The European Medicines Agency (EMA) has released a draft Reflection Paper exploring the role of artificial intelligence (AI) in the lifecycle of medicines. This paper emphasizes the value of AI in healthcare’s digital transformation and recognizes its potential to support data acquisition, analysis, and interpretation throughout the medicinal product lifecycle. However, it underscores the importance of using AI correctly and in compliance with existing data requirements.
EMA’s paper acknowledges the increasing use of AI and machine learning (ML) tools in clinical decision-making and imaging analysis. It advises marketing authorization applicants and holders on incorporating transparent, accessible, validated, and monitored AI and ML mechanisms across all stages of the medicines’ lifecycle. EMA stresses the responsibility of organizations to ensure that algorithms, models, datasets, and data processing pipelines align with ethical, technical, scientific, and regulatory standards.
The draft paper covers various stages of the medicines’ lifecycle where AI and ML may impact the benefit-risk of the medicinal product. EMA recommends developers engage with regulators as early as possible, especially when significant regulatory impact or risk is involved. For instance, when using AI/ML tools to select clinical trial participants or analyze trial data, the sponsor must adhere to statistical principles, ethical rules, and technical aspects. Marketing authorization applicants need to convince EMA or national authorities of the trustworthiness, reliability, and robustness of data obtained through AI/ML.
Technical parameters for AI/ML system usage throughout the medicinal product lifecycle are also addressed in the paper. Additionally, EMA evaluates AI/ML medical devices used in clinical trials to determine their capacity to generate reliable data supporting marketing authorization applications. The regulatory status of such AI/ML devices will also be assessed by relevant authorities alongside the forthcoming EU AI Act.
Ethical principles and risk assessments are emphasized, promoting compliance with trustworthy AI guidelines and relevant data protection legislation. The paper provides valuable guidance and recommendations for marketing authorization applicants and holders, including the systematic conduct of early-stage impact analyses.
With a comprehensive list of relevant guidance sources, EMA highlights the applicability of many recommendations and best practices to AI/ML. This further underscores the need for compliance throughout the medicinal product lifecycle.
The EMA’s draft Reflection Paper on AI’s role in medicine lifecycle is open for consultation until December 31, 2023. EMA will also host a workshop in November 2023 to discuss the draft paper in detail. Feedback from the consultation will aid in finalizing the paper and developing future comprehensive guidelines.
This Reflection Paper from EMA signifies a significant step toward ensuring the responsible and effective integration of AI in healthcare. By providing clarity on expectations and standards, EMA aims to facilitate the adoption of AI/ML technologies while safeguarding patient safety and product quality. As this paper evolves into finalized guidance, stakeholders in the pharmaceutical and healthcare industries can anticipate enhanced understanding and implementation of AI in their respective domains.
