AI/ML in Pharmacovigilance and Pharmacoepidemiology

Artificial Intelligence (AI) has emerged as a scientific field coming with great hopes to revolutionize various industries. However, AI also comes with great challenges which go well-beyond algorithmic performance issues, e.g., the validation of its results, explainability, risk management, etc. AI could have a profound impact within the domains of pharmacoepidemiology and pharmacovigilance, including the identification and validation of novel drug safety signals, drug development and drug safety management via the combination of heterogeneous data sources, and the application of novel algorithmic approaches. However, many challenges remain in the development of both symbolic AI (e.g., the use of Knowledge Engineering approaches, rule-based systems, ontologies, etc.) and non-symbolic AI (e.g. statistics-based Machine Learning - ML, neural networks, etc.), hindering its use in the industry.

This Research Topic, therefore, seeks research on the latest advancements on the use of AI for pharmacoepidemiology, pharmacovigilance, systems pharmacology, and related areas, aiming to present novel algorithmic approaches, innovative data representation schemes, and relevant software tools. We anticipate this special issue to contribute to the live discussion on how AI could support and advance drug safety processes.

Within the context of pharmacoepidemiology and pharmacovigilance, themes of interest include but are not limited to:

• Explainability/interpretability and visualization of AI methods/results.
• Real World Data/Evidence (RWD/RWE) and AI.
• Validation of “intelligent” systems and risk management of trustworthy AI systems.
• Combining AI with traditional approaches.
• Federated AI approaches for distributed data ecosystems.
• Usability and user experience and integration of AI in established operating procedures.
• AI and Knowledge Representation.