Artificial Intelligence in Pharmacovigilance: Scoping Points to Consider.

Artificial intelligence (AI), a highly interdisciplinary science, is an increasing presence in pharmacovigilance (PV). A better understanding of the scope of artificial intelligence in pharmacovigilance (AIPV) may be advantageous to more sharply defining, for example, which terms, methods, tasks, and data sets are suitably subsumed under the application of AIPV. Accordingly, this article explores relevant points to consider regarding defining the scope of AIPV and offers a potential working definition of the scope of AIPV.

Pharmacovigilance during the pre-approval phases: an evolving pharmaceutical industry model in response to ICH E2E, CIOMS VI, FDA and EMEA/CHMP risk-management guidelines.

Pharmacovigilance science has traditionally been a discipline focussed on the postmarketing or post-authorisation period, with due attention directed towards pre-clinical safety data, clinical trials and adverse events. As the biological sciences have evolved, pharmacovigilance has slowly shifted toward earlier, proactive consideration of risks and potential benefits of drugs in the pre- and peri-approval stages of drug development, leading to a maturing of drug safety risk management. Further advances in biology, pharmacology and improvements in computational applications to medicine have led to the development of more complex medicines previously unobtainable and have also permitted a more thorough assessment of risks and potential benefits even earlier in the development process. Elevated public concern with the safety of more sophisticated medicines, combined with new science, have led pharmaceutical innovators, regulators and healthcare professionals to collaborate to develop guidelines, which drive enhanced pharmacovigilance and safety risk management earlier in drug development. In this paper, we review international guidelines on pharmacovigilance planning applicable to the pre-approval phases of medicines development and provide author opinion on these guidelines' potential drug safety implications. We discuss the possible evolution of a pharmaceutical industry model to respond to these guidelines; a view on multidisciplinary safety management teams is provided to encourage refinement of safety-signal identification and risk assessment early in drug development and to communicate important safety concerns to internal research efforts, patients, investigators and regulators. We further describe these functions in the context of the complexities of vulnerable populations, including the example of medicines research for paediatric populations. We also discuss the special role of epidemiology in pre-approval drug development and the impact on epidemiological science of changes to the pharmacovigilance paradigm.