AI and Multiple Data Sources in Pharmacovigilance 

Saad Shakir writes:

I attended the DIA Europe meeting in Basel on 18-20 March 2025. 

The application of AI in all phases of the medicines lifecycle, including pharmacovigilance, was a major theme. The hopes are that AI will be part of signal detection and management, benefit/risk evaluation, decision-making support and more. Many of these will happen, though they will need some time to be applied. Moreover, the exact contribution of AI into all these processes will need time to be defined. 

Another area which was addressed at the meeting is the use of multiple data sources in pharmacovigilance, e.g. electronic health records. This is happening through the methods for integrating heterogeneous data sources, e.g. pooling, but systematic reviews and meta-analysis are yet to be fully developed for this purpose. For example, how to integrate electronic health records data with spontaneous reporting and information from social media. We will get there, but it will take longer than many people think. A huge amount of work is needed.