Ongoing drug safety monitoring is critical to containing risk and keeping patients safe, and with so many channels for reporting adverse events, the workload in processing cases is rising exponentially. But fit-for-purpose PV isn’t just about absorbing those extra volumes without driving up costs; it is also about maximising the positive impact of products for patients. This is especially true as ambitious new therapies enter markets with less predictable outcomes and side-effects over time. These converging requirements make harnessing AI non-negotiable, says Qinecsa’s Adam Sherlock.
As pharmacovigilance (PV) demands have soared over the years, the priority has been to optimise processing, handle more adverse event (AE) cases for less. Outsourcing arrangements, and use of technology, have been geared largely to enabling those improved efficiencies.
But as a raft of critical new therapies and drug applications enter the market, with less predictable long-term effects, there are other priorities driving the PV technology agenda. Such products include GLP-1 receptor agonist/weight loss injections (WHO plans to officially support their use to treat obesity in adults). They also include messenger ribonucleic acid (mRNA) technology in approaches to cancer. Then there are the pioneering COVID-19 vaccines of 2020–2021 that were approved at speed for use by significant populations around the world – populations that still need to be closely observed for emerging side-effects.
In such contexts, there is a heightened need to detect issues and emerging patterns swiftly. This is compounding the need for technology-enabled PV transformation – as a means to hone accuracy, precision and speed, in addition to operational efficiency. It is for these reasons combined that AI is starting to make its mark as a mature and viable solution to AE case processing.
The Rise of PV-Specific AI Applications
A number of AI solutions designed for PV are available now, and being put through their paces by leading pharma organisations, with promising results. Generally, AI-based PV tools have been shown to reliably handle large volumes of data, extract key information from various sources, and even detect subtle patterns that might be missed by human reviewers. (According to the US FDA, implementing AI in PV has improved the detection of potential drug risks by over 25%.)
The need for pharma companies to diversify as a means of new brand differentiation and long-term growth, on top of their already soaring AE case volumes, gives them little choice about harnessing next-generation, AI-driven process automation.
Applying AI: Best Practice Approaches
The best approach to implementing AI will depend on a company’s existing PV ecosystem, the volumes of work it underpins, and the existing technology infrastructure that’s in place. However impressive the promise from the tech vendor, individual organisations will need to understand how a solution would fit their ‘as is’ set-up, and also how its deployment might translate into tangible benefits, including cost reduction and improved productivity.
