Bruce Palsulich, vice president of Product Strategy, Oracle Health Sciences, explains how AI can help pharmaceutical companies gather more important data and boost pharmacovigilance.
Is it possible to have too much of a good thing? That is the question the pharmaceutical industry is currently asking itself. In an effort to bolster pharmacovigilance programmes, drug companies have succeeded in gathering an enviable amount of patient data – so much so that they have found themselves struggling under the weight of it, particularly around adverse event reports.
Aided partly by the digitisation of reporting, partly by the growth of marketing outreach programs, pharmaceutical companies are seeing a 10-20% rise in adverse event reports. And it’s not just the volume of information that’s changing, but also the nature of it. Due to the convenience of logging adverse events via email, or on online forums, there has been an upsurge in patient reporting. But it also means that the information logged is much less detailed. Drug companies could receive a notice that a patient taking a particular medication has felt dizzy and jittery, for example, with no further context provided.
Pharmaceutical companies want to gather the maximum volume of data, but it is becoming difficult for them to convert it into meaningful conclusions. Fragmentary reports like the “dizzy and jittery” symptoms don’t mean anything on their own – it’s only after discovering an unusual pattern or other such reports that the findings may become significant. This means there is an interest in processing reports as quickly as possible, which is leading drug companies to outsource these activities to third party providers. But this only creates further challenges, as costs rise and quality diminishes.
If drug companies could learn to make proper use of the volumes of data they find themselves with, the wealth of information would quickly become an asset. The answer isn’t necessarily in outsourcing the additional workload; it’s more about boosting the speed, capacity – and quality – of the processing. Artificial Intelligence (AI) could be a compelling solution to achieve just that.
Pharmaceutical companies, and healthcare practitioners in general, have so far been wary to introduce AI into their work. And yet, the benefits of doing so would be immeasurable. For example, Oracle Health Sciences has recently been working with one pharmaceutical company to bring efficiency to its pharmacovigilance programme. Together, we decided to integrate AI into the company’s case handling, and an early pilot has demonstrated the potential of decreasing its total processing time by up to 50%. Not only will the AI help the company handle cases in as little as one minute, it has also demonstrated an unprecedented accuracy of 90%.
Ultimately, bringing speed into pharmacovigilance allows drug companies to make sense of their findings quicker, putting them in a position to intervene when necessary. For example, a batch of medication can be recalled as soon as a drug is found to cause serious medical issues, helping to prevent a swathe of adverse reactions.
However, it’s not just the speed of processing, but also the depth of information processed, that is so powerful with AI. Processing data quicker quite simply allows more data to be processed. In turn, this enables practitioners to focus their efforts on more informed observations, for example detecting unusual patterns, or perceiving disproportionate reactions. AI can assist these observations by also integrating third-party data sources such as patient healthcare data or academic literature in areas like chemical structure, pharmacodynamics or genomics. This helps practitioners make more finely-tuned evaluations, cross-checking patterns of data against the latest scientific research.
Further, pharmacovigilance isn’t just about determining when a drug should be taken off the market – it’s also about pinpointing when drugs are suitable for use. By leveraging the analytical capabilities of AI, practitioners can determine which subsets of patients respond well to certain medications, even if other subsets don’t. This helps drug companies avoid the complete market recall of medications and vaccines, when they could continue to benefit the lives of patients responding without symptoms.
With AI, pharmaceutical companies could turn their data from a challenge into an opportunity. Although the healthcare industry can understandably be risk-averse in the face of new technologies, it’s only a matter of time until AI becomes a commonplace solution in the day-to-day workings of every pharmacovigilance program. With the evidence trickling in from frontrunners like the company we’re working with, drug companies the world over are starting to see the real-world benefits that AI technology could bring to their own operations. Not only will AI bring speed, efficiency and accuracy to pharmacovigilance programs, but it will also help practitioners focus on what they do best – evaluate the safety and effectiveness of their pharmaceutical products.