (ISTAND) pilot program, and the Model-Informed Drug Development (MIDD) pilot program. increase. For post-market safety surveillance, the CDER Sentinel System, the CBER Biologics Efficacy and Safety (BEST) System, and the CDRH National Evaluation System for Health Technology (NEST) efforts are focused on AI/ML to improve existing systems. I am looking for an approach.
3.FDA understands the importance of developing standards and practices for the use of AI/ML. The U.S. government and the international community have committed to stepping up efforts to accelerate AI innovation and adoption. Regulators and standards bodies have developed and published standards to facilitate progress in ethical AI. For example, in August 2019, the National Institute of Standards and Technology (NIST) published “U.S. Leadership in AI: Plans for Federal Government Engagement in the Development of Technical Standards and Related Tools.” Additionally, in October 2021, the FDA, Health Canada and the UK Medicines and Healthcare Products Regulatory Agency (MHRA) will jointly inform the development of Good Machine Learning Practices (GMLP) for Medical Devices using AI/ML. announced 10 guidelines to
Four. FDA has identified important questions about AI/ML in drug development.of FDA aims to initiate discussions with stakeholders in three key areas and will provide specific questions to solicit feedback.
Human-driven governance, accountability and transparency
- Which specific use case or application of AI/ML in drug development requires greater regulatory clarity?
- In your experience, what are the main barriers and facilitators of transparency when (and under what circumstances) AI/ML is used during the drug development process?
- How are pre-specification activities managed and changes acquired and monitored to ensure the safe and effective use of AI/ML in drug development?
Data quality, reliability and representativeness
- What additional data considerations are there for AI/ML in the drug development process?
- What are some key practices that stakeholders use to ensure data privacy and security?
- What processes do developers use to identify and manage bias?
Model development, performance, monitoring and validation
- What practices and documentation are used to inform and record data source selection and inclusion or exclusion criteria?
- In what use cases are stakeholders addressing explainability, and how do you balance performance and explainability considerations?
- What are some examples of current tools, processes, approaches, and best practices that stakeholders are using: choosing model types and algorithms for specific uses, validating models and measuring their performance in specific contexts? Deciding when to use a particular approach, assessing transparency, explainability, improving model transparency, etc.?
Five.FDA wants your feedback. FDA is seeking feedback on the opportunities and challenges of using AI/ML in drug and medical device development. FDA has included a set of questions for feedback in the discussion paper, and workshops are planned with stakeholders to provide opportunities for further engagement. Comments must be submitted by August 9, 2023 (docket number FDA-2023-N-0743).
