“Artificial Intelligence” for Pharmacovigilance: Ready for Prime Time?

Abstract

There is great interest in the application of ‘artificial intelligence’ (AI) to pharmacovigilance (PV). Although US FDA is broadly exploring the use of AI for PV, we focus on the application of AI to the processing and evaluation of Individual Case Safety Reports (ICSRs) submitted to the FDA Adverse Event Reporting System (FAERS). We describe a general framework for considering the readiness of AI for PV, followed by some examples of the application of AI to ICSR processing and evaluation in industry and FDA. We conclude that AI can usefully be applied to some aspects of ICSR processing and evaluation, but the performance of current AI algorithms requires a ‘human-in-the-loop’ to ensure good quality. We identify outstanding scientific and policy issues to be addressed before the full potential of AI can be exploited for ICSR processing and evaluation, including approaches to quality assurance of ‘human-in-the-loop’ AI systems, large-scale, publicly available training datasets, a well-defined and computable ‘cognitive framework’, a formal sociotechnical framework for applying AI to PV, and development of best practices for applying AI to PV. Practical experience with stepwise implementation of AI for ICSR processing and evaluation will likely provide important lessons that will inform the necessary policy and regulatory framework to facilitate widespread adoption and provide a foundation for further development of AI approaches to other aspects of PV.

The evolution of pharmacovigilance ecosystems: Does Moore's law invite the use of Occam's razor?

“…Lack of qualitative databases, weak technology, insufficient human resources and insufficient support from governments are just a few examples 46 . While current algorithms are not sufficient for complete automation, they can still be applied to improve efficiency, value and consistency if included into a system with human intelligence in control 47,48 …”

“…46 While current algorithms are not sufficient for complete automation, they can still be applied to improve efficiency, value and consistency if included into a system with human intelligence in control. 47,48 The value of mass media, social media and the Internet, as sources of case reports, has been debated and intensely investigated.…”

Exploring opportunities for AI supported medication error categorization: A brief report in human machine collaboration

“…More recent works have explored the contributions of NLP to support adverse event reporting workflows and operational insights through the use of decision support systems and information visualization applications (Botsis et al, 2016;Spiker et al, 2020). These recent works aims to tackle the important challenges of translating technical advances into operations to support workflow (Ball and Dal Pan, 2022).…”

An open-source platform integrating emerging data sources to support multi-modal active pharmacovigilance

“…Hence, there is an increasing need for the development of new tools aiming to integrate "intelligent" data processing approaches to support drug safety (Trifirò et al, 2018;Bate and Hobbiger, 2020). Moreover, the emergence of additional data sources such as biochemical databases, electronic health records (EHRs), insurance claims or other "Real-World Data" (RWD) and social media (Hussain, 2021; Knowledge Base workgroup of the Observational Health Data Sciences and Informatics (OHDSI) collaborative, 2017) have led to relevant research initiatives (Natsiavas et al, 2019b;Ball et al, 2022). To this end, Machine Learning (ML) algorithms are also under investigation (Lee et al, 2022;Imran et al, 2022), including the use of Natural Language Processing (NLP) which is elaborated to identify ADR mentions in EHRs/clinical notes or other free text/unstructured data.…”